Skip Navigation

Federal Communications Commission

English Display Options

Commission Document

Media Ownership Study 2-Revised Study

Download Options

Released: July 27, 2011

CONSUMER VALUATION OF MEDIA AS A FUNCTION

OF LOCAL MARKET STRUCTURE1


Final Report to the Federal Communication Commission’s 2010 Quadrennial
Media Ownership proceeding - MB Docket No. 09-182.

Scott J. Savage and Donald M. Waldman
University of Colorado at Boulder
Department of Economics
Campus Box 256
Boulder, CO 80309-0256

Submitted to the Federal Communication Commission, May 30, 2011

Peer review received July 19, 2011

Final report submitted to the Federal Communication Commission, July 20, 2011

Abstract


This study examines the effects of media market structure on consumer demand and welfare. A
differentiated-product model is used to estimate demand for the local media environment,
described by the offerings from newspapers, radio, television, the Internet and Smartphone.
Results show that the representative consumer values more different viewpoints in the reporting
of information on news and current affairs, more information on community news, and more
information that reflects the interests of women and minorities. Consumers have a distaste for
advertising. Demand estimates are used to calculate the expected change in consumer welfare
from a merger between two television stations that lowers the amount of diversity and
advertising in local media environments. Welfare decreases following the merger, but these losses
decrease with the number of television stations in the market. For example, the average consumer
in a “small market” loses $0.11 per month, whereas the average consumer in a “large market”
loses $0.04 per month. These losses are equivalent to $6 million annually for all small-market
households in the U.S. and $1.4 million annually for large-market households. If the merger
occurs in all markets, aggregate consumer welfare losses would be about $91 million.



Key words: differentiated product, market structure, merger, welfare

JEL Classification Number: C9, C25, L13, L82, L96




1 The Federal Communications Commission (“FCC”) provided funding for this research. We are grateful to
Jessica Almond, Jonathan Levy and Tracy Waldon for their assistance during the completion of this project.
Nicholas Flores, Edward Morey and Gregory Rosston provided helpful comments and contributions. Any opinions
expressed here are those of the authors and not those of the FCC.
0


0.

Executive Summary

As part of the Federal Communications Commission’s (“FCC”) 2010 Quadrennial Media
Ownership proceeding, we have been asked to help determine consumer valuations of the
localism and diversity features of their local media environment. We estimate a differentiated-
product model of demand that measures consumer benefits from their local media environment.
We use our demand estimates to calculate the effects on expected consumer welfare from a merger
of two television stations that results in quality differences in localism and diversity between the
pre- and post-merger markets. This report outlines our methodology, sample and results.
We use data from a sample of 5,131 respondents, obtained from a nationwide survey
conducted in March 2011, to estimate consumer demand for their local media environment.
Local media environment describes the offerings from newspapers, radio, television, the
Internet and a web-enabled mobile telephone. Household data, obtained from choices in a real
market and an experimental setting, are combined with a discrete-choice model to estimate the
marginal willingness-to-pay (WTP) for improvements in four features of the media
environment: the diversity of opinion in reporting information; the amount of information on
community news and events; the coverage of multiculturalism, that is, ethnic, gender, and
minority related issues; and the amount of advertising.
Respondents are presented with eight scenarios, and in each scenario, must choose
between a pair of new media environment options that differ by the levels of their features.
The information in these choices is enriched with market data by having respondents indicate
whether they would stay with their current media environment at home or switch to the new
option they had just selected. The parameters of the representative consumer’s utility function,
and their WTP, are then estimated from all observed choices.
1


Results show that diversity of opinion and community news and events are important
features of the local media environment. The representative consumer is willing to pay $13 per
month for more viewpoints in the reporting of information on news and current affairs, and $14
per month for more information on community news and events. Consumers also value more
information that reflects the interests of women and minorities, although the WTP for this is
relatively small at about $2 per month. Consumers have a distaste for advertising and are
willing to pay $8 per month for a decrease in the amount of advertising in their overall media
environment. Willingness-to-pay for diversity of opinion and community news increase with
age, education and income, while WTP for multiculturalism decreases with age. White male
consumers do not value the multiculturalism feature.
We use our demand estimates to calculate the changes to consumer welfare from a merger
between two television stations that results in quality differences in localism and diversity in the
pre- and post-merger markets. We conduct a simple experiment that simulates the merger by
reducing the number of independent television voices in the market by one, all other things held
constant. Diversity of opinion, coverage of multicultural issues, and the amount of advertising
decrease following the merger. These changes in the quality of local media environments lead
to welfare losses, but these losses decrease with the number of television stations in the market.
For example, the average consumer in a “small market” loses $0.11 per month, whereas the
average consumer in a “large market” loses $0.04 per month. These losses are equivalent to $6
million annually for all small-market households in the U.S. and $1.4 million annually for all
large-market households. If the merger occurs in all markets, aggregate consumer welfare
losses would be about $91 million.


2


1.

Introduction

Two goals of U.S. media policy have been ensuring there are sufficient opportunities for
different, new and independent viewpoints to be heard (“diversity”), and that media sources
respond to the interests and needs of their local communities (“localism”). An interesting
question is how do media ownership rules enhance diversity and localism in a way that satisfies
consumers? As part of the FCC’s 2010 Quadrennial Media Ownership proceeding, we have
been asked to help determine consumer valuations of their local media environment. Local media
environment describes household consumption of information on news and current affairs from
newspapers, radio, television, the Internet, and a web-enabled mobile telephone we will call
Smartphone. This report outlines a methodology that measures satisfaction by how much
consumers value the diversity and localism features of their local media environment, and then
calculates how these valuations vary between different market structures. We do not provide
policy recommendations.

Media can be crucial for democracy.2 Because information on news and current affairs
can raise political awareness, and promote ideological diversity, many societies have charged
policy makers with ensuring there are sufficient independent voices within media markets. In the
U.S., the FCC has traditionally limited the amount of common ownership of radio and
television stations, and the amount of cross ownership between newspapers, radio and
television stations serving the same community. When ownership limits prevent market share
from being concentrated around a few corporations, standard theory predicts that competition
between many independent media sources can promote diversity of opinion, and incent owners
to respond to the interests and needs of their local communities.

2 Media is also important for economic development. By providing an efficient flow of information, media sources
reduce transaction costs and alleviate the need for economic regulation of many sectors of the economy.
3


More recently, legislators and the FCC have refocused their attention on market forces, for
example, consumer preferences and new media, such as satellite radio and television, the Internet,
and Smartphone, in order to deliver their competition, diversity and localism goals. The
Telecommunications Act of 1996 (“Act”) relaxed the limit on the number of radio and
television stations a firm could own nationwide, and permitted greater within-market common
ownership by allowing a firm to own more local radio stations. The Act also required the FCC
to review its ownership rules every four years and to “determine whether any of such rules are
necessary in the public interest as the result of competition.” Given the increase in choices
through new media, supporters of greater ownership concentration argue that traditional media
should be free to merge and use the efficiencies to provide more diverse and local
programming. Opponents question whether such efficiencies are achievable, and argue that
consolidated media corporations are not flexible enough to serve the interests and needs of
local and minority communities.3 Furthermore, many segments of the population to do not
have access to new media and even if they did, most of the original news on the Internet, for
example, is originated by newspapers, radio and television.4 5

Formal evaluation of these arguments requires, among other things, measurement of the
expected societal benefits that arise from increased media diversity and localism. For example,
policy makers may want to use the most recent estimates of demand to measure consumer

3 As noted by Owen and Wildman (1992), economies of scale in program distribution support the supply of non-
local content. Given a fixed cost of producing news content, multiple station owners can spread these costs over
more stations by distributing the same, non-locally oriented content across many communities. However,
economies of scale can also support the supply of more minority content. Siegelman and Waldfogel (2001) argue
that because individuals with similar tastes help defray the fixed costs of programming they all prefer, minority
consumers will derive benefits from being in the same, larger market as others with similar preferences.
4 U.S. Census Bureau (2009) data show that 64 percent of households had Internet access at the end of 2009. Data
from Pew Internet and American Life surveys show that about 78 percent of adult Americans use the Internet at
May, 2010 (http://www.pewinternet.org/Static-Pages/Trend-Data/Internet-Adoption.aspx">See http://www.pewinternet.org/Static-Pages/Trend-Data/Internet-Adoption.aspx). About 24 percent
of the 234 million mobile phone subscribers owned a Smartphone at August, 2010 (ComScore, 2011).
5 During 2009, Pew Research Center (2010) monitored 53 Baltimore newspapers, radio and television stations,
their associated web sites, as well as Internet-only web sites. They found that traditional media accounted for 93
percent of the original reporting or fresh information on six major news stories during the week of July 19 to 25.
4


satisfaction with the various features of their local media environment. Because households do
not have identical preferences, they may also want to see how consumer valuations for specific
media features vary with observable demographics such as age, education, gender, income, and
race, and with differences in market structure. The economic construct of willingness-to-pay
(WTP) can provide a theory-based, dollar measure of the value consumers place on their local
media environment, as well as the amount they would be willing to pay for improvements in
the individual features that comprise their environment. Since media environment is a mixture
of private and public goods, indirect valuation methods, such as those used in the environmental
and transportation choice literature, are appropriate.
This report uses data from a nationwide survey conducted during March 2011 to
estimate a differentiated-product model of consumer demand for their local media environment.
The report expands the work of Savage and Waldman (2008) and Rosston, et. al. (2010) by
combining household data, obtained from choices in a real market and an experimental setting,
with a well-specified discrete-choice model to estimate the marginal WTP for improvements in
four local media environment features. The features are the:
 diversity of opinion in reporting information (DIVERSITY OF OPINION);
 amount of information on community news and events (COMMUNITY NEWS); and
 coverage of multiculturalism, that is, ethnic, gender, and minority related issues
(MULTICULTURALISM); and
 amount of advertising (ADVERTISING).
We measure consumer satisfaction with diversity in media markets by their WTP for
DIVERSITY OF OPINION and MULTICULTURALISM. We measure consumer satisfaction
with local programming in media markets by their WTP for COMMUNITY NEWS. We
5


measure the full cost of their media environment by their monthly payments for media sources
(COST) and the amount of advertising.
The empirical methodology proceeds as follows. First, we designed a survey that
describes the local media environment by the offerings from newspapers, radio, television, the
Internet and Smartphone. The survey was fielded on a nationally representative sample that
adequately covered the range of market structures in U.S. television markets, as defined by the
FCC (2011). A carefully designed choice experiment manipulates the features for a series of
hypothetical media options to obtain the optimal variation in the data needed to estimate the
marginal utility parameters precisely. Respondents are presented with eight choice scenarios.
In each scenario, they choose between a pair of new media environment alternatives that differ
by the levels of their features. The information in these choices is enriched with market data by
having respondents indicate whether they would stay with their current media environment at
home or switch to the new option they had just selected.
Next, we used the data from these choices and the method of maximum likelihood to
estimate the structural parameters (“marginal utilities”) of the representative household’s utility
function. These structural parameters are used to calculate consumer valuations (and WTP) for
each of the various features of their media environment, and the variation in consumer
valuations by demographic groups.
Finally, we used these estimated consumer valuations and government furnished
information from the FCC (2011) to calculate the impact on expected consumer benefits from a
change in media ownership. By relating consumer valuations of DIVERSITY OF OPINION,
MULTICULTURALISM, and COMMUNITY NEWS to measures of market structure, we are
6


able to indirectly assess the extent to which ownership rules address the FCC’s goals of
diversity and localism.
Our empirical results show that the average price for a media environment was about
$111 per month. Diversity of opinion and community news and events are important features
of the local media environment. The representative consumer is willing to pay $13 per month
for more viewpoints in the reporting of information on news and current affairs, and $14 per
month for more information on community news and events. Consumers also value more
information that reflects the interests of women and minorities (i.e., multiculturalism), although
the willingness to pay for this is relatively small, about $2 per month, and is less precisely
estimated. Consumers have a distaste for advertising and are willing to pay $8 per month for a
decrease in the amount of space and/or time devoted to advertising in their overall media
environment. Willingness-to-pay for diversity of opinion and community news increase with
age, education and income, while WTP for multiculturalism decreases with age. White
consumers do not value the multiculturalism feature of their local media environment. In
contrast, non-white consumers are willing to pay about $5 per month for more information that
reflects the interests of women and minorities. More specifically, non-white males and non-
white females are willing to pay about $3.50 and $6 per month, respectively, for more
information that reflects the interests of women and minorities.
Our results also suggest declining marginal utility with respect to diversity of opinion,
multiculturalism and community news. The representative consumer is willing to pay about
$20 per month for an improvement in diversity of opinion (or community news) from a low to
a medium level defined by us in the survey (see below), but only an additional $6 to move to a
high level of diversity of opinion (or community news). Consumers value an improvement in
7


information that reflects the interests of women and minorities from low to medium about the
same as an improvement from low to high, that is, about $4 per month for each. In other
words, consumers would like some multiculturalism, but are unwilling to pay for a high level
when they feel they have adequate coverage.
We use our demand estimates to calculate the changes to consumer welfare from a merger
between two television stations that results in quality differences in localism and diversity in the
pre- and post-merger markets. We conduct a simple experiment that simulates the merger by
reducing the number of independent television voices in the market by one, all other things held
constant.6 Diversity of opinion, coverage of multicultural issues, and the amount of advertising
decrease following the merger. These changes in the quality of local media environments leads
to welfare losses, but these losses decrease with the number of television stations in the market.
For example, the average consumer in a “small market” (i.e., five television stations) loses $0.11
per month, whereas the average consumer in a “large market” (i.e., 20 television stations) loses
$0.04 per month. These losses are equivalent to $6 million annually for all small-market
households in the U.S. and $1.4 million annually for all large-market households. If the merger
occurs in all markets, aggregate consumer welfare losses would be about $91 million.
The report is organized as follows. Section 2 reviews the previous literature. Section 3
describes the random utility model of media environment choice. The econometric method
used to estimate the random utility model, calculate WTP, and to indirectly measure the value
to society from a change in media market structure is contained therein. The experimental
design, survey questionnaire and data are described in Section 4. Section 5 presents demand
estimates and calculates consumer valuations, and Section 6 presents a merger simulation that
predicts how these valuations vary between different market structures. Section 7 concludes.

6 See page 25 and page 26 for a description of how the number of independent television voices is measured.
8


2.

Literature Review

Numerous studies in communications, economics and political science examine the effects of
technology change and ownership rules in media markets. These studies can be classified into
two broad themes: new media and the information environment; and ownership, competition
and the supply of programming. We review the relevant papers from these themes below.

2.1
New media and the information environment
Baum and Kendall (1999) present ratings data that showed the average percentage of
households who watched prime-time presidential television appearances declined from 48
percent in 1969 to 30 percent in 1998. They offer two explanations for this trend; the rise of
political disaffection and the growth of cable television. Using data from the National Election
Study (NES), and controlling for demographics and political affection, Baum and Kendall
estimate the effect of cable television on the individual’s probability of viewing the 1996
presidential debate. They find that cable subscribers were less likely to have viewed the second
debate by nine percentage points and conclude that because they have more viewing choices,
cable subscribers with a preference for entertainment do not stay tuned to the President.
Because of the increased availability of entertainment content, Prior (2002) argued that
people with a preference for entertainment now consume less political information than they
used to. He uses data from the NES and Pew Media Consumption Surveys from 1996 and
2000 to examine the relationship between access to cable television and the Internet, and
knowledge about congressional house incumbents. Using a logistic regression model that
controls for demographics and general political knowledge, Prior finds that among people who
9


prefer entertainment, greater access to new media is associated with lower recall of house
candidates names and their voting record.7
Using survey data from over 16,000 adults in the Washington, D.C. area between 2000
and 2003, Gentzkow (2007) estimated how the entry of online newspapers affected the welfare
of consumers and newspaper firms. Estimates from a structural model of the newspaper
market, comprising of The Washington Post’s print and online versions and The Washington
Times, suggest that the online and print versions of the Post are substitutes. The online
newspaper reduced print readership by 27,000 per day at a cost of $5.5 million in print profits.
For consumers, the entry of the online newspaper generated a per-reader surplus of $0.30 per
day, equivalent to about $45 million in annual consumer welfare.
Byerly et. al. (2006) interviewed 196 subjects in the D.C. area during 2006 to
investigate the consumption of news by minorities. They found that commercial television
news and newspapers were the most important sources of local news and information, while
radio and the Internet were among the least important. Subjects who identified the Internet as a
new media source indicated that it was a supplement to other traditional media, rather than a
sole source of news. The most popular preferences for important media sources were
“completeness of information” and “a stronger focus on local issues with a minority angle.”
Nielson Media Research (NMR) and Pew Internet and American Life provide results
from periodic surveys of large numbers of households that provide a timeline for studying
consumer preferences and new technologies in media markets. For example, NMR (2007)
surveyed over 100,000 households during May and June, 2007 to find out how they get their
news and information. They found that new media, such as cable television and the Internet,

7 Relative entertainment preference is entertainment viewing (i.e., average daily viewing of various combinations
of Jeopardy, Wheel of Fortune, Dr. Quinn, Medicine Women, Entertainment Tonight, Hardcopy, and MTV) as a
share of entertainment plus news viewing (i.e., average daily viewing of local national news).
10


have made substantial inroads into traditional media’s market share. Cable news channels,
such as CNN and MSNBC, were the most important household sources for breaking news, in-
depth information on specific news and current affairs, and national news, while the Internet
was the second most important source for in-depth information on specific news and current
affairs. Broadcast television stations, such as ABC, and local newspapers remain the most
important sources of local news and current affairs.
Purcell (2011) provided survey results from 2,251 households that show that almost half
of all American adults get at least some of their local news and information on their cellphone
or tablet computer. Compared with other adults, these mobile local news consumers are
younger, have higher income, live in urban areas, and tend to be parents of minor children.
One-quarter of mobile local news consumers report having an “app” that helps them get
information about their local community. Because local app users also indicate they are not
necessarily more interested in general or local news than other adults, these findings suggest
that the convenience of mobile online news consumption, rather than quantity, is an important
aspect of their preferences.

2.2
Ownership, competition and programming
An alternative theme in the literature focuses on the effects of market structure, as
measured by media ownership and market competition, on the quantity and quality of
information supplied by newspapers, radio and television stations. Dubin and Spitzer (1993)
investigate the relationship between minority ownership and radio station format. Using a
cross section of stations at 1987, they find that minority-owned stations are more likely to
broadcast minority-targeted programming than white-owned stations. However, they are
11


unable to show whether minority ownership increased the total available amount of minority-
targeted programming or replaced an equivalent amount of white-owned, minority-targeted
programming. Siegelman and Waldfogel (2001) control for this “crowding out” effect with a
cross section of 244 radio markets at 1993 and 1997. They find that minority ownership
increased the net amount of minority-targeted programming. Even though most minority-
targeted stations are white-owned, markets with more minority-owned stations also have more
minority-targeted stations, which means that minority-owned stations added to the total
programming available to minority listeners.
Alexander and Brown (2004) estimate the impact of local ownership on the provision of
local stories during each television station’s half-hour local news broadcast. News is defined as
local when the story takes place in the station’s television market, and the mean consumer in
the market perceives the story as local relative to consumers in other markets. Using data from
275 stations during 1998, their ordinary least squares (OLS) estimates show that stations with
their headquarters located within the television market (“local ownership”) provide about five
and a half more minutes of local news, and over three more minutes of local on-location news.8
They suggest that because they have lower costs of monitoring local events and personnel,
local owners can efficiently cover more local news.
Yan and Napoli (2006) examine how market structure affects the provision of local
public affairs programming. They use a two-week sample of the schedules from 285 television
stations in 2003, provided by Tribune Media Services (TMS), to calculate the minutes of local
programming provided by each station. Because some stations never provide local
programming, while others provide none during the sample period, they estimate a zero-

8 Using time-series, cross-section station data from 2002 to 2005, Shiman (2007) obtains a similar result for total
scheduled news and public affairs programming. Among other things, he finds that TV stations owned by a big
four network provided 13 percent more news programming.
12


inflated negative binomial model. The model relates local programming to station controls, the
number of television stations in the market, whether the station is commercial, private or
public, and several measures of ownership (e.g., affiliation with one of the big four networks,
national audience share). Market competition did not correlate with the quantity of local
programming. They concluded that increased competition is unlikely to force individual
stations to provide more local programming and that commercial broadcasters are more likely
to cede that programming to their public counterparts.
Crawford (2007) employs TMS data from 2004 to 2006 to analyze the relationship
between the ownership of television stations and the quantity and quality of their programming.
His sample is comprised of 1,479 stations, for each of the three years, from 27 broadcast
networks and 192 cable networks. Quantity was measured by the number of minutes supplied
in each of the following categories: local news and public affairs; minority; children; family;
indecent; violent; and religious. Quality was measured by the number of minutes of advertising
included in the programming.9 Crawford estimates several econometric specifications that
relate quantity and quality to local ownership, whether the parent corporation of the television
station also owned a newspaper or radio station within the same market, and various other firm
and market controls. Two interesting findings are that newspaper cross-owned stations provide
more local news, and that ownership has no economic effect on advertising.
Milyo (2007) examined the effects of newspaper and radio cross ownership on the
quantity and quality of local content for late-evening local television news during early
November, 2006. Using data from 104 stations for 27 markets, and controlling for station
characteristics, his within-market estimates show that newspaper cross-owned stations supply

9 TMS provided data on every channel in the U.S., along with the “program type” and “category” fields that were
used to classify the programming into the seven types. Data on the number and length of advertisements in each
program were obtained from TNS.
13


about seven to ten percent more local news than non-cross-owned stations. On average, cross-
owned stations also provide about 25 percent more coverage of state and local politics. He
found no difference in quality between cross-owned stations and other major network-affiliated
stations in the same market, where quality is measured by political slant, although there is
evidence that the slant of local news is associated with voting preferences.10 Because supply
has increased without loss of quality, Milyo suggests that market forces, specifically consumer
preferences and economies of scope, are driving station’s decisions on local news coverage.
Chipty (2007) studied the effects of consolidation on non-music programming for
commercial radio stations and found that more concentrated markets are associated with less
program variety. Using market- and station-level data from 2005, she estimates several
empirical models of the number of formats offered and format concentration.11 Chipty shows
that more concentrated markets have less “pile-up” of stations on individual formats, and large
national radio owners offer more formats and have less pile up. Consolidation also affects the
composition of non-music programming, with owners of multiple local stations offering longer,
uninterrupted blocks of sports programming in the evening.
Using panel data from 1998 to 2001, Sweeting (2010) investigated the impact of within-
and between-market common ownership on programming and listeners for 1,095 contemporary
music radio stations.12 He found that commonly-owned stations within the same market and
music category differentiate their respective playlists by up to 20 percent and gain listeners.

10 Slant is measured by the speaking time allowed to candidates of either party, candidate coverage, partisan issue
coverage, and opinion polls favoring one party or the other.
11 Edision Media Research characterized the formats as: advertising; announcements; dead air/unknown;
entertainment; leisure or DJ banter; fundraising and charity; music; news; other; public affairs; religious (non-
music); sports; and static/interference.
12 Several papers have used the variation in ownership rules from the Act to examine whether common ownership
increased aggregate radio variety. Berry and Waldfogel (2001) compared the number of formats before and after
the Act, and find that common ownership increased variety. The FCC (2001),with an alternative format
classification, and Chambers (2003), with a cross section of playlist data, find evidence that common ownership
did not increase variety.
14


Such differentiation makes their stations more like rivals, who lose a similar amount of
listeners. Cross-market common ownership results in some homogenization of programming,
which is consistent with a common owner enjoying economies of scope in offering similar
programming in different markets. Because stations owned by large national corporations have
been able to play more commercials without losing listeners, Sweeting suggests that
homogenization has increased quality as these companies use information from different
markets to identify which songs appeal to listeners. In effect, competition for listeners provide
stations with strong incentives to serve local preferences.
Gentzkow and Shaprio (2010) focus on how newspapers dedifferentiate themselves
with political slant. They construct an index of slant that measures the similarity of a
newspaper’s language to that of a congressional Republican or Democrat. Using a sample of
433 newspapers at 2005, and measuring market structure at the zip code level, they estimate a
model of newspaper demand and supply that accounts for slant and newspaper owner ideology.
Ideology is measured by the share of each newspaper firm’s total political contributions going
to Republicans. They find that readers have a preference for like-minded news, and that firms
respond to market forces by slanting their news towards preferences. By contrast, the identity
of the newspaper’s owner explains less variation in slant. They conclude that ownership
diversity may not be an important precondition for ideological diversity in the media.

2.3
Summary
Previous studies provide insights on consumer preferences for information on news and
current affairs, the quantity and quality of the information provided in the market, and how the
provision of information is affected by new media, ownership and market competition. Most of
15


these studies were based on data obtained prior to 2007, and most focus on measuring
programming outcomes for one on the media sources that comprise the local media
environment: newspaper, radio, television, the Internet, or Smartphones. Moreover, because
they employ a wide range of collinear market structure variables, many of the studies are
unable to identify any systematic effects from cross ownership and market competition.
This report uses the methodology described by Savage and Waldman (2008) and
Rosston, et. al. (2010), and survey data obtained during March, 2011 to estimate consumer
valuations for improvements in the DIVERSITY OF OPINION, MULTICULTURALISM and
COMMUNITY NEWS features of their local media environment. A new measure of market
structure is defined that counts the number of independent television voices within the market.
By investigating how estimated consumer valuations vary with the number of independent
voices in the market, the report helps shed light on the question of how media ownership rules
and merger guidelines enhance diversity and localism in a way that satisfies consumers.

3.

Methodology

3.1.
Research Questions
We specify a relevant performance metric with respect to the FCC’s diversity and localism
goals, and then analyze the impact of differences in media market structure on the performance
metric. Our performance metric for consumer satisfaction with diversity in media markets is
their willingness-to-pay for DIVERSITY OF OPINION and MULTICULTURALISM. Our
performance metric for consumer satisfaction with local programming in media markets is their
willingness-to-pay for COMMUNITY NEWS. The questions of interest are:
16


i.
what are consumers willing to pay for their local media environment features,
DIVERSITY OF OPINION, MULTICULTURALISM and COMMUNITY NEWS;
ii.
how does WTP vary between different demographic groups; and
iii.
how does WTP vary between different market structures?
A two-step approach is employed to answer the research questions. Step one will
estimate consumer demand and valuations for marginal improvements in the localism and
diversity features with choice data. Step two will estimate how a hypothetical merger between
two television stations affects the market’s provision of the diversity and localism features of
media environment alternatives. Expected consumer benefits are calculated from the difference in
their valuations for media environment alternatives provided before and after the merger.

3.2
Estimating Consumer Valuations
The random utility model is used to estimate marginal utilities and calculate WTP.
Survey respondents are assumed to maximize their household’s utility of the media
environment option A or B conditional on all other consumption and time allocation decisions.
A linear approximation to the household conditional utility function is:
U* = 1COST + 2DIVERSITY OF OPINION + 3COMMUNITY NEWS
+ 4MULTICULTURALISM + 5ADVERTISING +

(1)
where U* is utility, β1 is the marginal disutility of COST, β2, β3 and β4 are the marginal utilities
for the media environment features DIVERSITY OF OPINION, COMMUNITY NEWS and
MULTICULTURALISM, β5 is the marginal disutility of ADVERTISING, and  is a random
disturbance. Table 1 describes the levels of the features that comprise the media environments,
A and B. COST is the dollar amount the household pays per month for their home media
17


environment. That is, the total of monthly subscriptions to all media sources, plus any
contributions to public radio or public TV stations. DIVERSITY OF OPINION is the extent to
which the information on news and current affairs in the household’s overall media
environment reflects different viewpoints. COMMUNITY OF NEWS is the amount of
information on community news and events in the household’s overall media environment.
MULTICULTURALISM is the amount of information on news and current affairs in the
household’s overall media environment that reflects the interests of women and minorities.
ADVERTISING is the amount of space and/or time devoted to advertising in the household’s
overall media environment.
The marginal utilities have the usual partial derivative interpretation; the change in
utility, or, satisfaction, from a one-unit increase in the level of the feature. Given “more is
better”, our a priori expectation for DIVERSITY OF OPINION, COMMUNITY NEWS and
MULTICULTURALISM is β2, β3, β4 > 0. For example, an estimate of β2 = 0.2 indicates that a
one unit improvement in DIVERSITY OF OPINION, measured by a discrete improvement from
“Low = 1” to “Medium = 2”, increases utility by 0.2 for the representative household. A higher
cost and a higher amount of advertising provide less satisfaction so β1, and β5 < 0 are expected.
Since the estimates of marginal utility, such as an increase in utility of 0.2 described
above, do not have a understandable metric, it is convenient to convert these changes into
dollars. This is done by employing the economic construct of willingness-to-pay. For
example, the WTP for a one unit increase in DIVERSITY OF OPINION (i.e., the discrete
improvement from “Low” to “Medium”) is defined as how much more the local media
environment would have to be priced to make the consumer just indifferent between the old
18


(cheaper but with only one viewpoint) media environment and the new (more expensive but
with a few different viewpoints) media environment:
1COST + 2DIVERSITY OF OPINION + 3COMMUNITY NEWS
+ 4MULTICULTURALISM + 5ADVERTISING +  =
1(COST + WTPD) + 2(DIVERSITY OF OPINION + 1) + 3COMMUNITY NEWS
+ 4MULTICULTURALISM + 5ADVERTISING +


(2)
where WTPD is the willingness-to-pay for an improvement in DIVERSITY OF OPINION.
Solving algebraically for WTPD in equation 2 gives the required change in cost to offset an
increase of 2 in utility:13
WTPD = -2/1








(3)
For example, estimates of β2 = 0.2 and β1 = -0.01 indicate that the WTP for an improvement in
diversity of opinion from “Low” to “Medium” is $20 (= -0.2/0.01). Note that the model
specification in equation 1 implies that the representative household would also be willing to
pay the same amount ($20) for an improvement in diversity of opinion from “Low” to
“Medium” as it would to move from “Medium” to “High.” This constraint can be relaxed
during econometric estimation so that the marginal utility for an improvement in diversity of
opinion from “Medium” to “High” can be different from the marginal utility for an
improvement in diversity from “Low” to “Medium.”
This approach to estimating consumer valuations is used for all other features of the
local media environment. The WTP for COMMUNITY NEWS, MULTICULTURALISM and
ADVERTISING is the negative of the ratio of its marginal utility to the marginal disutility of
COST. In summary, the WTP construct provides a theory-driven, intuitive (dollar) measure of

13 The discrete-choice model actually estimates 2/σ and 1/σ, where σ is the scale parameter. The WTP calculation
is not affected by the presence of the scale parameter because –(2/σ)/(1/σ) = -2/1.
19


the value consumers place on their local media environment and the specific features that
comprise the environment. Moreover, by determining, for example, how much more valuable
the medium level of the feature is than the low level, it is possible to quantify household
benefits from changes in market conditions that affect the provision of the underlying features of
media environment alternatives (see Section 3.3).
Households may not have identical preferences. Preferences towards multiculturalism,
for example, may differ because of observable demographic characteristics, or may be not be
observable or measurable. It is possible to estimate differences in the marginal utility of all
features to different households by estimating the random utility model on various subsamples
of the data according to demographic variables, such as age, education, gender, income, and
race. We have this ability because of the large number of respondents answering our survey.
The utility of each media environment U* is not observed. What is known is which
option has the highest utility. For instance, when a respondent chooses the new media
environment option A over B and then the status quo (SQ) over A, it is assumed that
*
U >
*
U
A
B
and
*
U
>
*
U . For this kind of dichotomous choice data, a suitable method of estimation is
SQ
A
maximum likelihood (i.e., a form of bivariate probit) where the probability of the outcome for
each respondent-choice occasion is written as a function of the data and the parameters. The
experimental design uses variation both between and within subjects. The within subject
variation comes from the eight repeated A-B choice questions plus the follow up A or B vs. SQ
choice questions, i.e., the choice between current home media environment and option A or B.
The between subjects variation comes from the cross section of households that are surveyed.
For details on the econometric model and calculation of WTP, see Appendix A.

20


3.3
Consumer Valuations and Market Structure
Estimates of utility from a differentiated-product model are useful for measuring the
change in benefit, or consumer surplus, from a change in the features of a product alternative.
For example, Rosston et. al. (2010) calculate the benefit of broadband Internet service to rural
households by comparing their utility from before and after a change from a dial-up to a high-
speed connection. By comparing these valuations to the cost of service provision, policy
makers can make a more accurate judgment of the potential subsidy required, or not required,
for individual broadband adoption and/or deployment in rural areas. Alternatively, policy
makers may want to consider how a change in regulations or market structure affect the
market’s provision of product alternatives, and ultimately, how consumers value these changes
in features between alternatives. For example, Crawford (2000) estimated the expected welfare
gain per household from the changes in programming alternatives offered by cable television
systems in response to the price caps from the Cable Act of 1992. Similarly, Economides et. al.
(2008) quantified consumer welfare gains from the changes in product diversity and quality
resulting from entry in New York local telephone markets following the Act.
An interesting question in media markets is how FCC policies on market structure
affect the diversity and localism features of the media environment in a way that satisfies
consumers. Ideally, we would like to directly estimate the value to society from a change in
media ownership rules or merger guidelines. For example, consider the value from a merger
that reduces the number of independent television voices in a market as it impacts the single
media environment feature diversity of opinion (d). Below, we use this example to show what
cannot and what can be estimated with our data.
21


A simple representation of the diversity of opinion produced for a consumer in
television market j = 1, 2, …, J is:
d *
ij = α + δXj + γZi + vij






(4)
where d *
ij is the (unobserved) continuous index of respondent i’s diversity of opinion feature in
market j, Xj is a vector of variables that measure media market structure and the size of the
market, Zi is a vector of demographic and media source variables that control for heterogeneity
across consumers, α, δ, and γ are parameters to be estimated, and v is an independently and
identically normally distributed error term with zero mean and constant variance σ 2
v .

The respondent reports one of three possible levels for the diversity of opinion feature,
low, medium or high, based upon her level of d *
ij :
*


*



=







(5)

*


where µ is the unknown level of d *
ij above which respondents report a high level for diversity
of opinion. Given v is normally distributed, equation 4 and equation 5 represent the
conventional ordered probit model, which can be estimated by maximum likelihood (Zavoinia
and McElvey, 1975). However, since the scale of diversity of opinion is not observed, we are
unable to directly estimate the policy parameter of interest, δ. Rather, what is estimated is the
ratio of δ to σv, where σv is the standard deviation of the errors in equation 4.14
Suppose that the number of independent television voices in the market is represented
by X1 and its associated parameter in equation 4 is δ1. It is tempting to multiply the estimated
δ1 from equation 4 by the estimated 2 from equation 1, to calculate the value to society from a

14 The standard deviation has the same unit of measurement as d *
ij , the index of diversity of opinion.
22


change in the number of independent television voices that affects the market’s provision of
diversity of opinion. However, this would result in:
β

δ










(6)
σ
and not the actual effect that we are interested in. The problem is that we cannot observe the
scale of diversity of opinion.
An alternative approach to equation 6 is to use our sample estimates from equation 1
and equation 4 to predict how changes in the number of independent television voices affect
consumer’s expected benefit from the diversity of opinion in their local media environment.
Let the expected benefit from diversity of opinion for consumer i in market j be:
E[B
*
*
*
d(X1)] = PdL(X1)bdL + PdM(X1)bdM + PH(X1)bdH


(7)
where PdL(X1) is the probability that the consumer will be in the low diversity of opinion state,
PdM(X1) is the probability that the consumer will be in the medium diversity of opinion state,
PdH(X1) is the probability that the consumer will be in the high diversity of opinion state, and
b *
*
*
dL , bdM and bdH are consumer valuations for low, medium and high diversity of opinion.

We do not observe b *
*
*
dL , bdM and bdH . However, we are able to estimate from Step one
the value the consumer places on a move from low to medium diversity of opinion (ΔbdM), and
the value of a move from low to high diversity (ΔbdH). Substituting these valuations into the
expected benefit equation 7 gives:
E[B
*
*
*
d(X1)] = PdL(X1)bdL + PdM(X1)(bdL + ΔbdM) + PdH(X1)(bdL + ΔbdH).

(8)
The effect of a change in X1 on the expected benefit from diversity of opinion is:
= b * + (b * + Δb
(b * + Δb



) +

)






= + + )b * + Δb
Δb


+






23


= Δb
Δb

+






(9)




where
and
measure the effects of a change in X

1 on the predicted probability of



being in the medium and the high diversity of opinion states, and
+ + = 0.



Note that the marginal effects sum to zero, which follows from the requirement that the three
probabilities sum to one.15
Equation 9 provides the basis for calculating the value to society from a merger between
two television stations that affects the market’s provision of diversity of opinion in local media
environments. Estimates of ΔbdM and ΔbdH for the typical consumer are obtained from the


demand estimates and marginal WTP calculations in Step one. Estimates of
and
for


each individual consumer are obtained from the ordered probit model in Step two. Here, we
use our estimated coefficients from the ordered probit model and the sample data to calculate
the predicted pre-merger probability distributions for low, medium and high diversity of
opinion. Holding all other things constant, we then reduce the number of independent

15 The derivation of this result shows clearly that expected net benefit is a function of only willingness-to-pay for a
change out of the low level of a feature, and a function of only the changes in probability for the medium and high
levels of the feature. The other relevant quantities, the WTP for a change from the medium to the high level, and
the change in probability for the low level of the feature, are not present in the formula. Equation 9 is valid
essentially because of the two identities, Δb b + Δb , where Δb is the value the consumer places on
a move from low to high diversity of opinion, Δb is the value the consumer places on a move from medium to
high diversity of opinion and + + = 0, so that only two of the three quantities on the left hand sides
of equation 8 and equation 9 above are in play.
A hypothetical but concrete example may be useful. Suppose from Step one we find b = $5 and
b = $8. Further, suppose the initial distribution of ten individuals into the low, medium, and high regimes is
(PL0, PM0, PH0) = (0.3, 0.2, 0.5). As a result of a policy change, the new distribution is (PL1, PM1, PH1) =
(0.4, 0.4, 0.2), resulting in = 0.2 and = -0.3. The high regime has lost three members, the medium
regime has gained two members, and the low regime has gained a single member. According to equation 9,
expected net benefit is ($5  0.2) + ($8  -0.3) = -$1.40 per household. One way this can occur is if three
households in the high regime move to the low regime, at a cost of 3  $ -8 = $ -24, and two members of the low
regime move to the medium regime, at a benefit of 2  $5 = $10, for a total net benefit of -$14. But b = $5 and
b = $8 implies Δb = $3. Another way the final distribution can be obtained is if two members of the high
regime move to the medium regime and one member of the high regime moves to the low regime. This results in
the same net benefit calculation of (-$3  2) + ($ -8  1) = -$14.
24


television voices by one in the sample data to simulate the merger, and re-calculate the
predicted post-merger probability distributions for low, medium and high diversity of opinion.


The difference in predicted probabilities are used to form
and
. These calculations


are repeated for the multiculturalism, community news, and advertising features of the media
environment, and then scaled to reflect the general population. See Section 6 for a more
detailed description and example of the consumer benefit calculations.
FCC (2011) data are used to measure the media market structure variables in equation
4. There are many possible descriptions of media market structure reflecting the number of
newspapers, radio stations, and television stations, and/or the extent to which these individual
media outlets represent independent voices in the market place. Because we are interested in
measuring the effects of a television merger on consumer demand and welfare, we use a
parsimonious description of market structure, that is: the number of television stations in the
market (TV STATIONS); the number of independent television voices in the market (TV
VOICES); and the interaction between these two variables (TV STATIONS TV VOICES). The
addition of the interaction term makes it possible to examine the different impacts from a
merger in say, small versus large markets.
The FCC (2011) measure TV VOICES by first combining all the television outlets
within each individual market. The voiceprint of each outlet, that is, the listing of the unique
parent company identifiers of all attributable owners of an outlet, is then created, sorted
alphabetically, and duplicate voiceprints are eliminated. The parent identifier is then used to
count the number of voices in the voiceprint for each outlet. Voiceprints composed of a single
voice are added to the voice count of the market, while any voiceprint that includes one of the
voices counted at the previous stage of the calculation are eliminated. These are voices that are
25


not independent because their voice has been heard on another outlet. This process is
sequentially repeated based on the number of voices in the voiceprint.16
The survey questionnaire provides the household data used to construct the vector of
demographic and media source controls, Z. The vector Z measures the head of the household’s
age (AGE), education (EDUC), gender (GENDER), household income (INCOME), and race
(RACE). In addition, Z also includes 15 dummy variables to control for the different media
source combinations that can be comprised from a newspaper, radio, television, the Internet,
and Smartphone.17
Because it is comprised of demand- and supply-side terms for alternative features, the
parameters on market structure in equation 9 are difficult to sign and they remain an empirical
question. However, given positive marginal WTPs for the diversity and localism features, it is
possible to focus more closely on the effects of a change in the number of independent
television voices on the market’s provision of product features. Theory and previous studies
guide a priori expectations of the supply-side effects from a change in media market structure
(e.g., Spence and Owen, 1977; Owen and Wildman, 1992; Chipty, 1995; Waldfogel, 2004;
Shiman, 2007; and Sweeting, 2010). These expectations largely follow the efficiency versus

market power arguments surrounding merger analysis. For example, the finding of
< 0

indicates that a decrease in TV VOICES is associated with an increase in the provision of more
diversity of opinion in local media environments. This would be taken as evidence supporting

16 Note that the FCC uses several different measures of “independent voices” when discussing ownership rules,
and these measures include newspapers, radio and/or cable television. In Section 6, we report the results from
sensitivity analysis of the effects of a television merger on consumer demand and welfare using an alternative
description of market structure that controls for newspapers and radio stations.
17 The survey administrator provides sample households without Internet access with a laptop computer and free
Internet access to complete surveys. As such, the default media source combination is “Internet only”, which
comprises 1.1 percent of our sample. See Section 4.2 for the description of survey administration, and Table 6 for
the distribution of media source combinations.
26


the hypothesis that parent companies use the efficiency gains from mergers to supply
consumers with more diverse and local information, as measured by DIVERISTY OF OPINION
and COMMUNITY NEWS. Among other things, efficiencies arise from economies of scale in
the production of information for a larger geographic area (e.g., regional, domestic, and foreign
news), and from the economies of scope from sharing information with co-owned stations
and/or newspapers (Shiman, 2007). Size may also confer larger profits and increased ability to
attract low-cost capital and devote more resources to research and development.

The finding of
> 0 indicates that a decrease in TV VOICES is associated with the

provision of less diversity of opinion. This would be taken as evidence supporting the
hypothesis that consolidation leads to the softening of competition and the provision of less
diverse and more non-local information. More specifically, parent companies exercise market
power by providing fewer viewpoints and by limiting their investments in new and diverse
information programming. There may also be decreases in managerial efficiency that increase
the costs of production and affect the supply of programming. For example, Farrell (1983) and
Nalebuff and Stiglitz (1983) show that a decrease in the number of competitors in the market
limits the effectiveness of both efficiency-improving tournaments and relative-performance
evaluations. de Bettignies and Ross (2011) show that less competition can lead to control
problems and higher agency costs for owners monitoring managers.
There are several other theoretical explanations for the relationship between TV
VOICES (TV STATIONS) and product quality, with the sign dependent on how entry affects
consumer responsiveness to changes in quality (Spence, 1975; Schmalensee, 1979; Chen and
Schwartz, 2010; Matsa, 2010). As noted by Matsa, lower profit margins under more
competition reduces the immediate costs of losing a “sale” (e.g., revenue from the purchase of a
27


daily newspaper, or advertising revenue from viewers consumption of a daily radio and/or
television shows), so media outlets may shade quality. However, in the long run, competition
may raise the likelihood that unhappy consumers switch to a new media provider, so media
outlets improve quality.18 Chen and Schwartz also outline conditions whereby the incentive to
add a new, higher-quality product can be greatest under monopoly. The monopolist loses more
profit on the old product (“diversion effect”) but may earn more profit on the new one
(“coordination effect”) because it prices the old product in a way that internalizes the effect on
the new one. The relative strength of these opposing diversion and coordination effects depends
on the particular properties of demand.

4.

Data

4.1
Experimental Design
The WTP for local media environment features are estimated with data from an online
survey questionnaire employing repeated discrete-choice experiments. The questionnaire
begins with the cognitive buildup section that describes the respondent’s local media
environment in terms of the offerings from newspapers, radio, TV, the Internet, and
Smartphone. Respondents are asked questions about their media sources, how much
information they consume from each source, the cost of their media sources, and the quality of
the four different features of their media environment described in Table 1.19 The descriptions
of the “Media Environment Features” as they appear in the survey are provided in Appendix B.

18 Matsa (2010) shows empirically that supermarkets facing more intense competition have more product
availability, and Mazzeo (2003) shows that average flight delays are longer in more concentrated airline markets.
In contrast, Domberger an Sherr (1989) find no correlation between the threat of new entry and customer’s self-
reported satisfaction with their attorney used for house purchases.
19 Respondents were asked to consider what is available in their local media environment, rather than what they
usually view or listen to. This represents a statement about the amount and quality of information programming
being produced by media sources for their consumption.
28


Cognitive buildup is followed by the choice experiments. Information from the
cognitive buildup questions is used to summarize each respondent’s actual “status quo” (SQ)
media environment at home in terms of the media sources they use to get their information, the
levels of the features of their environment: DIVERSITY OF OPINION, COMMUNITY NEWS,
MULTICULTURALISM and ADVERTISING features, and their COST. A table summarizing
the sources and features of the respondent’s actual media environment at home is presented
before the choice task (see Figure 1 for an example). The respondent is then instructed to
answer the eight choice scenarios within the choice task. In each choice scenario, a pair of new
media environment options, A and B, is presented. The two options provide information on
news and current affairs from the same set of media sources indicated by the respondent during
cognitive buildup, but differ by the levels of the features. Respondents indicate their preference
for choice alternative A or B. A follow-up question is then presented that asks respondents to
make an additional choice between their preferred alternative, A or B, and their actual SQ
media environment at home. See Figure 2 for a choice scenario example.
We used market data from newspapers, radio and television stations, Internet and
mobile telephone service providers, a pilot study and three focus groups to test and refine our
descriptions of the features for choice alternatives A and B. The first focus group, with a hard-
copy version of the survey, was held on December 9, 2010, in a room of the Economics
building at the University of Colorado at Boulder. Four individuals, two men and two women,
a local service employee and three staff members of the Economics Department (undergraduate
advisors and clerical workers) took the survey under supervision of the principal investigator
and answered detailed questions regarding how they interpreted the questions and what they
were thinking when they answered them. The second focus group, with an online survey, was
29


facilitated by RRC Associates in Boulder on February 2, 2011. The group consisted of five
diverse individuals with respect to age, gender, and work Internet experience, who completed
the survey sequentially in the presence of a professional facilitator. A report describing the
findings from the second focus group is provided in Appendix C.
Measures developed by Huber and Zwerina (1996) and Zwerina et. al. (1996) were used
to generate an efficient non-linear optimal design for the levels of the features that comprise the
media environment choice. A fractional factorial design created 72 paired descriptions of
media environment, A and B, that were grouped into nine sets of eight choice questions. The
nine choice sets were rebalanced to ensure that each household faced a range of costs that
realistically portrayed the prices for media sources in their local media environment. For
example, a respondent who indicated that they pay nothing for their local media environment
was exposed to a range of costs that included zero dollars per month. Accordingly, COST1
ranged from $0 to $50 for households that indicated that the total cost of their actual media
environment at home was less than or equal to $30 per month. COST2 ranged from $5 to $100
for households that indicated that their total cost was greater than $30 but less than or equal to
$70 per month. COST3 ranged from $5 to $150 for households that indicated that their total
cost was greater than $70 but less than or equal to $120 per month. COST4 ranged from $10 to
$200 for households that indicated that their total cost was greater than $120 but less than or
equal to $180 per month. COST5 ranged from $10 to $250 for households that indicated that
their total cost was greater than $180 per month.20

20 The upper limit of $250 per month is the total cost for a media environment with a seven-day subscription to a
premium newspaper, such as the New York Times or San Francisco Chronicle ($25), a “All of XM” subscription
to satellite radio ($20), a premier subscription to cable or satellite television ($110), a subscription to very-fast
Internet service ($45), an unlimited data subscription for a Smartphone ($30), and $10 monthly memberships to
both NPR and PBS.
30


The nine choice sets were randomly distributed across all respondents. Upon
completion of their cognitive buildup questions, an online algorithm calculated each
individual’s total cost of their local media environment and assigned the appropriate cost range
for the choices experiments, either COST1, COST2, COST3, COST4, or COST5. To account for
the possibility of order effects that could confound the analysis, the order of the eight A-B
choices questions within each of the nine choice sets were also randomly assigned across all
respondents.
The methodology used to estimate consumer utility has several important
characteristics. First, the experimental approach determines the levels of the features of each
media environment offered exogenously and avoids collinearity problems by offering non-
existing alternatives. For example, the levels for the diversity of opinion, community news
and/or multiculturalism features change independently in the hypothetical alternatives as
opposed to market data where they are often positively correlated.21 By asking respondents to
complete eight choice experiments, we are able to increase parameter estimation precision, and
reduce sampling costs by obtaining more information on preferences for each respondent.
Second, the choice data are used to estimate parameters of the representative household’s utility
function. This has the advantage that from estimates of these structural parameters, we can
construct estimates of the value of any variant of the current and future media environment
described by the levels of the five features.
Because some of our data are from choice experiments, we need to be concerned with
hypothetical bias and survey fatigue. Hypothetical bias arises when the behavior of the
respondent is different when making choices in a hypothetical market versus a real market. For

21 By holding all media sources and unobserved features of the media environment constant within each individual
household’s choice task, the discrete-choice model also eliminates any potential correlation between prices and the
unobserved quality of the media environment options.
31


example, if the respondent does not fully consider her budget constraint when making choices
between hypothetical options A and B, WTP may be overestimated, because the cost parameter
in the denominator of the WTP calculation (see Section 3.2) will be biased toward zero and the
marginal utility parameter in the numerator will be biased away from zero. We believe this
bias is less of a concern in this study as opposed to studies that ask consumers to value
environmental goods or advanced telecommunications services that are not provided in
markets. Because most consumers have paid for many different media sources in actual
markets, they should have a reasonable understanding of their preferences for their local media
environment, and how their choices are constrained by their budget and time. Nevertheless,
recent papers by Cummings and Taylor (1999), List (2001), Blumenschein et. al. (2008) and
Savage and Waldman (2008) have proposed methods for minimizing hypothetical this source
of bias. In this paper, we follow Savage and Waldman by employing a follow-up question that
asks respondents to make an additional choice between their new choice, A or B, and their
actual media environment at home. This additional non-hypothetical market information is
then incorporated into the likelihood function that is used to estimate marginal utility
parameters.
Survey fatigue can arise from a lengthy questionnaire and make estimates from later
scenarios differ from earlier scenarios. Carson et. al. (1994) review a range of choice
experiments and find that respondents are typically asked to evaluate eight choice scenarios.
Savage and Waldman (2008) found there is some fatigue in answering eight choice scenarios
when comparing online to mail respondents. To minimize survey fatigue in this study, we have
reduced the cognitive burden by dividing the choice task into two sub groups of four choice
32


scenarios. Here, the respondent is given a break from the overall choice task with an open-
ended valuation question between the first and second set of four scenarios.22

4.2
Survey Administration
Knowledge Networks Inc. (KN) administered the household survey online. KN panel
members are recruited through national random samples, almost entirely by postal mail. For
incentive, panel members are rewarded with points for participating in surveys, which can be
converted to cash or other rewards.23 An advantage of using KN is that it obtains high
completion rates and the majority of the sample data are collected in less than ten days. KN
also provides detailed demographic data for each respondent. Because these demographics are
previously recorded, the length of the field survey is shortened to less than 20 minutes, which
ensures higher quality responses from the respondents.
During the week of March 7, 2011, KN randomly contacted a gross sample of 8,621
panel members by email to inform them about the media environment survey.24 The survey
was fielded from March 11 to March 21. A total of 5,548 respondents from all 50 states and
the District of Columbia completed survey questionnaires for a response rate of 64.4 percent.
We trimmed the net sample by eliminating: 341 respondents with a completion time of less
than six and one-half minutes; 46 respondents who skipped any questions in the choice task; 14
respondents who indicated that they pay $500 or more per month for the media sources within

22 For a robustness check, we compared our baseline estimates of utility in Table 10 below for the AB-SQ
bivariate probit model with estimates on the data for the hypothetical A-B choices only, as well as with estimates
on the data for the first four choice questions only, and get similar results.
23 Unlike convenience panels that only include volunteers with Internet access, KN panel recruitment uses dual
sampling frames that includes both listed and unlisted telephone numbers, telephone and non-telephone
households, and cellphone-only households, as well as households with and without Internet access. If required,
households are provided with a laptop computer and free Internet access to complete surveys, but they do not
participate in the incentive program.
24 The invitation email has the subject heading and approximate time for completion, estimated from KN’s online
pretest on 39 respondents.
33


their local media environment; 11 respondents who provided incomplete cost information; and
five respondents who provided incomplete information on the features of their media
environment.25 The median completion time for our final sample of 5,131 respondents with
complete information was about 16 and three-quarter minutes. The panel tenure in months for
final sample respondents ranged from one to 136, with a mean of 41.18 and standard deviation
of 31.33.26
Table 2 presents a selection of demographics for the U.S. population, for all KN’s panel
members, and for panel members who were invited to participate in this survey (Knowledge
Networks, Inc., 2010; United States Census Bureau, 2009). The demographics for all KN panel
members are similar to those reported by the Census Bureau. Casual inspection of column four
and column five of Table 2 also show that, apart from race and employment status, the
demographics for the gross sample of panel members invited to participate in this study and the
final sample of respondents who completed questionnaires are also similar to those reported by
the Census Bureau. However, estimates from the probit model that compares respondent’s
characteristics between the gross sample and the final sample also indicate potential differences
in age, gender, education and Internet access between our final sample and the population. We
remedy this possible source of bias in our results in Section 5 by estimating with weighted
maximum likelihood, where the contribution to the log likelihood is the post-stratification
weight times the log of the bivariate probability for the individual choice occasion.27


25 Our pilot study, focus groups and other testing indicated that the minimum time needed to complete the survey
was about six or seven minutes. Because they may be shirking, we removed the 341 respondents in our survey
with a completion time of less than six and one-half minutes. Evidence from KN suggests that this behavior is not
specific to our survey style or content. Our sample’s distribution of interview duration in minutes is very similar to
other KN surveys with median completion times ranging from seven to 19 minutes.
26 See Dennis (2009) for a description of the within-panel survey sampling methodology.
27 See Appendix D for the probit model estimates and the procedures used to develop the post-stratification
weights used for weighted maximum likelihood estimation.
34


4.3
Media Environment at Home
Table 3 presents summary statistics for respondent’s media sources. Column two and
column three show that about 94 percent of sample respondents watch television, about 81
percent listen to the radio, and about 80 percent use the Internet. About 45 percent of
respondents read a paper or online newspaper regularly, and about 24 percent of sample
respondents own a Smartphone. On average, television viewers spend about 1.9 hours on a
typical day watching television to get information on news and current affairs, radio listeners
spend about 1.4 hours listening to the radio to get information on news and current affairs, and
Internet users spend about one hour online (e.g., MSN, Yahoo, radio and TV station web sites,
journalists’ blogs) to get information on news and current affairs. Newspaper readers also
spend about one hour on a typical day reading the newspaper, while Smartphone owners use
their phone to go online for about 0.6 hours to get information on news and current affairs
online. Table 4 shows that the most popular media source combinations are radio, television
and the Internet, about 30 percent of sample respondents, and newspaper, radio, television and
the Internet, about 26 percent of sample respondents.

Summary statistics for media environment features are presented in Table 5. These data
indicate that, on average, the levels of the DIVERISTY OF OPINION, COMMUNITY NEWS,
MULTICULTURALISM and ADVERTISING features were about “medium.” About 58 percent
of respondents indicated that they bundled their subscription television service with the Internet
and/or telephone service. The price (or, COST) for the typical media environment ranged from
zero to $447 per month, with an average of $111.20 per month. Interestingly, about ten percent
of the sample indicated that they have contributed to public radio stations and/or public TV
stations during the past twelve months at an average of $9.30 per month.
35



4.4
Satisfaction With Media Environment Features

Additional information on household satisfaction with their media environment was
obtained by asking respondents to indicate on a scale from one to five, with one indicating “not
satisfied” and five indicating “very satisfied”, how satisfied they were with each feature, and
their overall media environment. Table 6 and Table 7 show that the distribution of responses
are similar for DIVERSITY OF OPINION, COMMUNITY NEWS, and MULTICULTURALISM,
with almost 90 percent of households indicating a satisfaction score of at least three for each of
these features. Households are less satisfied with COST and ADVERTISING. About 60 percent
of households indicated a satisfaction score of at least three for each of these two features.
Table 8 presents summary statistics from FCC (2011) data that show considerable
variation in market structure between the 203 television markets in our sample.28 At
December, 2009, the total number of newspaper, radio, and television outlets (MEDIA
OUTLETS) ranged from four to 291, with an average of 139 per television market. On average,
about 81 percent of media outlets are radio stations, which partially reflects the fact that the
geographical definition of a television markets can include several radio markets. Moreover,
when examining the market structure data at the 75th percentile, we observe that most markets
are served by about 182 media outlets or less. The bottom panel in Table 8 shows a similar
pattern for small television markets with five or less television stations. At December, 2009,
the total number of newspaper, radio and television outlets in small markets ranged from four
to 86, with an average of 47 per market. On average, about 82 percent of media outlets in small

28 There are 210 television markets in the U.S. The seven markets not covered by our sample are: Bend, OR;
Fairbanks, AK; Grand Junction, CO; Missoula, MT; North Platte, NE; Ottumwa, MO; and Presque, ME. All seven
markets are small markets with five or less television stations. As shown in Table 10, the remaining small markets
in our sample cover 8.43 percent of households. FCC (2011) data show that 8.37 percent of population households
were in small markets at December, 2009.
36


markets are radio stations, and as indicated by the 75th percentile, most small markets are
served by about 57 media outlets or less.29
Evidence of the relationship between consumer satisfaction with their local media
environment and television market structure is presented in Table 9. We use the ordered probit
model to estimate the effect on consumer satisfaction, as measured by DIVERSITY OF
OPINION, COMMUNITY NEWS, MULTICULTURALISM, ADVERTISING or OVERALL
MEDIA ENVIRONMENT (i.e., how satisfied are you with your overall media environment),
from TV STATIONS, TV VOICES, and TV STATIONS TV VOICES, the vector of demographic
controls, X = [AGE, EDUC, GENDER, INCOME, RACE], and dummy variables that control for
the 15 media source combinations. The values for each of the dependent variables range from
one to five with one indicating “not satisfied” and five indicating “very satisfied.” The
estimated coefficients on the demographic variables indicate that the elderly are more satisfied
with the diversity of opinion, community news and multiculturalism features of their media
environment, but less satisfied with advertising. Similarly, educated respondents are more
satisfied with their diversity of opinion and multiculturalism features, but less satisfied with
advertising. White respondents are likely to be more satisfied with their diversity of opinion
and multiculturalism features, but are less satisfied community news and advertising. Only two
of the estimated coefficients on the individual market structure variables are significant at
conventional levels in all five specifications. For example, the positive coefficient on TV
VOICES in column three indicates that, all other things held constant, an increase in the number
of independent television voices in the market is associated with a increase in consumer’s
satisfaction with the multiculturalism feature of their local media environment.

29 See Appendix E for a description of the variables in Table 8.
37


There are several possible explanations for the insignificant coefficients on most of the
market structure variables in Table 9: there is no relationship between consumer satisfaction
with the media environment and television market structure; the five-point index is not a good
measure of consumer satisfaction; and/or the television variables may not provide an accurate
representation of market structure. In Section 5, we quantify consumer satisfaction in dollars
by estimating the marginal willingness-to-pay for DIVERSITY OF OPINION, MULTICUL-
TURALISM, COMMUNITY NEWS and ADVERTISING. In Section 6, we use the WTP
estimates to calculate how expected consumer welfare from the local media environment varies
between different market structures.

5.

Consumer Valuations

The choice data described in Section 4.1 are used to estimate a bivariate probit model of
household utility from their local media environment. Since each pair of binary choices, A vs.
B, and A or B vs. SQ, for each choice occasion represents information on preferences, the
starting maximum sample size for econometric estimation is n = 5,031 x 8 = 40,248. Table 2
showed some demographic differences our final sample and the population. We remedy this
possible source of bias in our results by estimating with weighted maximum likelihood, where
the contribution to the log likelihood is the post-stratification weight times the log of the
bivariate probability for the individual choice occasion.

5.1
Baseline Results and Robustness
Table 10 reports weighted maximum likelihood estimates of the baseline model of
household utility. Because consumers may have heterogeneous preferences for unmeasured
38


aspects of media environment alternatives, we estimate utility with a constant to capture
differences in tastes between the status quo and new A and B media options.30 Marginal utility
parameters (MU), asymptotic t-statistics for the marginal utilities (t), WTP calculations (WTP)
and standard errors for the WTP calculations (s.e.) are presented in columns two through five.
The estimate of the ratio of the standard deviation of the errors in evaluating the hypothetical
alternatives to the errors in the status quo alternative, λ = 1.49, is greater than one.
Respondents appear to have more consistency in choice when comparing the new media
environment options than when comparing a new option to a real alternative.31
The data fit the baseline model well as judged by the statistical significance of most
parameter estimates. The marginal utility parameters for DIVERSITY OF OPINION,
COMMUNITY NEWS, and MULTICULTURALISM are positive and are significant at the one
percent level. The marginal utility parameters for COST and ADVERTISING are negative and
statistically significant at the one percent level. The estimated signs for these media features
imply that the representative consumer’s relative utility increases when: the information on
news and current affairs from different viewpoints is increased; the amount of information on
community news and events is increased; the amount of information on news and current
affairs reflecting the interests of women and minorities is increased; the amount of space and/or

30 Holding all other features of the media environment constant, the difference in utility between the status quo and
the new media environment option can be interpreted as the consumer’s disutility from switching from the status
quo to the new media environment. Dividing this difference by the marginal disutility of COST provides an
estimate of the average consumer switching cost, here, about $26 (= 0.319/0.012) per month. Another way of
examining switching costs is by comparing them to respondent’s annualized average monthly cost of their media
environment, here $1,334 (= 111.2  12). The estimated switching cost is about 23 percent of annual consumer
expenditures on the media sources that comprise their media environment. For comparison, Shcherbakov (2007)
estimated that switching costs comprise about 32 and 52 percent of annual expenditures on cable and satellite
television services, respectively.
31 The parameter λ is generally estimated to be greater than one in most models in Table 10 through Table 18. We
report its estimate and the corresponding test statistic, but do not discuss it further.
39


time devoted to advertising is decreased; and the dollar amount the household pays per month
for their media environment is decreased.
DIVERSITY OF OPINION and COMMUNITY NEWS are important features of the local
media environment with consumers willing to pay $13.06 per month for different viewpoints in
the reporting of information on news and current affairs, and $13.95 for more information on
community news and events. Consumers also value MULTICULTURALISM, although the
willingness to pay for this feature is less precisely estimated. The results show that consumers
would be willing to pay an additional $1.82 per month for more information that reflects the
interests of women and minorities. As expected, consumers have a distaste for ADVERTISING.
The representative consumer would be willing to pay $8.18 per month for a marginal decrease
in the amount of advertising they have to listen to or view.
Because our data are from choice experiments, we check the sensitivity of our results
with respect to hypothetical bias and survey fatigue.32 Our first robustness check examines the
potential for hypothetical bias by observing how the omission of consumer choice data from
actual markets affected our results. Here, we compared baseline estimates of the bivariate
probit model on data for the A versus B and A or B versus SQ choices with estimates from the
univariate probit model on the data for the A-B choices only. The results, reported in column
six through column nine of Table 11, show that the two models produce reasonably similar
estimates of willingness to pay. Households have relatively similar valuations for DIVERSITY

32 The distributions of answers to the choice scenarios show that in 53 percent of the A-B choice questions,
respondents chose new media environment option A over B. Because the order of the choice scenarios is
randomized, this has no effect our results. In the follow up questions, respondents chose to stay with their actual
(status quo) media environment over the new option, A or B, in about 71 percent of the choice questions. About
12.7 percent of the choice occasions respondents chose to change to hypothetical media environment A, and about
16.4 percent of the time they chose to switch to hypothetical environment B. The number of seconds it took
respondents to answer the eight choice questions remained essentially constant over the eight choice occasions.
40


OF OPINION and COMMUNITY NEWS in the univariate probit model, while the valuation for
MULTICULTURALISM is lower.
Our second robustness check looks for any major differences in household valuations of
their media environment features as they progress through the choice task, perhaps reflecting
survey fatigue. Here, we compared the baseline estimates of the bivariate probit model on the
data for the first four choice questions versus the second four questions. The results, reported
in column six through column nine of Table 12, show reasonably similar estimates for the two
subsamples of data. There is no systematic pattern that could be taken as evidence of survey
fatigue.

5.2
Heterogeneous Preferences
Because they do not have identical preferences, it is possible that individual consumer’s
WTP for their media environment varies with observable demographics such as age, education,
gender, income, and race. For example, women and non-white households may have stronger
preferences for MULTICULTURALISM, and, because of a higher opportunity costs of time,
higher income households may have a stronger distaste for ADVERTISING.33
Table 13 reports estimates of the random utility model for subsamples of respondents
aged from 18 to 29 years, 30 to 44 years, 45 to 59 years, and respondents aged 60 years and
over. Willingness-to-pay for more information on community news and events increases with
age, from $8.96 to $20.78 per month. Willingness-to-pay for more information that reflects the
interests of women and minorities decreases with age, with the 60 years and over group placing
no value on this particular feature. Younger consumers have less distaste for advertising.

33 The likelihood ratio test statistics for Table 13 through Table 18, not reported, are large and reject the hypothesis
that the estimated marginal utilities are equal across different subsamples.
41


Respondents aged 18 to 44 years are willing to pay about $5 or $6 per month for a decrease in
the amount of advertising in their media environment, whereas respondents 45 years and over
are willing to pay about $9 to $12 per month.
The possibility that preferences vary with education is examined in Table 14 which
reports estimates for a subsamples of respondents with less than high school education, with a
high school education, some college, and with a bachelors degree or higher. Willingness-to-
pay for diversity of opinion, information on community news and events, and information that
reflects the interests of women and minorities increases with years of education. Respondents
with no college experience do not value information that reflects the interests of women and
minorities. Moreover, they are willing to pay about $4 or $6 per month for a decrease in the
amount of advertising in their media environment compared to educated respondents who are
willing to pay about $9 to $10 per month.
Table 15 reports estimates for a subsample of low income respondents (i.e., annual
household income less than $25,000), a subsample of middle income respondents (i.e., annual
household income more than $25,000 but less than $75,000) and a subsample of high income
respondents (i.e., annual household income greater than $75,000). Valuations for diversity of
opinion, more information on community news and events, and (less) advertising all increase
with income. Low-income respondents do not value information on news and current affairs
that reflects the interests of women and minorities, however, middle- and high-income
respondents are willing to pay about $1.50 to $2.50 per month for more information that
reflects the interests of women and minorities.
Estimates of utility for subsamples of male and female respondents are reported in
Table 16. The estimated willingness-to-pay for diversity of opinion, information on
42


community on news and events and less advertising are very similar across these groups.
However, while females are willing to pay about $3 per month for information on news and
current affairs that reflects the interests of women and minorities, males place no value on this
type of information from their local media environment. Estimates of utility and WTP for
subsamples of white and non-white respondents are reported in Table 17. White consumers are
willingness-to-pay more for diversity of opinion, information on community news and events,
and less advertising that non-white households. White consumers do not value information on
news and current affairs that reflects the interests of women and minorities. In contrast, non-
white consumers are willing to pay about $5 per month for more information that reflects the
interests of women and minorities. This relationship is explored further by estimating the
random utility model on subsamples of white versus non-white males and white versus non-
white females. The results, reported in Table 18, are similar in flavor to those reported for the
male and female subsamples. Non-white males are willing to pay $3.48 per month for more
information that reflects the interests of women and minorities, white females are willing to pay
$1.52 per month, and non-white females are willing to pay $6.16 per month.

5.3
Non-Linear Preferences
Up to this point the coding of the four non-price features in the household utility
function have been linear, which implies that the marginal utilities are the same when moving
from low to medium and from medium to high. We now relax this restriction by replacing each
of the four features with a pair of dichotomous variables. For example, MEDIUM DIVERISTY
OF OPINION equals one when DIVERISTY OF OPINION equals “medium” and zero
otherwise, and HIGH DIVERISTY OF OPINION equals one when DIVERISTY OF OPINION
43


equals “high” and zero otherwise. Here, the estimated parameter on MEDIUM DIVERISTY OF
OPINION measures the change in utility from moving from information on news and current
affairs in the household’s overall media environment reflecting only one viewpoint to a few
different viewpoints. The estimated parameter on HIGH DIVERISTY OF OPINION measures
the change in utility from moving from information on news and current affairs reflecting only
one viewpoint to many different viewpoints. This approach to estimating non-linear consumer
valuations is used for all of the other non-price features of the local media environment.
Estimates of the utility model with non-linear preferences are presented in Table 19.
We first note that the these estimates are very consistent with the marginal WTP estimates
calculated for linear preferences in Table 10. For example, Table 19 shows that the
representative consumer is willing to pay $25.28 per month for an improvement in diversity of
opinion from low to high medium, while Table 10 shows that the consumer would be willing to
pay $26.12 (= 2  $13.06) for the same improvement. Similar, consistent findings are also
found for COMMUNITY NEWS, MULTICULTURALISM and ADVERTISING.
The estimates from Table 19 suggest declining marginal utility with respect to diversity
of opinion, community news and multiculturalism. The representative household is willing to
pay $18.86 per month for an improvement in diversity of opinion from low to medium, but
only another $6.42 per month for an additional improvement to high diversity of opinion.
Similarly, the representative household is willing to pay $20.54 per month for an initial
improvement in information on community news and events from low to medium, but only
another $6.39 per month for an additional improvement to high. The marginal utility estimates
for MULTICULTURALISM indicate that households value an improvement in information that
reflects the interests of women and minorities from low to medium (i.e., WTP = $4.03) about
44


the same as an improvement from low to high (i.e., WTP = $3.96).34 In other words, a lot of
information reflecting the interests of women and minorities is valued the same as some
information. The marginal utility estimates for ADVERTISING indicate a similar pattern, albeit
in reverse, with respect to the amount of space and/or time devoted to advertising in the
household’s media environment. The representative household is willing to pay about $17.37
per month for a move from high to medium advertising, but would pay only an additional $2.85
per month to move from medium to low advertising.

6.

Change in Consumer Welfare From a Television Merger

The policy question of interest is how do FCC ownership rules and merger guidelines affect the
diversity and localism features of the media environment in a way that satisfies consumers We
shed light on this question in Step two of our empirical analysis by estimating equation 4,
which measures the effects of a change in TV VOICES on respondent i’s perceived quality of
her local media environment features in market j. We then use our demand estimates from Step
one to calculate the effects on consumer welfare form a one-unit decrease in TV VOICES that
results in quality differences between the pre- and post-merger markets.

6.1
Ordered Probit Model Estimates of Equation 4
Table 20 reports the estimated coefficients and standard errors from weighted maximum
likelihood estimation of equation 4. We use the ordered probit model to estimate the effect on
product quality, as measured by DIVERSITY OF OPINION, COMMUNITY NEWS,
MULTICULTURALISM, or ADVERTISING from TV STATIONS, TV VOICES, and

34 In statistical terms, the two WTP estimates for MEDIUM MULTICULTURALISM and
HIGH MULTICULTURALISM are not significantly different from one another.
45


TV STATIONSTV VOICES, the vector of demographic controls, X = [AGE, EDUC, GENDER,
INCOME, RACE], and dummy variables that control for the 15 media source combinations.
The values for each of the dependent variables range from one to three with one indicating
“low”, two indicating “medium” and three indicating “high.” The estimated coefficients on the
demographic variables indicate that the elderly, and educated respondents, are likely to
perceive higher quality in the diversity of opinion, community news and multiculturalism
features of their media environment, but lower quality with respect to advertising (i.e., too
much advertising). In contrast to Table 9, where the 5-point index of consumer satisfaction is
the dependent variable, the television market structure variables are reasonably good predictors
of product quality. The estimated coefficients on TV VOICE are positive and significant for
DIVERSITY OF OPINION, MULTICULTURALISM and ADVERTISING. This suggests that
following a decrease in the number independent television voices in the market, consumers are
more likely to be in a state with less diversity of opinion, less multiculturalism, and less
advertising. Moreover, when evaluated at the mean value for TV STATIONS of 12.74, linear
estimates of ∂β΄x/∂TV VOICES for DIVERSITY OF OPINION, MULTICULTURALISM and
ADVERTISING are significant at the five percent level.35 Because ∂β΄x/∂TV VOICES for
COMMUNITY NEWS is not precisely estimated, we exclude this feature from the merger
analysis below.



35 ∂β΄x/∂TV VOICES is the effect of TV VOICES on the index function in equation 4, β΄x, where x = [1, Xj, Zi], 1 is
a vector of ones, and β = [α, δ, γ]. We also evaluated ∂β΄x/∂TV VOICES over the range of values for
TV STATIONS from five to 20 for each of the four features. The effects are significant at the five percent level for
all values of TV STATIONS for DIVERSITY OF OPINION, MULTICULTURALISM and ADVERTISING. The
effects are not significant for COMMUNITY NEWS.
46


6.2
Merger Algorithm
The procedure to calculate the changes to consumer welfare from a merger between two
television stations is:
(i)
Estimate the marginal WTP per month for a move from a low level of the feature to
the medium level, and from the low level to the high level. These estimates are
from the non-linear specification of preferences reported in Table 19, and they are
replicated directly below for easier exposition. For example, row two-column two
below shows that the representative consumer is willing to pay $ 18.86 for a move
from low to medium diversity of opinion in her local media environment.
Feature
Low to medium (ΔbM)
Low to High (ΔbH)
DIVERSITY OF OPINION
$18.86
$25.28
MULTICULTURALISM
$20.54
$26.93
ADVERTISING
-$2.85
-$20.22

(ii)
Estimate the ordered probit model of equation 4 and obtain coefficient estimates of the
effect of TV VOICES (and other independent variables) on the quality of a
respondent’s media environment features. These coefficients are reported in Table
20.
(iii)
Given the existing (pre-merger) sample data, use the estimated coefficients from (ii)
above to predict each respondent’s pre-merger probability distribution of low,
medium and high values for each of the three media environment features, where:
PL0 is the pre-merger probability of a low level of the feature;
PM0 is the pre-merger probability of a medium level of the feature; and
PH0 is the pre-merger probability of a high level of the feature.
47


(iv)
Simulate the merger between two television stations by reducing the number of
independent television voices in the sample by one, all other things held constant.
Repeat (iii) above to predict each respondent’s post-merger probability distribution
of low, medium and high values for each of the three media environment features,
where:
PL1 is the post-merger probability of a low level of the feature;
PM1 is the post-merger probability of a medium level of the feature; and
PH1 is the post-merger probability of a high level of the feature.


(v)
Use the probabilities in (iii) and (iv) to form, for each respondent,
,
and
,


for each of the three local media environment features, where ΔPL = PL1 - PL0,

Δ

P
, Δ
M = PM1 - PM0
PH = PH1 - PH0, and ΔX1 = 1. For example, let
= 0.0089,



= 0.0074 and
= -0.0163 for the advertising feature. The interpretation is


that if the number of independent television voices is reduced by one, the
probability that respondent i is in a low advertising environment will increase by
0.0089, the probability that respondent i is in a medium advertising environment
will increase by 0.0074 and the probability that respondent i is in a high advertising
environment will decrease by 0.0163.36
(vi)
Use the estimates of marginal WTP in (i) and the change in predicted probabilities in
(v) to evaluate equation 9 for each respondent. This is the effect of a one-unit change

36 The sample means of the predicted probabilities are presented in Table 21. For our example of advertising, they
have the following interpretation. Following the reduction in the number of independent television voices by one,
the percentage of households in a low advertising environment will increase by 0.0089, the percentage of
households in a medium advertising environment will increase by 0.0074, and the percentage of households in a
high advertising environment will decrease by 0.0163.
48


in the number of independent television voices (X ) on the respondent’
1
s expected
consumer welfare from diversity of opinion, multiculturalism or advertising (a). To
continue the advertising example, equation 9 is:


= Δb
Δb




+





(9΄)



and the expected change in welfare for consumer i from a one-unit reduction in the
number of independent television voices is:




= (-$2.85  0.0074) + (-$20.22  -0.0163) = $0.31, which represents a gain

to consumer i of about 31 cents per month from less advertising in her media
environment.
(vii)
Sort the expected welfare changes in (vi) for each respondent by the number of
television stations, ranging from five to 20. Calculate the mean expected welfare
change per month for all respondents in a market with five stations, for all respondents
in a market with six stations, for all respondents in a market with seven stations, … ,
and all respondents in a market with 20 stations.
(viii) Use the FCC (2011) data to count the number of population households in a television
market with five stations, six stations, seven stations, … , and 20 stations.
(ix)
Calculate the aggregate annual change in consumer welfare from the merger between
two television stations by multiplying (vii) by 12 by (viii) for each level of the number
of television stations, i.e., five stations, six stations, seven stations, etc.


49


6.3
Merger Results
Table 22 presents the results from our hypothetical merger between two television stations.
The results show that, all other things held constant, diversity of opinion, the coverage of
multicultural issues, and the amount of advertising decrease following the merger. These
changes in quality in local media environments lead to lower total consumer welfare, but these
losses decrease with the number of television stations in the market.
Figure 3 plots average consumer welfare per month for all television markets with five
stations to 20 stations. The first interesting observation is that the welfare changes from a
television station merger are dependent on the number of television stations that existed before the
merger. Here, the negative effects from a one-unit decrease in the number of independent
television voices on consumer’s perception of quality in their local media environments are more
pronounced in markets with relatively fewer television stations. As such, the average consumer in
a small market, i.e., five television stations, loses $0.11 per month (see also column six of Table
22), whereas the average consumer in a large market, i.e., 20 stations, loses $0.04 per month. As
shown in the right-hand panel of Table 22, these losses are equivalent to $6.0 million annually
for all small-market households in the U.S. and $1.4 million annually for all large-market
households.37 By summing the individual values in column 10 of Table 22, we that if the
merger occurs in all markets, aggregate consumer welfare losses would be about $91 million.
The other interesting observation from the merger analysis is the potential tradeoff
between the quality of diversity features and the amount of advertising in local media
environments. Both Table 22 and Figure 3 show that consumers lose from the merger because
there is less diversity of opinion and less coverage of multiculturalism issues, but they gain

37 For small markets, $5,975,856 = 4,469,100 12 -0.1114. For large markets, $2,348,092 = 1,391,092 12
-0.0396.
50


from lower advertising. More specifically, columns seven, eight and nine of Table 22 show
that more than 75 percent of the annual consumer losses from less diversity of opinion and
multiculturalism in each television market are offset by less exposure to advertising in local
media environments.
Finally, we examined how sensitive our results are to an alternative specification of
media market structure that controls for the number of daily newspapers in the market
(NEWSPAPERS), the number of radio stations in the market (RADIO STATIONS), the number
of households in the market (HOUSEHOLDS), and the geographical size of the market in
square miles (AREA). Specifically, we use the ordered probit model to estimate the effect on
product quality, as measured by DIVERSITY OF OPINION, COMMUNITY NEWS,
MULTICULTURALISM, or ADVERTISING from TV STATIONS, TV VOICES, TV STATIONS
TV VOICES, NEWSPAPERS, RADIO STATIONS, HOUSEHOLDS and AREA, the vector of
demographic controls, X = [AGE, EDUC, GENDER, INCOME, RACE], and dummy variables
that control for the 15 media source combinations. The results from the ordered probit model
and the merger analysis, not reported here, have a qualitatively similar pattern to those
presented in Table 20 and Table 22.

7.

Conclusions

This study examined the effects of media market structure on consumer demand and welfare. A
differentiated-product model was used to estimate demand for the local media environment,
described by the offerings from newspapers, radio, television, the Internet, and Smartphone.
Results show that the representative household is willing to pay $13 per month for different
viewpoints in the reporting of information on news and current affairs, and $14 per month for
51


more information on community news and events. Consumers value more information that
reflects the interests of women and minorities, although willingness-to-pay for this is only
about $2 per month. Consumers have a distaste for advertising and are willing to pay $8 per
month for a decrease in the amount of advertising in their media environment.
We used our demand estimates to calculate the changes to consumer welfare from a
merger between two television stations that resulted in quality differences in diversity and
advertising in the pre- and post-merger markets. We conducted a simple experiment that
simulates the merger by reducing the number of independent television voices in the market by
one. The two interesting empirical observations from our analysis are that the consumer welfare
effects from a television station merger depend on the number of pre-existing television stations in
the market, and there are quality tradeoffs from the changes to product offerings by media outlets
in the post-merger markets.
All other things held constant, diversity of opinion, the coverage of multicultural issues,
and the amount of advertising decrease following the merger. These changes in quality in local
media environments lead to lower consumer welfare, which decreases with the number of
television stations in the market. For example, the average consumer in a small market loses $0.11
per month, whereas the average consumer in a large market loses $0.04 per month. These losses
are equivalent to $6 million annually for all small-market households in the U.S. and $1.4
million annually for all large-market households. If the merger occurred in all markets,
aggregate consumer welfare losses would be about $91 million.


52


References

Alexander, P., and Brown, K. 2004. “Do Local Owners Deliver More Localism? Some
Evidence from Local Broadcast News,” working paper, available at
http://www.fcc.gov/ownership/materials/already-released/doownersdeliver070004.pdf">http://www.fcc.gov/ownership/materials/already-released/doownersdeliver070004.pdf.
Baum, M., and Kernell, S. 1999. “Has Cable Ended the Golden Age of Presidential
Television,” American Political Science Review, 93(1), p. 99-113.
Berry, S., and Waldfogel, J. 2001. “Do Mergers Increase Product Variety? Evidence from
Radio Broadcasting,” Quarterly Journal of Economics, 116(3), 1009-1025.
Blumenschein, K., Blomquist, G., and Johannesson, M.. 2008. “Eliciting Willingness to Pay
without Bias using Follow-up Certainty Statements: Comparisons between
Probably/Definitely and a 10-point Certainty Scale.” Economic Journal, 118, 114-137.
Byerly, C., Langmia, K., and Cupid, J. 2006. “Media Ownership Matters: Localism, the Ethnic
Minority News Audience and Community Participation,” In Does Bigger Media Equal Better
Media? Four Academic Studies of Media Ownership in the United States, Benton Foundation
and Social Science Research Council, http://www.ssrc.org/programs/media.
Carson, R., Mitchell, R., Haneman, W., Kopp, R., Presser, S., and Ruud, P. 1994. Contingent
Valuation and Lost Passive Use: Damages from the Exxon Valdez. Resources for the future
discussion paper, Washington, D.C.
Chambers, T. 2003. “Radio Programming Diversity in the Era of Consolidation,” Journal of
Radio Studies, 10(1), 33-45.
Chipty, T. 2007. “Station Ownership and Programming in Radio,” FCC Media Ownership
Study #5, Washington, D.C.
53


Chipty, T., 1995. “Horizontal Integration for Bargaining Power: Evidence from the Cable
Television Industry,” Journal of Economics and Management Strategy 4 (1995), 375-397.
ComScore, 2011. “ComScore Reports August 2010 U.S. Mobile Subscriber Market Share,”
accessed on March 31, 2011 at
http://www.comscore.com/Press_Events/Press_Releases/2010/10/comScore_Reports_August_2010_U.S._Mobile_Subscriber_Market_Share/(language)/eng-US">http://www.comscore.com/Press_Events/Press_Releases/2010/10/comScore_Reports_August
http://www.comscore.com/Press_Events/Press_Releases/2010/10/comScore_Reports_August_2010_U.S._Mobile_Subscriber_Market_Share/(language)/eng-US">_2010_U.S._Mobile_Subscriber_Market_Share/(language)/eng-US
Crawford, G. 2007. “Television Station Ownership Structure and the Quantity and Quality of
TV Programming,” FCC Media Ownership Study #3, Washington, D.C., FCC.
Crawford, G. 2000. “The Impact of the 1992 Cable Act on Household Demand and Welfare,”
RAND Journal of Economics, 31, 422-449
Cummings, R., and Taylor, L. 1999. “Unbiased Value Estimates for Environmental Goods: A
Cheap Talk Design for the Contingent Valuation Method.” American Economic Review, 89,
649-665.
de Bettignies, J., and Ross, T. 2011. Mergers, Managerial Efficiency, and Social Welfare,”
working paper, University of British Columbia.
Dennis, M., 2009. “Description of Within-Panel Survey Sampling Methodology: The
Knowledge Networks Approach.” Government and Academic Research, Knowledge
Networks.
Domberger, S., and Sherr, A. 1989. “The Impact of Competition on Pricing and Quality of
Legal Services,” International Review of Law and Economics, 9, 41-56.
Dubin, J., and Spitzer, M. 1993. “Testing Minority Preferences in Broadcasting,” working
paper 856, California Institute of Technology.
54


Economides, N., Seim, K., and Viard, B. 2008. “Quantifying the Benefits of Entry into Local
Phone Service,” RAND Journal of Economics, 39(3), 699-730.
Farrell, J. 1983. “Monopoly Slack and Competitive Rigor: A Simple Model,” in Readings in
Games and Information, Rasmusen, E., ed., 169-172, Blackwell Publishing.
FCC. 2011. Government Furnished Information on Media Market Structure, February 17,
Media Bureau, FCC, Washington, D.C.
FCC. 2001. Review of the Radio Industry, Mass Media Bureau Policy and Rules Division,
FCC, Washington, D.C.
Gentzkow, M., and Shapiro, J. 2010. “What Drives Media Slant? Evidence from U.S. Daily
Newspapers,” Econometrica, 78(1), 35-71.
Gentzkow, M. 2007. “Valuing New Goods in a Model with Complementary Online
Newspapers,” American Economic Review, 97(3), 713-744.
Huber J, Zwerina K. 1996. “The Importance of Utility Balance in Efficient Choice Designs.
Journal of Marketing Research, 33, 307–317.
Knowledge Networks Inc. 2010. KnowledgePanel®Demographic Profile January 2010.
Government and Academic Research, Knowledge Networks.
Kling, C., and Sexton, R. 1990. “Bootstrapping in Applied Welfare Analysis,” American
Journal of Agricultural Economics, 72, 406–418.
Krinsky, L. amd Robb. A. 1986. “On Approximating the Statistical Properties of Elasticities,”
Review of Economics and Statistics, 68, 715-719.
List, J. 2001. “Do Explicit Warnings Eliminate Hypothetical Bias in Elicitation Procedures?
Evidence from Field Auctions for Sportscards.” American Economic Review, 91, 1498-1507.
55


Matsa, D. 2010. “Competition and Product Quality in the Supermarket Industry,” Quarterly
Journal of Economics, forthcoming, available at SSRhttp://ssrn.com/abstract=1440414">N: http://ssrn.com/abstract=1440414.
Mazzeo, M. 2003. “Competition and Service Quality in the U S Airline Industry,” Review of
Industrial Organization, 22, 275-296.
Milyo, J. 2007. “The Effects of Cross-Ownership on the Local Content and Political Slant of
Local Television News,” FCC Media Ownership Study #6, Washington, D.C., FCC.
Morey, E., and Waldman, D. 1994. “Functional Form and the Statistical Properties of Welfare
Measures–A Comment,” American Journal of Agricultural Economics, 76(4), 954-57.
Nalebuff, B., and Stiglitz, J. 1983. “Information, Competition, and Markets,” American
Economic Review, 73(2), 278-283.
NMR. 2007. “How People Get News and Information,” FCC Media Ownership Study #1,
Washington, D.C.
Owen, B., and Wildman, S. 1992. Video Economics, Harvard University Press, Cambridge,
MA.
Pew Research Center. 2010. “How News Happens - Still A Study of the News Ecosystem of
Baltimore,” Pew Research Centre Publication, Accessed on March 31, 2011 at
http://pewresearch.org/pubs/1458/news-changing-media-baltimore">http://pewresearch.org/pubs/1458/news-changing-media-baltimore.
Prior, M. 2002. “Efficient Choice, Inefficient Democracy? The Implications of Cable and
Internet Access for Political Knowledge and Voter Turnout.” In Communications Policy and
Information Technology: Promises, Problems, Prospects, eds. Cranor, F., and S. Greenstein,
The MIT Press, Cambridge, MA.
Purcell, K. 2011. “Trends to Watch: News and Information Consumption.” Presented to the
Catholic Press Association, Annual Meeting, March 24, 2011.
56


Rosston, G., Savage, S., and Waldman, D. 2010. “Household Demand for Broadband Internet
in 2010,” The B.E. Journal of Economic and Policy Analysis (Advances), 10(1), Article 79.
Available at:http://www.bepress.com/bejeap/vol10/iss1/art79"> http://www.bepress.com/bejeap/vol10/iss1/art79.
Savage, S., and Waldman, D. 2008. “Learning and Fatigue During Choice Experiments: A
Comparison of Online and Mail Survey Modes,” Journal of Applied Econometrics 23(3),
351-371.
Schmalensee, R. 1979. “Market Structure, Durability, and Quality: A Selective Survey,”
Economic Inquiry, 17, 177-196.
Shcherbakov, O. 2007. “Measuring Consumer Switching Costs in the Television Industry,”
mimeo, University of Arizona.
Shiman, D. 2007. “The Impact of Ownership Structure on Television Stations’ News and
Public Affairs Programming,” FCC Media Ownership Study #4, Section I, FCC, Washington,
D.C.
Siegelman, P., and Waldfogel, J. 2001. “Race and Radio: Preference Externalities, Minority
Ownership, and the Provision of Programming to Minorities.” In Baye, M., and Nelson, J.,
eds., Advances in Applied Microeconomics, Vol. 10: Advertising and Differentiated Products,
Elsevier, Amsterdam.
Spence, M. 1975. “Monopoly, Quality, and Regulation,” Bell Journal of Economics, 6, 417-
429.
Spence, M., and Owen, B. 1977., “Television Programming, Monopolistic Competition, and
Welfare,” Quarterly Journal of Economics, 91(1), 103-126.
Sweeting, A. 2010. http://econ.duke.edu/~atsweet/SWEETING_mergersJan10.pdf">“The Effects of Mergers on Product Positioning: Evidence from the Music
http://econ.duke.edu/~atsweet/SWEETING_mergersJan10.pdf">Radio Industry,” RAND Journal of Economics, 41(2).
57


United States Census Bureau. 2009. American Factfinder. Washington, D.C.: United States
Census Bureau.
Waldfogel, J., 2004. Who Benefits Whom in Local Television Markets? in Brookings Wharton
Papers on Urban Affairs: Gale, W., and Rothenberg Pack, J., eds., Brookings Institution
Press, pp. 257-305.
Yan, M., and Napoli, P. 2006. “Market Competition, Station Ownership, and Local Public
Affairs Programming on Broadcast Television,” Journal of Communications, 56(1), 795-812.
Zavoina, R. and W. McKelvey. 1975. “A Statistical Model for the Analysis of Ordinal Level
Dependent Variables,” Journal of Mathematical Sociology, 103-120.
Zwerina, K., Huber, J., and Kuhfeld, W. 1996. “A General Method for Constructing Efficient
Choice Designs. in Marketing Research Methods in the SAS System.” Version 8 edition,
SAS Institute, Cary, North Carolina.

58


APPENDIX A: ECONOMETRIC DETAILS


We add subscripts to the random utility model to explicate the econometrics, and
indicate choices A and B with numbers 1 and 2. The unobserved utility of media
environment 5
>2
>2
534
34 œ "ß
# to the 3 person on the 4 choice occasion, Y , is modeled as a
34
linear combination of the choice characteristics and a random error:
Y 534 œ
534
"w 5

B

34
%
ß 3 œ "ß á ß 8; 4 œ "ß á )ß 5 œ "ß #
34
34
34
34
,
(A1)
where the elements of the vector " are the marginal utilities of the elements of the vector

B

34: DIVERSITY OF OPINION, COMMUNITY NEWS, MULTICULTURALISM,
ADVERTISING, and COST . The disturbances %534
34 are assumed to be independent,
identically distributed, mean zero normal random variables, uncorrelated with

B

34, with
constant unknown variance 5#. Under these assumptions, the probability of choosing
%
media environment option ,
" for example, is:
T 1 œ T Ð "
#
Y
Y Ñ œ
w
F’
" Ð #

B

B

1 ÑÎÈ# %“
34
34
34
34
34
5 ,
A2
and similarly for media environment 2, where F( † ) is the univariate standard normal
cumulative distribution function. This is the usual univariate probit model for
dichotomous choice, where the unit of observation is an 3ß 4 pair. The likelihood is the
product of the ) ‚
8 probabilities like equation A1.

After choosing 534, individuals answer a question stating whether media
environment 534 would be chosen over the status quo (as in Figure 2, Question 2). Utility
for the status quo, Y ! under the model assumption (equation 0) is given by:
3
Y ! œ
w
!
!
!
"
" B
% ß
3
3
3
A3
where %! are disturbances and !

B

are the characteristics of the status quo. We include the
3
3
constant "! and not in the utility of the hypothetical choices to capture the real and
psychic costs of switching to a new media environment. The %! are assumed to be
3

independent, identically distributed normal random variables with zero expectation and
5
variance 5#, uncorrelated with % 34 .
!
34

Two choices are then made for each choice occasion. The probability of choosing
media environment 534 Ð"ß #Ñ over media environment $
534 Ð#ß "Ñ, and then choosing
that alternative over the status quo is the bivariate probability:
T ÐY 534
Y $ 534ß Y 534
!
Y Ñ
34
34
34
3
A4
œ T
$ 534
534
w
Š%
%
" Ð $ 534
534

B

B

Ñß !
534
w
%
%
" Ð !
534

B

B

Ñ
3
34 ‹
34
34
34
34
3
34
œ
w
"
w
!
534
#

B

B

#
É
#
F ’
Ð $ 534
534

B

B

Ñ È
Î
#5%ß
"
5
5
0
% 3“
34
34
Ð
Ñ
3
34 Î
à
$ 534
534
5
where 3 is the correlation between %
% and %!
% 34
34
34
3
34 ,
5#
5
3 œ
%
œ
%
,
A5
5
# # 5
Ð #
5#Ñ
5
#Ð #
È
È
5#Ñ
%
0
%
0
%
and F# is the standard bivariate normal cumulative distribution function. Similar
expressions give the probability of choosing media environment 534 over alternative
$
534 and then choosing the status quo over that media environment.

One normalization is required and one additional parameter is identified. Let
5
È
#
#
#
#
% œ "Î
#. Define - œ 5 /5 œ 5
#
Þ The parameter -
!
%
, while not essential to this
!
research, is interesting in its interpretation: the ratio of the variance of the random part of
utility in the evaluation of the current forecast to that of the hypothetical forecasts. Then
equation A3 can be written:
T ÐY 534
Y $ 534ß Y 534
!
Y Ñ œ
34
34
34
3
A6
w
" Ð !
534

B

B

Ñ
w
"
F
3
34

B

B

# ’
" Ð $ 534
534 Ñß
à

34
34
È
.
Ð"
-#ÑÎ# È#Ð"
-#Ñ
The likelihood is again the product of the ) ‚
8 probabilities like equation 5.

To examine WTP estimates, we rewrite the utility model for media environment
as:

Y ‡ œ
w
534

B

-
" -34
"-w 34
3
% 4ß 3 œ "ß á ß 8à 4 œ "ß á ß )
34
.
A7
where -34 is cost and "-w is the vector of characteristics of the media environment other
than cost. The estimated of WTP for these characteristics are

A

s œ
s
" /s .
-w "-
Since the
estimates of willingness-to-pay are nonlinear function of parameter estimates, their exact
standard errors are unknown. While it would be possible to bootstrap the distribution of
these estimators, since the normally distributed estimator of "- is the denominator, the
simulation would not converge to anything useful (see Kling and Sexton, 1990; Morey and
Waldman, 1994). Instead, we use a linear approximation to the variance (sometimes
known as the “delta method”). This approximation for elasticities has been examined in
Krinsky and Robb (1986). Let

D

s be the estimated variance-covariance matrix of "s. The
linear approximation to the variance of

A

s is
`

A

w
`

A

Z
s Ð Ñ

A

¸
s
s

D



A8
`"
`"
where the derivatives are evaluated at the parameter estimates. The square root of the
diagonal elements of Z
s Ð s

A


Ñ are the estimated standard errors of the estimates of WTP.
These derivatives are
Î
"
s+"
"
#
!
! á
! Ñ
Ð
"
s
"
s
-
-
Ó
Ð
"
s+
Ó
Ð
#
#
!
"
!
á
! Ó
Ð
"
s
"
s
-
-
Ó
` s

A

Ð
Ó
œ Ð
ã
ã
ã
ã Ó
A9
` s
"
Ð
Ó
Ð
"
s+O
Ó
Ð
#
!
!
!
á
!
s
Ó
Ð
"-
Ó
Ï
"
s=#
!
!
!
á
"
s
s Ò
"
"
-
-

APPENDIX B: SURVEY QUESTIONS FOR FEATURES AT HOME

Your overall media environment has a monthly cost. This is the total of any monthly
subscriptions to newspapers, satellite radio, cable or satellite TV, and the Internet (including
the data feature of a Smartphone contract), as well as any contributions to public radio or TV.

Some companies bundle TV, the Internet, and/or telephone service into a service plan with a
single monthly bill. Do you pay a single monthly bill for a bundle of services? (mark one
answer)


Yes



No


About how much do you pay per month for paper or online newspapers? (write “0” if you read
free paper or online newspapers)

$ __________ per month

About how much do you pay per month for satellite radio service?

$ __________ per month

About how much do you pay per month for the TV service at your home? (If you bundle TV
with other services, just write the TV portion of your bundle)


$ __________ per month

During the past 12 months, about how much did you contribute to public radio stations (e.g.,
NPR) and/or public TV stations (e.g., PBS)? (write “0” if you made no contributions)

$ __________

About how much do you pay per month for the Internet service at your home? (If you bundle
Internet with other services, just write the Internet portion of your bundle)


$ __________ per month

About how much do you pay per month for the data feature of your Smartphone contract? (If
you bundle a Smartphone contract with other services, just write the Smartphone portion of
your bundle; write “0” if you do not subscribe to a data feature)


$ __________ per month


59


Besides cost, there are some other features of your media environment. They are the . . .

(1) diversity of opinion in reporting information;
(2) amount of information on community news and events;
(3) coverage of multiculturalism: i.e. ethnic, gender, and minority related issues; and
(4) amount of advertising

We are especially interested in what is available in your media environment, rather than what
you usually view or listen to. It is possible that your media environment provides you with
many different news options. This is what we would like to know, even if you choose to view
or listen to the same news programs every day. Please bear this in mind when completing the
next four questions.

A *low* diversity of opinion media environment provides information on news and current
affairs from only one viewpoint. For example, if only a single local newspaper was available,
the diversity of opinion from newspapers would be low. In a *medium* diversity of opinion
environment the information available could come from a few different viewpoints. For
example, if you could watch CNN and Fox news for different opinions about a national issue,
the diversity of opinion from TV would be medium. As another example, if you could listen to a
few radio talk shows with different opinions about an issue, the diversity of opinion from radio
would also be medium. A *high* diversity of opinion environment provides information from
many different viewpoints.
Please indicate the level of diversity of opinion from your overall media environment. (mark
one answer)


Low


Medium

High

Now consider the information available to you on community news and events. Examples are
reports on: school sporting results, local council meetings, city/county elections, neighborhood
crime, local heroes who give their time to the community, or job layoffs at a local factory.
A media environment with *low* community news provides very little or no information on
community news and events. For example, if you live in a town without a local newspaper or
local TV station, the amount of community news from your media environment may be low.
With *medium* community news, there is some information on community news and events.
With *high* community news, there is a lot of information on community news and events.
Please indicate the level of community news from your overall media environment. (mark one
answer)


Low


Medium

High

Next, consider the information available to you that reflects the interests of women and
minorities. We will call this multiculturalism. Examples are reports on: Black History month,
the Cinco de Mayo celebration, female wage inequality, or programs that help people with
disabilities find a job.
A media environment with *low* multiculturalism provides very little or no information that
reflects the interests of women and minorities. With *medium* multiculturalism, the information
reflects some of the interests of women and minorities. With *high* multiculturalism, the
information reflects many of the interests of women and minorities.
60


Please indicate the level of multiculturalism from your overall media environment. (mark one
answer)


Low


Medium


High

Finally, consider the amount of advertising.
With *low* advertising, the amount of space on a newspaper or web page, or the amount of air
time devoted to commercial advertising on radio or TV, is barely noticeable. With *medium*
advertising
, the space or time devoted to advertising is more noticeable. With *high*
advertising
, the space or time devoted to advertising is very noticeable, to the point of being
annoying when you are viewing or listening to your media source.

Is the advertising in your overall media environment barely noticeable (Low), noticeable
but not annoying (Medium), or annoying (High)? (mark one answer).

Low


Medium


High



61


APPENDIX C: FOCUS GROUP INVITATION, PROTOCOLS, QUESTIONS,

NOTES

Invitation to a survey “Focus Group”


I will pay you $25, cash, to take a 15 minute survey and then
discuss it with me in an informal setting. Total time: about one
hour. Refreshments will be provided!
The survey is about your use of newspapers, TV, radio, and the
Internet.
Date:_Thursday, December 9, TODAY_
Time:_About 6:10 pm__
Directions: come to Room 5 in the basement of the Economics
building (in front and to the left, as you walk to campus through
the Broadway underpass at College Ave).
Thanks very much,
Don Waldman
Professor, Department of Economics
Your Name:___________________________
Telephone #:___________________________
(I will call you to remind you, if needed)
Miscelleneous comments and issues from focus groups
Opening pages: news stand should be newsstand
newspaper time question: add "(for 1/2 hour, enter .5, for 15 minutes, enter .25, etc.)"

"On a typical day, about how much of the 1 hour do you spend listening to the radio to get
information on news and current affairs?" should be "On a typical day, about how much
of the 1 hour you spend listening to the radio is to get information on news and current
affairs?" Same for TV.
"Answer '2.1' is not in range '0 - 2'. " Sounds like a math lesson. It would be better if it
read "You answered that you [listen to radio/watch TV/spend using the Internet] for 2
hours. The amount of time getting news and information from [listen to radio/watch
TV/spend using the Internet] must be less than 2 hours."
After one prompt that time getting news is too large, go to next question.
Difference between cable and satellite TV important? Does everyone know the
difference?
How much do you pay per month for the data featur
e of your Smartphone contract?
(Italics needed).
TV and phone bundle: typed 90. TV portion: 99. Took it.
In summary of media env. and follow up questions, round off amount paid to nearest
dollar. $159.833333333333 per month
Need to redo design, to eliminate $5 costs for the full environment. Some crazy choices:
$5, high, medium, high, medium vs. $180 low, medium, medium, low? Redo by hand.
After WTP question, remind respondent they are comparing the full media environment.
In description of localism, and elsewhere, we have the phrase "from your media
environment" which could be dropped.
"All" spelled "Al" in first choice question.
Rewrite descriptions of certainty levels - see Freya's email.
In follow up questions, instead of "I prefer option B", "I prefer option B to my media
envirnment," or, "I still prefer option B."
"We would like to have your answer to this question" to "We would really like to have
your answer to this question"
Number the choice questions, or have a status bar, to let respondent know how much they
have to do.

Smartphone question "a Android . . ." should be "an Android . . ."
Add smartphone data contract to intro on cost
How much do you pay per month for the data feature of your Smartphone contract?
Italics needed. Position of question wrong.
Try Table 1 inputs for differences
Internet capitalized throughout? See intro to choice questions.
"Journalistic viewpoints" ???
"pretty" certain
Need to check the congruency of version number and attribute levels
Ask subjects if they like the text balloons
Tell subjects to ignore the first screen, and the two "Data only" screens that appear later.
Both Dave and Ken had even less trouble with the survey than the women. Ken is a KN
panelist. They thought our descriptions and examples were good. They understood the
exercise completely. Number of choice questions OK, maybe helps to divide up 4 and 4
with the payment card question. Shading good (Ken), didn't realize (Dave). Cheap talk
on taking cost seriously. Some problems with things that will not be a problem with the
online version.
Dave was fairly certain about all but one choice, for which he was certain, and Ken was
half very certain, half fairly certain; Susan was two very certain, six fairly certain, one
somewhat certain; Ann Marie's answers were spread out, although no one had picked the
least certain category.
Cost of media environment: Dave: $93, Ken: $25; Ann Marie: $100; Susan: $15
Switch/Stay results:
Ann Marie
Susan
Ken
Dave
Switch
8
2
0
0
Stay
1
7
9
9
Both Dave and Ken had even less trouble with the survey than the women. Ken is a KN
panelist. They thought our descriptions and examples were good. They understood the
exercise completely. Number of choice questions OK, maybe helps to divide up 4 and 4
with the payment card question. Shading good (Ken), didn't realize (Dave). Cheap talk

on taking cost seriously. Some problems with things that will not be a problem with the
online version.
Dave was fairly certain about all but one choice, for which he was certain, and Ken was
half very certain, half fairly certain; Susan was two very certain, six fairly certain, one
somewhat certain; Ann Marie's answers were spread out, although no one had picked the
least certain category.
Cost of media environment: Dave: $93, Ken: $25; Ann Marie: $100; Susan: $15
Switch/Stay results: Ken would not switch on any choice, which was consistent with his
home media environment ($45, Low, High, Medium, High). Dave likewise would not
switch, which was inconsistent with one of his choices; Ann Maria switched on all but one;
Susan stayed 7 times, switched twice. Susan and Ann Marie were consistent.
Time: about 22 minutes for Ken and Dave.
Media Ownership Study
Focus Group Protocol
Online Administration at RRC Associates, Boulder

We will conduct two focus groups. The first will have five subjects and be
administered by and on the premises of RRC Associates of Boulder, CO. The second will
have approximately 10 subjects and be in the computer lab of the Economics building at
the University of Colorado.
First focus group
Subjects for the first focus group will be recruited by RRC from the general population of
Boulder County to be heterogenous with respect to gender, age, income, and education.
They are paid approximately $50 for their participation. As in the survey, they will be
asked to bring with them any media bills that they have on hand. We will sit behind a one-
way mirror observing the session. At any point we can interrupt. In the past we have
usually waited until the end of the session to introduce ourselves as the researchers and
ask any questions that may have arisen during the survey. The entire session is recorded
on a DVD for further study.

Before the subjects arrive we meet with the facillitator to review the survey and
discuss our general objectives and goals for the session. The facillitator has previously
had access to the survey and our list of questions. They provide input on their role in the
procedure.
Format


Subjects come to RRC's offices and meet with the facillitator. They sit in a small
room at a desk with a computer, with the facillitator sitting next to them. The instructions
spoken by the facillitator are approximately:
"Imagine you have agreed to take a short survey on the Internet. Proceed through the
survey as you would at home. From time to time I will ask for your reactions to some of
the questions. At any point while taking the survey, please tell me if you are unsure of
what any question is asking."
Questions
Particular questions that the facillitator will ask. Here the format is somewhat
freewheeling. The facillitator is allowed to follow up on responses, and to tailor questions
to individuals.
1.
After the section of the survey "Sources of Information:" "How difficult was it for
you to estimate the number of hours you spend each week using the different media
sources? Would anything make it easier for you to guess at the answers to these
questions?"
2.
After the section of the survey "Cost of Your Media Environment:" "How certain
were you about the dollar amounts you entered? Do you think you were able to
differentiate any combined bills you may have for TV, the Internet, etc."
3.
After they read the information under "Features of Your Media Environment:” "Is
it clear to you what we mean by 'features of your media environment?' "
4.
After Q39: "We asked for a lot of information in this question! We're you able to
figure out how to mark the table? Do you think you were able in your answers to
distinguish the amount of advertising in the different sources, or do you think you kind of
generally answered from your overall feeling about the amount of advertising you see or
hear?"
5.
After Q40,if they checked "No": "Did we miss by much?" "Were you at all put off
by us for telling you what we thought your overall advertising exposure was?" If they
checked "Yes": "Do you really think we got it about right, or were you not sure and you
were you just agreeing?"
6.
After Q42: "Is it clear what we mean by diversity of opinion?" Again, how
difficult was it for you to distinguish the levels of diversity for the four media sources?"
7.
Question 6 is repeated for the features localism (Q45-47) and multiculturism
(Q48-50).

8.
After Q51: "Is it clear what we are asking here?" Probing: "instead of the levels of
the attributes advertising, etc., we are now inquiring about your level of satisfaction with
the features. Was that clear to you? Do you think this question is burdensome?"
9,
After they have viewed Table 3, with time to roll their mouse pointer over a
feature of a level: "Do you think you have a good feel for what we mean by your media
environment? Do you have a good idea of what your own media environment is like from
answering all these questions?"
10.
After Q52: "Is it clear what you are choosing in this question? How difficult was it
for you to decide between the two choices? Did either environment appeal to you?"
Then, if they had to be prompted to make a choice: "Would you have liked to skip this
question? After the question returned to the screen, did you simply mark a choice to
move on, or would you really prefer the choice you were given?"
11.
This line of questioning is repeated after Q53.
12.
Now we let recruits answer a few more choice questions, and then ask them again
questions like 10. This is repeated once more toward or at the end of the eight questions.
13.
After the end of the choice questions: "Did you feel there were too many of these
choices to make? Were you thinking as hard as you were in the beginning when you
made these choices?"
14.
After Q70: "Was it clear what we were asking for in this question? You answered
eight choice questions before this one. What are your thoughts on the following: you
answer four choice questions, then this one, then four more choice questions. Would that
have made any difference for you?"
Second Focus Group
This is a different approach altogether. We recruit 10 or so people, paying $25/person,
some students, some other campus area people (secretaries, professors, shop keepers in
the immediate campus area). They all sit at the computer and take the survey
uninterrupted, at the same time. After they are finished answering the questions . . .
At the conclusion of the survey, they write their remarks on any aspect of the survey. We
prompt them with a screen of issues to guide their remarks.

APPENDIX D: PANEL RECRUITMENT AND NON-RESPONSE38

D1.
Panel Recruitment
KN panel members are recruited by either formerly random digit dialing (RDD) sampling or
the current address-based sampling (ABS) methodologies. To offset attrition, multiple
recruitment samples are fielded evenly throughout the calendar year.

Random digit dialing
KN’s recruitment methodology has used the standards established by selected RDD
surveys conducted for the Federal government, such as the CDC-sponsored National
Immunization Survey. KN employed list-assisted RDD sampling techniques based on a sample
frame of the U.S. residential landline telephone universe. For purposes of efficiency, KN
excludes only those banks of telephone numbers (a bank consists of 100 numbers) that had
fewer than two directory listings. Additionally, an oversampling was conducted within a
stratum of telephone exchanges that had high concentrations of African American and Hispanic
households based on Census data.
A telephone number for which a valid postal address can be matched occurred in about
67 to 70 percent of each sample. These address-matched cases were all mailed an advance
letter informing them that they had been selected to participate in KN’s panel. Following the
mailings, telephone recruitment by trained interviewers begins for all sampled telephone
numbers. All members of the household are enumerated, and some initial demographic and
background information on prior computer and Internet use was collected.

38 Documentation on KN’s sampling, data collection procedures, weighting, and IRB-bearing issues are available
at: http://www.knowledgenetworks.com/ganp/reviewer-info.html
http://www.knowledgenetworks.com/knpanel/index.html">http://www.knowledgenetworks.com/knpanel/index.html; and
http://www.knowledgenetworks.com/ganp/irbsupport/.
62


Households with a home computer and Internet access were asked to take KN surveys
using their own equipment and Internet connection. Households without Internet access are
provided with a laptop computer and free monthly Internet access. Each computer is custom-
configured prior to shipment with individual email accounts so that it is ready for immediate
use by the household. Most households are able to install the hardware without additional
assistance, although KN maintains a toll-free telephone line for technical support.
Once a household is recruited and each household member’s e-mail address is either
obtained or provided, panel members are sent survey invitations linked through a personalized
e-mail message. This permits surveys to be fielded quickly and economically, and also
facilitates longitudinal research. In addition, this approach reduces the burden placed on
respondents, since e-mail notification is less intrusive than telephone calls and allows research
subjects to participate in research when it is convenient for them.

Address-Based Sampling
In 2009, Knowledge Networks introduced the ABS sample frame to panel recruitment
to reflect the recent changes in society and telephony. These changes, which have reduced the
long-term scientific viability of landline RDD sampling methodology, are: declining
respondent cooperation in telephone surveys as reflected in “do not call” lists, call screening,
caller-ID devices, and answering machines; dilution of the RDD sample frame as measured by
the working telephone number rate; and the emergence of cell phone-only households which
are excluded from the RDD frame.
ABS involves probability-based sampling of addresses from the U.S. Postal Service’s
Delivery Sequence File. Randomly sampled addresses are invited to join the KN panel through
63


a series of mailings and, in some cases, telephone follow-up calls to non-responders when a
telephone number can be matched to the sampled address. Operationally, invited households
have the option to join the panel by one of several ways: completing and returning a paper form
in a postage-paid envelope; calling a toll-free hotline maintained by KN; or going to a
dedicated KN web site and completing an online recruitment form.
After initially accepting the invitation to join the panel, respondents are then “profiled”
online by answering key demographic questions about themselves. This profile is maintained
through the same procedures that were established for RDD-recruited panel members.

D2.
Non-Response Bias
Multivariate Analysis of Survey Non Response
Table D1 presents summary statistics for the gross sample of 8,621 panel members
invited to participate in the media environment survey. Column two and column three of Table
D2 report the estimates from the probit model that relates FULL SAMPLE (equals one if the
respondent participated in the survey and zero otherwise) to selected demographic and regional
variables Column four and five of Table D2 report the estimates from the probit model that
relates FINAL SAMPLE (equals one if the respondent is included in the final sample and zero
otherwise) to selected demographic and regional variables. The results indicate potential
differences in age, gender, education and Internet access between our final sample and the
population. We remedy this possible source of bias in our results in Section 5 by estimating
with weighted maximum likelihood, where the contribution to the log likelihood is the post-
stratification weight times the log of the bivariate probability for the individual choice
occasion.
64


Sample Weights
Post-stratification weights were used to adjust sample estimates for any survey non-
response, as well as any non-coverage or under- and over-sampling resulting from the study-
specific sample design. Demographic and geographic distributions for the non-
institutionalized, civilian population ages 18 and over from the most recent CPS were used as
benchmarks in this adjustment. The benchmark distributions used for the post-stratification
adjustment are: age/gender (18 to 29 years male, 18 to 29 years female, 30 to 44 years male, 30
to 44 years female, 45 to 59 years male, 45 to 59 years female, 60 years and over male, and 60
years and over female); race/ethnicity (White/non-Hispanic, Black/non-Hispanic, other/non-
Hispanic, two or more races/non-Hispanic, Hispanic); education (less than high school, high
school, some college, bachelor degree and beyond); census region/metropolitan area (northeast
metro, northeast non-metro, midwest metro, midwest non-metro, south metro, south non-metro,
west metro, west non-metro); and Internet Access (yes, no).
Comparable distributions were calculated by using all completed cases from the field
data minus the dropped interviews (i.e., n = 5,131). Since study sample sizes are typically too
small to accommodate a complete cross-tabulation of all the survey variables with the
benchmark variables, a raking procedure is used for the post-stratification weighting
adjustment. Using the base weight as the starting weight, this procedure adjusts the sample
data back to the selected benchmark proportions.39 Through an iterative convergence process,
the weighted sample data are optimally fitted to the marginal distributions. After this final

39 The KN panel sample has several known sources of deviation from an equal probability of selection design.
These are corrected in the base weight and are: under-sampling of telephone numbers unmatched to a valid
mailing address; RDD selection proportional to the number of telephone landlines reaching the household;
oversampling of Chicago and Los Angeles in early pilot surveys; early oversampling the four largest states and
central region states; under-sampling of households not covered by the MSN® TV service network; RDD
oversampling of African American and Hispanic telephone exchanges; and ABS phone match adjustment and
oversample stratification adjustment.
65


post-stratification adjustment, the distribution of the calculated weights are examined to
identify and, if necessary, trim outliers at the extreme upper and lower tails of the weight
distribution. The post-stratified and trimmed weights are then scaled to the sum of the total
sample size of all eligible respondents.

Table D1. Summary Statistics for KN Gross Sample


Demographic
Description
Obs
Mean
s.d.
Min
Max
FULL SAMPLE
1 if the respondent is in the full
8,621
0.644
0.479
0
1
sample, 0 otherwise.

AGE
1 if 18-24 years; 2 if 25-34; 3 if
8,621
3.815
1.686
1
7
35-44; 4 if 45-54; 5 if 55-64; 6 if
65-74; 7 if 75 years or over.

RACE
1 if white; 0 otherwise.
8,621
0.697
0.459
0
1

GENDER
1 if female; 0 otherwise.
8,621
0.530
0.499
0
1

MARITAL STATUS
1 if married; 0 otherwise.
8,621
0.535
0.499
0
1

EDUCATION
1 if less than high school; 2 if
8,621
2.783
0.982
1
4
high school; 3 if some college; 4
if bachelors degree or more.

HOUSEHOLD INCOME
1 if less than $10,000; 2 if
8,621
3.593
1.293
1
5
$10,000-$24,999; 3 if 25,000-
$49,999; 4 if $50,000-$74,999; 5
if $75,000 or more.

EMPLOYMENT
1 if in work force; 0 otherwise.
8,621
0.568
0.495
0
1

INTERNET
1 if Internet access is not
8,621
0.800
0.400
0
1
provided by KN, 0 otherwise.
NORTHEAST
1 if respondent resides in the
8,621
0.180
0.385
0
1
Northeast census region; 0
otherwise.

MIDWEST
1 if respondent resides in the
8,621
0.234
0.424
0
1
Midwest census region; 0
otherwise.

SOUTH
1 if respondent resides in the
8,621
0.356
0.479
0
1
South census region; 0 otherwise.

WEST
1 if respondent resides in the
8,621
0.229
0.420
0
1
West census region; 0 otherwise.

NOTES. Obs is number of observations. s.d. is standard deviation. Min is minimum value. Max is maximum value.
66



Table D2. Probit Estimates of Gross Sample Respondents


Full sample
Final sample

dF/dx
|z|
dF/dx
|z|
AGE
0.041***
12.07
0.060***
16.93
RACE
0.095***
8.00
0.095***
7.80
GENDER
-0.053***
5.08
-0.043***
3.92
MARITAL STATUS
-0.013
1.14
-0.010
0.81
EDUCATION
0.013**
2.25
0.016**
2.67
HOUSEHOLD INCOME
0.001
0.83
0.002
0.38
EMPLOYMENT
-0.031***
2.69
-0.021*
1.75
INTERNET
0.053***
3.53
0.035**
2.28
NORTHEAST
0.008
0.65
-0.000
0.01
MIDWEST
0.023
1.47
0.031**
1.96
SOUTH
-0.015
1.06
-0.005
0.34
Likelihood
-5,435.6

-5,559.8

Observations
8,261

8,261

NOTES. “Full sample” equals one when the respondent participated in the survey, zero otherwise. “Final sample” equals
one when the respondent is included in the final sample for analysis. dF/dx is the effect of a marginal change in the
independent variable on the probability of being in the final sample. z is the z value, calculated from robust standard
errors. ***denotes significant at the one percent level. **denotes significant at the five percent level. *denotes significant
at the ten percent level.





67



APPENDIX E. MEDIA MARKET STRUCTURE VARIABLES

Variable
Description
HOUSEHOLDS
Households in the market.
NEWSPAPERS
Number of daily newspapers with a city of publication located in a county in the market.
RADIO STATIONS
Number of radio stations in the market.
TV STATIONS
Number of television stations in the market.
MEDIA OUTLETS
NEWSPAPERS plus RADIO STATIONS plus TV STATIONS.
NEWSPAPER VOICES
Number of parent entities owning a daily newspaper in the market.
RADIO VOICES
Number of independent radio voices in the market.
TV VOICES
Number of independent television voices in the market.
VOICES
NEWSPAPER VOICES plus RADIO VOICES plus TELEVISION VOICES.
TV-NEWSPAPER VOICES
Number of independent newspaper and television voices in the market.
TV-RADIO VOICES
Number of independent radio and television voices in the market.
NOTES. All data are from the FCC (2011).



68



Table 1. Media Environment Features

Characteristic
Levels
The total of monthly subscriptions to all of the household’s media
COST
sources, plus any contributions to public radio or public TV stations
(ranging from $0 to $250 per month).
The extent to which the information on news and current affairs in the
household’s overall media environment reflects different viewpoints.
DIVERSITY OF OPINION
Low: only one viewpoint.
Medium: a few different viewpoints.
High: many different viewpoints.
The amount of information on community news and events in the
household’s overall media environment.
COMMUNITY NEWS
Low: very little or no information on community news and events.
Medium: some information on community news and events.
High: much information on community news and events.
The amount of information on news and current affairs in the
household’s overall media environment that reflects the interests of
women and minorities.
Low: very little or no information reflecting the interests of women
MULTICULTURALISM
and minorities.
Medium: some information reflecting the interests of women and
minorities.
High: much information reflecting the interests of women and
minorities.
The amount of space and/or time devoted to advertising in the
household’s overall media environment.
ADVERSTISING
Low: barely noticeable.
Medium: noticeable but not annoying.
High: annoying.










69



Table 2. Demographic Distributions (%)



Census
KN panel
KN sample



Gross sample
Full sample
Final sample
Omitted
(Invited)
(Completed)
(Completed)
(Completed)
Region






Northeast
18.4
18.5
18.0
18.5
18.3
21.8
Midwest
21.8
22.1
23.4
24.7
24.9
21.3
South
36.5
35.9
35.6
34.2
34.4
31.4
West
23.2
23.5
22.9
22.6
22.4
25.4
Age






18-24 years
12.6
10.7
10.7
8.6
7.6
21.8
25-34 years
17.8
17.4
15.0
12.7
11.3
29.7
35-44 years
17.8
18.9
16.6
16.1
15.8
20.1
45-54 years
19.5
18.5
20.1
20.3
20.9
13.2
55-64 years
15.5
18.5
20.3
22.2
23.3
9.1
65 years or over
16.8
16.0
17.3
20.1
21.2
6.0
Race






Non-white
18.9
20.9
30.3
26.0
25.4
33.3
White
81.1
79.1
69.7
74.0
74.6
66.7
Gender






Female
51.7
52.6
53.0
50.8
51.1
47.5
Male
48.3
47.4
47.0
49.2
49.9
52.5
Marital status






Married
55.1
52.5
53.5
55.1
55.8
46.3
Not married
44.9
47.5
46.5
44.9
44.2
53.7
Education






< High school
13.8
12.9
10.8
9.6
9.5
10.8
High school
30.7
29.6
29.0
30.2
30.0
32.6
Some college
28.2
29.1
31.3
29.8
30.0
27.6
Bachelors degree or
27.4
28.3
28.9
30.4
30.5
29.0
higher
Household income






< $10,000
6.6
7.0
7.1
6.3
6.3
6.7
$10,000-$24,999
16.8
16.1
15.1
14.4
14.3
14.6
$25,000-$49,999
26.2
26.1
24.3
24.8
24.6
27.1
$50,000-$74,999
19.5
20.3
18.3
18.8
19.1
15.9
> $75,000-
30.8
30.4
35.2
35.7
35.7
35.7
Employment






In labor force
66.1
67.3
56.8
55.2
54.7
60.7
Not in labor force
33.9
32.7
43.2
44.8
45.3
39.3
Internet access
64.0
66.0
73.0
81.2
80.6
88.2
Observations
n.a.
n.a.
8,621
5,548
5,131
417
NOTES. Census data are from December, 2009. KN panel data are from January, 2010. Remaining data are from March, 2011.
SOURCE
. United States Census Bureau (2009); Knowledge Networks, Inc. (2010).




70



Table 3. Summary Statistics for Media Environment Sources


Media source
Obs
Sample
Mean
s.d.
Min
Max
share (%)
Newspaper
2,342
45.6
1.015
1.766
0
24
Radio
4,154
81.2
1.423
1.873
0
24
Satellite radio
558
10.9
1.522
2.221
0
24
Television
4,856
94.6
1.953
2.172
0
24
Cable television
2,736
53.4
1.976
2.210
0
24
Satellite television
1,381
27.0
2.071
2.197
0
24
Own Internet
4,135
80.6
1.074
1.659
0
24
Smartphone
1,270
24.8
0.580
1.344
0
24
NOTES. Obs is number of observations. Sample share is the percentage of the sample that uses the media
source. s.d. is standard deviation. Min is minimum value. Max is maximum value. Own Internet is home
Internet service not provided by KN.



71



Table 4. Distribution of Media Source Combinations


Newspaper
Radio
Television
Internet
Smartphone
Count
Sample






share (%)





7
0.1




8
0.2








15
0.3






21
0.4






29
0.6




56
1.1







59
1.1







56
1.1






80
1.6







87
1.7








336
6.5






411
8.0







504
9.8




558
10.9






1,345
26.2






1,559
30.4
NOTE
S. 5,131 observations. Count is the number of respondents that use the media source or
combination of media sources. Sample share is the percentage of the sample that use the media
source or combination of media sources.



72



Table 5. Summary Statistics for Levels of

Media Environment Features


Feature
Obs
Mean
s.d.
Min
Max
DIVERSITY OF OPINION
5,131
2.09
0.655
1
3
COMMUNITY NEWS
5,131
1.99
0.711
1
3
MULTICULTURALISM
5,131
1.83
0.705
1
3
ADVERTISING
5,131
2.29
0.682
1
3
COST ($ per month)
5,131
111.2
76.03
0
447
CONTRIBUTION ($ annual)
535
111.5
161.5
0.25
1,500
BUNDLE
3,688
0.576
0.494
0
1
NOTES. 1 = “low”, 2 = “medium” and 3 = “high” for DIVERSITY OF OPINION, COMMUNITY
NEWS, MULTICULTURALISM, and ADVERTISING. CONTRIBUTION is value of contributions to
public radio and public television stations during the past 12 months. BUNDLE = 1 when subscription
television service is bundled with Internet service and/or other telephone services. Obs is number of
observations. s.d. is standard deviation. Min is minimum value. Max is maximum value.



73



Table 6. Distribution of Satisfaction with Media

Environment Features


Feature
Obs
Not



Very
satisfied



Satisfied
1.
2.
3.
4.
5.
DIVERSITY OF OPINION
5,128
3.1 %
7.4 %
40.1 %
30.2 %
19.2 %
COMMUNITY NEWS
5,128
2.7 %
9.2 %
35.0 %
33.0 %
20.1 %
MULTICULTURALISM
5,127
3.2 %
9.6 %
42.6 %
27.5 %
17.0 %
ADVERTISING
5,124
15.0 %
23.3 %
37.5 %
15.7 %
8.5 %
COST
5,126
14.9 %
25.8 %
35.6 %
13.1 %
10.7 %
OVERALL ENVIRONMENT
5,125
2.8 %
8.0 %
44.9 %
31.0 %
13.1 %
NOTES. Data are from responses to the question: your media environment provides you with information on
news and current affairs from the following sources; [INSERT TEXT FROM LOOKUP TABLE 2]. On a scale from 1 to
5, with 1 indicating “not satisfied” and 5 indicating “very satisfied”, how satisfied are you with each feature?
Also, how satisfied are you with your overall media environment? Obs is number of observations. All other
cells are percent of respondents indicating 1, 2, 3, 4, and 5.



74



Table 7. Summary Statistics for Satisfaction with

Media Environment Features


Feature
Obs
Mean
s.d.
Min
Max
DIVERSITY OF OPINION
5,128
3.55
0.983
1
5
COMMUNITY NEWS
5,128
3.59
0.994
1
5
MULTICULTURALISM
5,127
3.46
0.987
1
5
ADVERTISING
5,124
2.80
1.135
1
5
COST
5,126
2.79
1.169
1
5
OVERALL ENVIRONMENT
5,125
3.43
0.916
1
5
NOTES. Data are from responses to the question: your media environment provides you with information
on news and current affairs from the following sources; [INSERT TEXT FROM LOOKUP TABLE 2]. On a scale
from 1 to 5, with 1 indicating “not satisfied” and 5 indicating “very satisfied”, how satisfied are you with
each feature? Also, how satisfied are you with your overall media environment? Obs is number of
observations. s.d. is standard deviation. Min is minimum value. Max is maximum value.



75



Table 8. Summary Statistics for Media Market Structure

Variable
Markets
Mean
s.d.
Min
25th
75th
Max
All markets







HOUSEHOLDS
203
1,670,158
1,842,396
4,145
447,396
2,228,143 7,444,659
SMALL MARKETS
203
0.084
0.278
0
n.a.
n.a.
1
MEDIA OUTLETS
203
138.7
71.25
4
80
182
291
VOICES
203
73.11
35.97
3
44
97
152
NEWSPAPERS
203
12.76
8.206
0
6
19
32
RADIO STATIONS
203
113.2
59.41
3
64
157
241
TV STATIONS
203
12.74
5.879
1
8
17
27
NEWSPAPER VOICES
203
7.634
4.076
0
4
10
19
RADIO VOICES
203
55.12
28.75
2
31
73
119
TV VOICES
203
10.36
4.626
1
7
13
22
TV-NEWSPAPER VOICES
203
11.91
4.758
1
8
15
24
TV-RADIO VOICES
203
63.06
30.95
2
38
85
129






Small markets (five or less television stations)
HOUSEHOLDS
68
195,814
98,806
4,145
116,273
264,844
395,620
MEDIA OUTLETS
68
46.97
15.90
4
37
57
86
VOICES
68
26.36
8.695
3
20
34
41
NEWSPAPERS
68
4.160
2.347
0
2
6
11
RADIO STATIONS
68
38.60
13.85
3
30
48
75
TV STATIONS
68
4.211
1.060
1
4
5
5
NEWSPAPER VOICES
68
3.308
1.900
0
2
4
8
RADIO VOICES
68
19.00
6.608
2
14
25
31
TV VOICES
68
4.046
1.097
1
3
5
5
TV-NEWSPAPER VOICES
68
5.734
1.302
1
5
7
8
TV-RADIO VOICES
68
22.54
7.316
2
17
28
35
NOTES. Markets is the number of television markets. s.d. is standard deviation. Min is minimum value. Max is maximum value.
25th is 25th percentile. 75th is 75th percentile. 5,123 observations are used to calculate summary statistics for all markets. 432
observations are used to calculate summary statistics for small markets. n.a. is not applicable.



76



Table 9. Ordered Probit Estimates of Media Environment Satisfaction

and Television Market Structure

DIVERSITY OF
COMMUNITY
MULTI-
ADVERTISING
OVERALL
OPINION
NEWS
CULTURALISM
MEDIA
ENVIRONMENT
TV STATIONS
-0.016
-0.001
-0.021*
-0.016
0.0003
(0.012)
(0.012)
(0.012)
(0.012)
(0.012)
TV VOICES
0.026
0.011
0.040**
0.021
-0.004
(0.017)
(0.017)
(0.017)
(0.017)
(0.017)
TV STATIONS
0.0001
-0.0005
-0.0003
-0.00004
0.0001
 TV VOICES
(0.0005)
(0.0005)
(0.0005)
(0.0005)
(0.0005)
AGE
0.041***
0.044***
0.035***
-0.043***
0.017*
(0.009)
(0.009)
(0.009)
(0.009)
(0.090)
EDUC
0.083***
0.002
0.044***
-0.041**
0.003
(0.017)
(0.016)
(0.016)
(0.016)
(0.017)
GENDER
0.014
0.033
-0.013
0.073**
0.063**
(0.030)
(0.030)
(0.030)
(0.030)
(0.030)
INCOME
0.004
-0.013
0.014
-0.026**
-0.027**
(0.013)
(0.013)
(0.013)
(0.013)
(0.013)
RACE
0.098***
-0.093***
0.136***
-0.223***
-0.055
(0.034)
(0.034)
(0.034)
(0.034)
(0.034)
Likelihood
-6,817.3
-6,935.1
-6,919.3
-7,511.8
-6,610.0

Observations
5,120
5,120
5,119
5,116
5,117

Markets
203
203
203
203
203
NOTES. The values for each dependent variable, DIVERSITY OF OPINION, COMMUNITY NEWS, MULTICULTURALISM, ADVERTISING
and OVERALL MEDIA ENVIRONMENT are 1 through 5, with 1 indicating “not satisfied” and 5 indicating “very satisfied.” Markets is the number
of television markets. Coefficient estimates for the 15 media source combination dummy variables are not reported. Estimated by weighted
maximum likelihood. Standard errors in parentheses. ***denotes significant at the one percent level. **denotes significant at the five percent level.
*denotes significant at the ten percent level.




77



Table 10. Baseline Estimates of Utility



MU
t
WTP
s.e.
DIVERSITY OF OPINION
0.160
44.83
$13.06
$1.35
COMMUNITY NEWS
0.171
50.45
$13.95
$1.35
MULTICULTURALISM
0.022
6.18
$1.82
$1.30
ADVERTISING
-0.100
23.37
$8.18
$1.33
COST
-0.012
129.7


CONSTANT
0.319
35.21



1.487
67.53


Likelihood
-1.092



Respondents
5,131



Observations




NOTES. MU is estimate of marginal utility. t is t ratio for MU estimate. WTP is estimate of
willingness to pay. s.e. is standard error of WTP estimate. λ is the estimate of the ratio of the
standard deviation of the errors in evaluating the status quo alternative to the errors in
evaluating the hypothetical alternatives. Likelihood is mean log likelihood.



78



Table 11. Probit Model Estimates of Utility



Bivariate Probit Model
Probit Model
(A vs. B & A or B
(A vs. B data)
vs. SQ data)

MU
t
WTP
s.e.
MU
t
WTP
s.e.
DIVERSITY OF OPINION
0.160
44.83
$13.06
$1.35
0.116
15.85
$11.35

COMMUNITY NEWS
0.171
50.45
$13.95
$1.35
0.145
21.58
$14.21
$1.40
MULTICULTURALISM
0.022
6.18
$1.82
$1.30
0.031
4.65
$3.04
$1.39
ADVERTISING
-0.100
23.37
$8.18
$1.33
-0.118
18.08
$11.55

COST
-0.012
129.7


-0.010
94.58


CONSTANT
0.319
35.21







1.487
67.53






Likelihood
-1.092



-0.258



Respondents
5,131







Observations








NOTES. MU is estimate of marginal utility. t is t ratio for MU estimate. WTP is estimate of willingness to pay. s.e. is standard
error of WTP estimate. λ is the estimate of the ratio of the standard deviation of the errors in evaluating the status quo
alternative to the errors in evaluating the hypothetical alternatives. Likelihood is mean log likelihood.



79



Table 12. Bivariate Probit Model Estimates of Utility

for Choice Questions 1-4 and 5-8

Questions 1-4
Questions 5-8

MU
t
WTP
s.e.
MU
t
WTP
s.e.
DIVERSITY OF OPINION
0.159
32.71
$12.67
$0.96
0.172
30.44
$14.65
$0.96
COMMUNITY NEWS
0.197
44.05
$15.77
$0.96
0.129
24.56
$10.92
$0.95
MULTICULTURALISM
0.023
4.191
$1.87
$0.93
0.008
1.585
$0.67
$0.78
ADVERTISING
-0.087
13.83
$6.91
$0.94
-0.121
20.07
$10.28
$0.94
COST
-0.013
90.44


-0.012
-7.62


CONSTANT
0.323
24.75


0.322
26.08



1.560
48.90


1.378
44.40


Likelihood
-1.072



-1.111



Respondents
5,131



5,131



Observations








NOTES. MU is estimate of marginal utility. t is t ratio for MU estimate. WTP is estimate of willingness to pay. s.e. is standard
error of WTP estimate. λ is the estimate of the ratio of the standard deviation of the errors in evaluating the status quo
alternative to the errors in evaluating the hypothetical alternatives. Likelihood is mean log likelihood.



80



Table 13. Estimates of Utility by Age



18 to 29 years
30 to 44 years

MU
t
WTP
s.e.
MU
t
WTP
s.e.
DIVERSITY OF OPINION
0.169
26.18
$12.16
$0.86
0.144
21.84
$11.53
$0.73
COMMUNITY NEWS
0.125
19.83
$8.96
$0.85
0.142
22.69
$11.43
$0.73
MULTICULTURALISM
0.036
5.45
$2.58
$0.83
0.031
4.72
$2.52
$0.71
ADVERTISING
-0.091
11.64
$6.51
$0.84
-0.068
8.45
$5.48
$0.71
COST
-0.014
81.40


-0.012
75.87


CONSTANT
0.330
20.31


0.264
15.44



1.503
42.98


1.396
36.62


Likelihood
-1.771



-1.276



Respondents
678



1,100



Observations









45 to 59 years
60 years or more

MU
t
WTP
s.e.
MU
t
WTP
s.e.
DIVERSITY OF OPINION
0.155
20.46
$12.59
$0.67
0.176
21.21
$16.39
$0.49
COMMUNITY NEWS
0.182
25.73
$14.79
$0.67
0.223
28.22
$20.78
$0.49
MULTICULTURALISM
0.025
3.28
$2.05
$0.64
-0.004
0.47


ADVERTISING
-0.116
12.82
$9.43
$0.66
-0.135
13.55
$12.59
$0.48
COST
-0.012
60.13


-0.011
47.02


CONSTANT
0.330
20.31


0.264
15.44



0.358
19.37


0.339
15.48


Likelihood
-0.900



-0.877



Respondents
1,540



1,643



Observations








NOTES. MU is estimate of marginal utility. t is t ratio for MU estimate. WTP is estimate of willingness to pay. s.e. is
standard error of WTP estimate. λ is the estimate of the ratio of the standard deviation of the errors in evaluating the
status quo alternative to the errors in evaluating the hypothetical alternatives. Likelihood is mean log likelihood.



81



Table 14. Estimates of Utility by Education



Less Than High School
High School Diploma

MU
t
WTP
s.e.
MU
t
WTP
s.e.
DIVERSITY OF OPINION
0.057
6.34
$3.98
$0.70
0.098
15.23
$7.53
$0.77
COMMUNITY NEWS
0.128
15.38
$8.91
$0.70
0.195
31.58
$14.92
$0.77
MULTICULTURALISM
-0.014
1.63


0.007
1.02
$0.51
$0.59
ADVERTISING
-0.050
4.66
$3.49
$0.68
-0.085
11.04
$6.52
$0.75
COST
-0.014
58.40


-0.013
76.83


CONSTANT
0.389
18.98


0.348
20.43



1.509
30.06


1.680
41.09


Likelihood
-1.479



-1.148



Respondents
486



1,538



Observations









Some College
College Degree

MU
t
WTP
s.e.
MU
t
WTP
s.e.
DIVERSITY OF OPINION
0.168
24.22
$13.83
$0.66
0.276
37.78
$24.91
$0.59
COMMUNITY NEWS
0.175
25.96
$14.41
$0.66
0.172
25.63
$15.52
$0.59
MULTICULTURALISM
0.029
4.02
$2.36
$0.65
0.060
8.42
$5.44
$0.59
ADVERTISING
-0.126
14.85
$10.39
$0.65
-0.104
12.21
$9.42
$0.58
COST
-0.012
67.20


-0.011
59.30


CONSTANT
0.284
15.49


0.240
12.68



1.484
35.34


1.372
30.31


Likelihood
-1.019



-0.966



Respondents
1,540



1,567



Observations








NOTES. MU is estimate of marginal utility. t is t ratio for MU estimate. WTP is estimate of willingness to pay. s.e. is
standard error of WTP estimate. λ is the estimate of the ratio of the standard deviation of the errors in evaluating the
status quo alternative to the errors in evaluating the hypothetical alternatives. Likelihood is mean log likelihood.



82


Table 15. Estimates of Utility by Income



Low Income < $25,000
$25,000 ≤ Middle Income < $75,000
$75,000 ≤ High Income

MU
t
WTP
s.e.
MU
t
WTP
s.e.
MU
t
WTP
s.e.
DIVERSITY OF OPINION
0.098
14.86
$6.19
$1.21
0.158
29.11
$11.82
$0.91
0.224
31.82
$23.11
$0.36
COMMUNITY NEWS
0.155
25.13
$9.83
$1.21
0.183
35.15
$13.74
$0.91
0.174
26.19
$17.90
$0.36
MULTICULTURALISM
0.013
1.97
$0.82
$0.99
0.034
6.20
$2.55
$0.89
0.015
2.12
$1.54
$0.35
ADVERTISING
-0.063
7.93
$3.97
$1.15
-0.097
15.08
$7.28
$0.89
-0.133
15.34
$13.67
$0.35
COST
-0.016
76.90


-0.013
92.80


-0.010
60.08


CONSTANT
0.551
42.64


0.252
17.15


0.034
1.26



1.236
35.31


1.564
48.61


1.780
30.73


Likelihood
-1.342



-1.076



-0.932



Respondents
1.058



2,241



1,820



Observations












NOTES. MU is estimate of marginal utility. t is t ratio for MU estimate. WTP is estimate of willingness to pay. s.e. is standard error of WTP estimate. λ is the estimate of
the ratio of the standard deviation of the errors in evaluating the status quo alternative to the errors in evaluating the hypothetical alternatives. Likelihood is mean log
likelihood.

83



Table 16. Estimates of Utility by Gender



Female
Male

MU
T
WTP
s.e.
MU
t
WTP
s.e.
DIVERSITY OF OPINION
0.155
31.02
$12.75
$0.90
0.165
32.53
$13.41
$1.01
COMMUNITY NEWS
0.176
36.91
$14.42
$0.90
0.166
34.49
$13.49
$1.01
MULTICULTURALISM
0.037
7.26
$3.03
$0.89
0.007
1.47
$0.61
$0.81
ADVERTISING
-0.092
15.08
$7.52
$0.89
-0.109
-18.06
$8.84
$1.00
COST
-0.012
92.02


-0.012
-91.72


CONSTANT
0.297
22.06


0.340
27.96



1.602
49.55


1.371
45.22


Likelihood
-1.113



-1.070



Respondents
2,620



2,511



Observations








NOTES. MU is estimate of marginal utility. t is t ratio for MU estimate. WTP is estimate of willingness to pay. s.e. is standard
error of WTP estimate. λ is the estimate of the ratio of the standard deviation of the errors in evaluating the status quo
alternative to the errors in evaluating the hypothetical alternatives. Likelihood is mean log likelihood.



84


Study 2 Consumer valuation of media

Table 17. Estimates of Utility by Race



White
Non White

MU
t
WTP
s.e.
MU
t
WTP
s.e.
DIVERSITY OF OPINION
0.176
37.99
$14.39
$0.98
0.127
22.67
$10.34
$0.94
COMMUNITY NEWS
0.184
41.74
$15.02
$0.98
0.143
26.87
$11.62
$0.93
MULTICULTURALISM
0.005
1.01
$0.39
$0.69
0.059
10.37
$4.78
$0.92
ADVERTISING
-0.114
20.44
$9.32
$0.97
-0.066
9.80
$5.38
$0.91
COST
-0.012
100.5


-0.012
82.31


CONSTANT
0.287
23.04


0.376
28.60



1.524
52.39


1.445
42.90


Likelihood
-1.072



-1.386



Respondents
3,826



1,305



Observations








NOTES. MU is estimate of marginal utility. t is t ratio for MU estimate. WTP is estimate of willingness to pay. s.e. is standard
error of WTP estimate. λ is the estimate of the ratio of the standard deviation of the errors in evaluating the status quo
alternative to the errors in evaluating the hypothetical alternatives. Likelihood is mean log likelihood.





85

Study 2 Consumer valuation of media

Table 18. Estimates of Utility by Gender and Race



White Male
Non-White Male

MU
t
WTP
s.e.
MU
T
WTP
s.e.
DIVERSITY OF OPINION
0.171
25.55
$13.66
$0.72
0.157
19.99
$13.06
$0.71
COMMUNITY NEWS
0.186
29.20
$14.84
$0.72
0.127
17.20
$10.64
$0.71
MULTICULTURALISM
-0.010
1.449


0.042
5.25
$3.48
$0.70
ADVERTISING
-0.122
15.10
$9.75
$0.71
-0.077
8.37
$6.44
$0.70
COST
-0.013
70.97


-0.011
56.65


CONSTANT
0.298
17.25


0.412
24.55
34.42


1.436
35.16


1.275
28.02


Likelihood
-0.959



-1.390



Respondents
1,871



620



Observations









White Female
Non-White Female

MU
t
WTP
s.e.
MU
T
WTP
s.e.
DIVERSITY OF OPINION
0.182
28.15
$15.10
$0.67
0.098
12.15
$7.77
$0.61
COMMUNITY NEWS
0.183
29.86
$15.21
$0.67
0.160
20.85
$12.68
$0.61
MULTICULTURALISM
0.018
2.78
$1.52
$0.64
0.078
9.58
$6.16
$0.61
ADVERTISING
-0.106
13.73
$8.83
$0.66
-0.056
5.63
$4.42
$0.59
COST
-0.012
71.07


-0.013
59.42


CONSTANT
0.276
15.34


0.338
16.49



1.613
38.10


1.609
32.12


Likelihood
-1.019



-1.380



Respondents
1,995



665



Observations








NOTES. MU is estimate of marginal utility. t is t ratio for MU estimate. WTP is estimate of willingness to pay. s.e. is
standard error of WTP estimate. λ is the estimate of the ratio of the standard deviation of the errors in evaluating the
status quo alternative to the errors in evaluating the hypothetical alternatives. Likelihood is mean log likelihood.




86

Study 2 Consumer valuation of media

Table 19. Estimates of Utility with Non-Linear Preferences



MU
t
WTP
s.e.
MEDIUM DIVERSITY OF OPINION
0.240
33.85
$18.86
$1.34
HIGH DIVERSITY OF OPINION
0.322
44.50
$25.28
$1.33
MEDIUM COMMUNITY NEWS
0.261
35.35
$20.54
$1.34
HIGH COMMUNITY NEWS
0.343
50.20
$26.93
$1.33
MEDIUM MULTICULTURALISM
0.051
7.25
$4.03
$1.31
HIGH MULTICULTURALISM
0.050
6.93
$3.96
$1.29
MEDIUM ADVERTISING
-0.036
5.84
$2.85
$1.26
HIGH ADVERTISING
-0.257
28.07
$20.22
$1.32
COST
-0.013
130.7


CONSTANT
0.289
30.21



1.569
68.54


Likelihood
-1.087



Respondents
5,131



Observations




NOTES. MU is estimate of marginal utility. t is t ratio for MU estimate. WTP is estimate of willingness
to pay. s.e. is standard error of WTP estimate. λ is the estimate of the ratio of the standard deviation of the
errors in evaluating the status quo alternative to the errors in evaluating the hypothetical alternatives.
Likelihood is mean log likelihood.





87

Study 2 Consumer valuation of media

Table 20. Ordered Probit Estimates of Media Environment Features

and Television Market Structure

DIVERSITY OF
COMMUNITY
MULTI-
ADVERTISING
OPINION
NEWS
CULTURALISM
TV STATIONS
-0.009
0.024*
-0.035***
-0.019
(0.013)
(0.013)
(0.013)
(0.013)
TV VOICES
0.052***
0.009
0.081***
0.052***
(0.018)
(0.018)
(0.018)
(0.018)
TV STATIONS  TV VOICES
-0.0007
0.001***
-0.0007
-0.0007
(0.0005)
(0.0005)
(0.0005)
(0.0005)
AGE
0.053***
0.096***
0.031***
0.112***
(0.010)
(0.010)
(0.010)
(0.010)
EDUC
0.123***
0.084***
0.125***
0.111***
(0.018)
(0.018)
(0.017)
(0.018)
GENDER
-0.048
0.114***
0.024
0.023
(0.032)
(0.032)
(0.032)
(0.032)
INCOME
0.056***
0.004
-0.031**
0.033**
(0.014)
(0.014)
(0.014)
(0.014)
RACE
0.083**
-0.128***
-0.053
0.179***
(0.036)
(0.036)
(0.036)
(0.036)
β΄x/∂TV VOICES
0.043**
-0.010
0.071***
0.044**

(0.016)
(0.016)
(0.016)
(0.016)
Likelihood
-4,822.7
-5,189.6
-5,186.2
-4,951.4

Observations
5,123
5,123
5,123
5,123

Markets
203
203
203
203
NOTES. The values for each dependent variable, DIVERSITY OF OPINION, COMMUNITY NEWS, MULTICULTURALISM and
ADVERTISING are low, medium and high. ∂β΄x/∂VOICES is the effect of TV VOICES on the index function, β΄x, where
x = [1, Xj, Zi ], 1 is a vector of ones, and β = [α, δ, γ]. Markets is the number of television markets. Coefficient estimates for the 15 media
source combination dummy variables are not reported. Estimated by weighted maximum likelihood. Standard errors in parentheses.
***denotes significant at the one percent level. **denotes significant at the five percent level. *denotes significant at the ten percent level.




88

Study 2 Consumer valuation of media

Table 21. Mean Change in Predicted Probabilities



Diversity of Opinion
Multiculturalism
Advertising
ΔPL
0.0106
0.0260
0.0089
ΔPM
0.0028
-0.0081
0.0074
ΔPH
-0.0134
-0.0179
-0.0163
NOTES. Sample means are calculated from each individual respondent’s predicted
probabilities.




89

Study 2 Consumer valuation of media

Table 22. Changes in Consumer Welfare From a Merger between

Two Television Stations

Market:
Pop.
Diversity
Multi-
Advertising
Total
Diversity
Multi-
Advertising
Total
Number
share
of opinion culturalism


of opinion
culturalism


of TV
(%)
Average consumer welfare per month
Annual aggregate consumer welfare in market
Stations

(dollars per month)
(annual dollars in millions)
5
5.0
-$0.34
-$0.12
$0.34
-$0.11
-$18.10
-$6.18
$18.31
-$5.98
6
6.1
-$0.33
-$0.11
$0.34
-$0.11
-$21.94
-$7.53
$22.14
-$7.33
7
9.1
-$0.33
-$0.11
$0.34
-$0.11
-$32.28
-$11.23
$33.05
-$10.46
8
8.1
-$0.32
-$0.11
$0.33
-$0.10
-$27.91
-$9.82
$29.24
-$8.48
9
9.5
-$0.31
-$0.11
$0.33
-$0.09
-$31.83
-$11.34
$33.72
-$9.45
10
5.6
-$0.30
-$0.11
$0.32
-$0.09
-$18.38
-$6.59
$19.55
-$5.42
11
9.9
-$0.30
-$0.11
$0.32
-$0.09
-$31.71
-$11.51
$33.80
-$9.42
12
6.9
-$0.30
-$0.11
$0.31
-$0.09
-$22.00
-$7.99
$23.02
-$6.98
13
2.4
-$0.28
-$0.10
$0.31
-$0.08
-$7.20
-$2.65
$7.85
-$1.99
14
9.3
-$0.27
-$0.10
$0.30
-$0.06
-$26.93
-$9.94
$30.47
-$6.40
15
3.0
-$0.27
-$0.10
$0.29
-$0.07
-$8.78
-$3.27
$9.63
-$2.42
16
7.9
-$0.25
-$0.10
$0.29
-$0.06
-$21.74
-$8.33
$25.12
-$4.95
17
7.2
-$0.25
-$0.10
$0.29
-$0.07
-$19.80
-$7.50
$22.24
-$5.06
18
4.3
-$0.25
-$0.10
$0.28
-$0.07
-$11.69
-$4.52
$12.97
-$3.23
19
2.6
-$0.24
-$0.10
$0.28
-$0.06
-$6.83
-$2.75
$7.98
-$1.60
20
3.2
-$0.23
-$0.09
$0.28
-$0.04
-$7.99
-$3.10
$9.71
-$1.39
NOTES. The merger is a one-unit reduction in the number of independent television voices in the market, all other things held constant.
There are 90,193,905 population households in markets from five to 20 television stations (FCC, 2011). Pop. share is the number of
population households in the market divided by population households. For example, Pop. share for the five-station market is
4,469,100/90,193,905 = 5.0 %. Pop. share for the 20-station market is 2,924,767/90,193,905 = 3.2 %.




90

Study 2 Consumer valuation of media

Figure 1. Your Actual Media Environment Example

In summary, your answers indicate that you listen to the radio, watch television, and use the
Internet to get information on news and current affairs. Your overall media environment also
has the five features described in Table 1 below.

Table 1. Your actual media environment

http://qcsurveys.knowledgenetworks.com/SPSSMR/ImageCache/ImageCache.aspx?Project=S14328&File=en-US/levels_1.htm">Click here to review a summary of the levels of all the features (Table 2).
To see the description of an individual feature, place your cursor over that feature

Feature


Level

Description

Your media environment provides information on
Diversity of
Medium
news and current affairs from a few different
opinion
viewpoints.
Community
Your media environment provides some information
Medium
news
on community news and events.
Your media environment provides very little or no
Multiculturalism
Low
information that reflects the interests of women and
minorities.
The amount of space and/or time devoted to
Advertising
High
advertising in your overall media environment is
annoying.
The total of your monthly subscriptions to all of your
$135
Cost
media sources, plus any contributions to public radio
per month or public TV stations.



















91

Study 2 Consumer valuation of media

Figure 2. Choice Scenario Example


1. C [Fix
onsid up whe
er the f n we
ollowi ha
ng tve
w f
o u
mll sample …
edia envir
Know
onment optledge
ion
Network
s, A and B
s
, whIinc. (
ch p KN
rovi ) a
de dministere
news and d the
current
affairs from your media sources: radio, television, and the Internet. The two options differ by the
online
l surve
evels o y
f . KN
diver pa
sity ne
of l membe
opinio
rs a
n, comre
m dra
unit wn b
y
y random di
news, multicultu g
rait
li dialing
sm, adv o
ertf list
isinged and unli
, and by cos sted
t.

telephone
For t house
his firs holds
t que , wi
sti
th a suc
on, we hig c
hl e
i ss r
ght tate of
he dif fabout 45 t
erences i
o 50 pe
n the levelrcent.
s of th F
e fi or
v in
e f c
ea e
t nti
ure ve
s i , pa
n r nel
ed. For
some of these five features, there may be no difference. Check the media environment option you
would prefer.
members are rewarded with points for participating in surveys, which can be converted to cash

http://qcsurveys.knowledgenetworks.com/SPSSMR/ImageCache/ImageCache.aspx?Project=S13229&File=en-US/table2.htm&_1">Click here to review a summary of the levels of all the features.
or various non-cash prizes. KN contacted a gross sample of 799 panel members on January
To see the description of an individual feature, place your cursor over that feature
24, 2003 informing them about the Internet service choice experiment. By February 12, 2003,

Option A

Option B

575 c
Div ompl
ersity ete que
of opini sti
on onna

ires were obtained w
Low it
h a effective unit response rate of
Me 32.4 to 36
dium
Community news
Medium
Low
percent (i.e., 575/79945 to 50 percent). 209 of the 575 questionnaires were excluded by us
Multiculturalism
Low
Low
from this analysis because they had been randomly assigned an additional Internet access
Advertising
High
Medium
attribute
Cost
as part of another study. Of the 366 c
$25 per m om
onthplete

d questionnaires remaining
$45 per f
m or us
onth e in
Option A is less expensive and has more
Option B has less advertising
thi
s study, 325 respondeints answe
nformati
red
on on all eig
comm ht
unitInte
y
rnet acc
news and e
evss choic
ents
e que
and m stions for
ore divers a
it n it
y of em
opinion
response r
Select t
ate of
he opti 88.8 p
on you ercent. The median compl

etion time for each mail questionna
ire was
prefer
I prefer option A
I prefer option B
about 19 m

inutes. ]
2. Since you currently have a media environment at home, we also ask if you would actually switch
t A se
o th
lec
e m
ti
edi on of
a envir sample d
onment, Bemog
, y
raphi
ou hav
c
e c s, alon
hosen. g
C with s
onsider im
t il
he a
f r da
eaturta f
es rom the U
of your act .S
ual . C
m ensus
edia
environment. Would you switch to the option B you chose previously?
Bur
eau (2003), are presented in Table 2. The sample covers 44 states. The typical respondent
http://qcsurveys.knowledgenetworks.com/SPSSMR/ImageCache/ImageCache.aspx?Project=S13229&File=en-US/table2.htm&_2">Click here to review a summary of the levels of all the features.
To see the description of an individual feature, place your cursor over that feature.
is a white, 50 year old male with either some college (no degree), who resides in a household
with 1.7 ot

her members. He was

Y

empl
our m oy
edied last m
a envir
onth at a
onment
location outside of

O

the home
ption B
, and
Diversity of opinion
Medium
Medium
has average annual household income $65,095. The sample is similar to the U.S. population
Community news
Medium
Low
with re
Multi spec
cult t t
ural o
i g
sme
ographic coverage, responde
Low nt's age, gender, employment status and
Low

Advertising
High
Medium

Cost
$135 per month
$45 per month
Select the option you



prefer


I prefer option A
I prefer option B


92

Study 2 Consumer valuation of media

Figure 3. Change in Average Consumer Welfare Per Month From

a Merger between Two Television Stations
0.40
0.30
0.20
Diversity of opinion
Multiculturalism
0.10
Advertising
Total
0.00
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
-0.10
-0.20
-0.30
-0.40

NOTES. Vertical axis is dollars per month and horizontal axis is number of television stations. The merger is a one-unit
reduction in the number of independent television voices in the market, all other things held constant.

SOURCE. Table 22.


93

Note: We are currently transitioning our documents into web compatible formats for easier reading. We have done our best to supply this content to you in a presentable form, but there may be some formatting issues while we improve the technology. The original version of the document is available as a PDF, Word Document, or as plain text.

close
FCC

You are leaving the FCC website

You are about to leave the FCC website and visit a third-party, non-governmental website that the FCC does not maintain or control. The FCC does not endorse any product or service, and is not responsible for, nor can it guarantee the validity or timeliness of the content on the page you are about to visit. Additionally, the privacy policies of this third-party page may differ from those of the FCC.