This paper was presented at the Association for Education in Journalism and
Mass Communication in San Antonio, Texas August 2005.
If you have questions about this paper, please contact the author
directly. If you have questions about the archives, email
rakyat [ at ] eparker.org. For an explanation of the subject line,
send email to
[log in to unmask] with just the four words, "get help info aejmc," in the
body (drop the "").
(Feb 2006)
Thank you.
Elliott Parker
====================================================================
Web Repertoires and Audience Concentration
Jungsu Yim
Submitted to the Mass Communication and Society Division of AEJMC
2005 Convention.
Author: Jungsu Yim, Ph.D.
(Assistant Professor, Department of Communication & Media,
Seoul Women's university)
Address: Prof. Jungsu Yim
Department of Communication & Media, School of Social Science,
Seoul Women's University.
126, Gongnuen-dong, Nowon-gu
Seoul, Korea, 139-174.
82-2-970-5586
[log in to unmask]
Web Repertoires and Audience Concentration
Abstract
This study focuses on presenting the evidence of an association
between Web repertoires and audience concentration that has been
hypothetically suggested in some past studies. The result is that Web
repertoires formed in an individual respondent level lead to audience
concentration in an aggregate level. The result implies that
television audiences in the multi-item media environment will face
the similar environment to the Web.
Web Repertoires and Audience Concentration
Many studies on audience selection in a multi-item option environment
dealt with television cases (e.g., Heeter, 1985; Webster & Lin, 2002;
Youn, 1994). 'Multi-item option' in the digital age means the bigger
number of over 100, whereas it meant 20-50 channels in the analogue
age. The Internet offers the much bigger number of items, literally
millions of Web sites. Selecting one out of the limited number of
items might result in different audience behavioral patterns from
selecting one out of several items available.
As the number of television channels increased to 20-50, it became
useful for viewers to be guided by TV Guide for selecting programs.
As the number of channels exceeded 100, searching a television
guidebook for some interesting programs became a cumbersome and
time-consuming task. It is also hard for viewers to schedule a TV
viewing because they do not have enough information about all
programs available to them. Television viewers want to spend less
time and effort in selecting television programs.
The notion of channel repertoires can be suggested as an audience
behavioral pattern explaining how audiences select channels or
programs in an economic way responding item abundance (Heeter, 1985;
Heeter & Greenberg, 1988; Ferguson & Perse, 1993; Webster & Phalen,
1997). Audience concentration can be defined as the extent to which
audience shares are unequally distributed across items of content
within a given medium at particular points in time (Yim, 2003).
Webster & Phalen (1997) argued that as the number of channels
increase, audiences' preferences concentrate on specific channels. By
employing the notion of audience concentration in the Web studies,
Webster & Lin (2002) and Yim (2003) argued that only a small
percentage of Web sites are visited by Internet users,
The hypothetical relationship between channel repertoires and
audience concentration was mentioned by Yim (2002). However, no
empirical study has been done so far. This study attempts to show
empirically how Web repertoires lead to audience concentration, and
ultimately to present audiences' responses to the media environment
that offers item diversity. The Web offers the largest items out of
all media used by humankind.
The first part of this study presents that the notion of channel
repertoires suggested to explain television channel selection can be
applied to Web audience behavior. The evidence of Web repertoires in
the everyday Web uses might show how audiences would select items in
the digital media environment that offers a large number of items. In
addition, this study examines the difference of Web repertoires
between heavy user and light users.
The second part examines the extent to which Web users have common
items in their own Web repertoires, and the degree to which items
included in many Web users' repertories concentrate on a small
percentage of the Web universe. This examination is expected to show
if Web repertoires formed in an individual respondent level lead to
audience concentration in an aggregate level. Channel repertoires and
audience concentration have been considered separate concepts, and
the relationship between them remains hypothetical so far.
Literature Review
Channel repertoires
Channel repertoires mean subgroups of channels or items frequently
used by individual audience members. Heeter (1985), Heeter &
Greenberg (1988) and Ferguson & Perse (1993) established the notion
of channel repertoires to explain audiences' responses to a
multi-channel environment provided by cable television. These studies
showed that cable television viewers watched some specific channels
intensively, rather than used all available channels randomly.
Webster & Phalen (1997) employed channel repertoires to explain
audience behavioral patterns related to channel selection. Yim (2002)
also employed channel repertoires as an important concept showing a
mechanism that audiences adapt to a multi-channel environment.
Channel repertoires do not directly explain audience concentration
because it concerns audience behavior in an individual level.
Nonetheless, if some channels are included in most audience members'
repertoires, and some others in much fewer audience members'
repertoires, audience concentration can occur. Yim (2003) found that
a positive relationship between item abundance and the degree of
audience concentration. It means that a vast majority of items will
be overlooked by most people especially in the Web.
In general, a channel in a narrow sense is the term used in broadcast
television networks, basic cable television networks, and radio. This
study attempts to apply the notion to the Web. The term "channels" is
rarely used in the cases of books, magazines, or Web sites.
Accordingly, I suggest that the alternative term "item" be used. This
study operationalizes "channels available" or "items available" as
the choices available within a medium to audience members at one time
point. That is, one can select only one channel from channel choices
available, as defined in this study, at one time point. For example,
a reader selects a magazine title out of many available titles,
rather than a magazine publisher out of many publishers. Likewise,
generally one cannot select two different television programs at one
time point. Hence, a television channel or a radio channel, not an
individual program, selected at one time point is considered a
"channel or item" (Yim, 2003).
Neuendorf, Atkin & Jeffres (2001) examined how frequently cable
television viewers use channel repertories by using three levels of
repertoires. According to them, the primary repertoire is defined as
the number of channels used at least one time every day, the
secondary repertoire as the number of channels used at least one time
every week, and the tertiary repertoire as the number of channels
used less than one time per week. Although these definitions of
channel repertoires have not yet been applied to Web studies, their
idea can be useful in dealing with Web repertoires.
Ferguson & Perse (2000) suggested two methods for measuring Web
repertoires: First, survey respondents select Web sites that they
visited from the 100 top-ranked site list by unique audience. This
method reduces errors by respondents' incorrect memories by
presenting 100 Web sites, but overestimates the ratio of Web
repertoires to all Web sites used. The second is to have respondents
keep the 3-day dairy recording their Web uses. This method
underestimates the repertoire size by an inadvertent omission.
This study suggests the notion of "the relative Web repertoire size",
which is defined as the ratio of the unduplicated number of the Web
sites visited for ? days to the unduplicated number of the Web sites
visited for a given days. For example, a user visited 20 Web sites
for 5 days, and visited 5 Web sites at least one time everyday, and 8
Web sites at least one time on any 4 days out of 5 days. In this
case, the relative Web repertoire size is 25% (= 5/20) if it is
defined as the ratio of the unduplicated number of the Web sites
visited everyday (5 days) to the unduplicated number of the Web sites
visited for 5 days. The relative Web repertoire size is 40% (= 8/20)
if it is defined as the ratio of the unduplicated number of the Web
sites visited on any 4 days to the unduplicated number of the Web
sites visited for 5 days. The relative repertoire size indicates the
degree of the dependence on Web repertoires in Web uses.
This study poses the following research question by introducing the
notion of channel repertoires to Web studies to explain how Web users
select a Web site out of millions of items.
Research Question1: How big repertoire do Web users have in their
everyday uses?
Lin (1994) and Perse, Ferguson & McLeod (1994) argued that heavy
cable television viewers tend to have bigger repertories. The second
research question concerns if there are any differences in the Web
repertoire size between heavy and light users.
In research question 2, Web repertoire size is defined in two ways.
One is repertoire size expressed in absolute numbers, and the other
is the relative repertoire size, which is the ratio of the repertoire
size in absolute numbers to all Web sites visited in a given time
period. Comparison of relative Web repertoire size between heavy and
light users is expected to reveal which group of users more relies on
their own repertoires.
Research Question2: Is there any difference in the Web repertoire
size between heavy and light users?
Channel repertoires and audience concentration
The second part of this study examines the relationship between Web
repertoires and audience concentration. The distribution of Web
audience shares expressed in the cumulative audience size shows how
audiences choose channels or items in a given time period (usually a
month for the Web and a week for television).
It is pretty rare that audience concentration is used as a separate
concept rather than an indicator of economic concentration in the
media. Substituting for income or ownership data, audience size has
been used to analyze economic concentration in the cable
multiple-system operator market (Chan-Olmsted, 1996), local
newspapers (Picard, 1988), film industry and television industry
(Waterman, 1991). This can be justified because the media industry is
the one producing audiences, rather than media content (Owen &
Wildman, 1992, p.3). However, audience concentration is not always
identical to economic concentration.
Consider a world in which 100 channels are available. Suppose that
each channel is owned by a separate firm and offers a distinct
program format. Initially, it appears that this world is
characterized by both diversity of ownership and diversity of
content. But, imagine that one channel dominates 99% of all viewing.
By introducing audience behavior into the equation it becomes obvious
that one channel, one owner, and one type of content has a monopoly
on the marketplace of ideas. Moreover, that one owner may have 99% of
market revenues as well. Although this example is extreme and
oversimplified, it highlights the importance of studying audience
concentration. As was the case in our hypothetical world, introducing
audience behavior provides insight into the nature of media diversity
and economic power (Yim, 2003, p.115).
Although less attention was paid to audience concentration, similar
notions have been mentioned by others. Napoli (1999) introduced the
construct, 'exposure diversity', referring to how many different
items audiences select. This idea is distinguished from source
diversity and content diversity (McQuail, 1992; Napoli, 1999, 2001).
In other words, audiences can be concentrated on some particular
items regardless of source or content diversity. Adamic & Huberman
(1999), Webster & Lin (2002) and Yim (2003) presented unequal
distribution of Web shares across Web sites available to audiences.
Audiences' channel selection is not random. Thus, some channels or
items are in most audience members' repertoires. For example, most
viewers include broadcast networks (NBC, ABC, CBS, FOX, etc.) in
their repertoires (Heeter & Greenberg, 1988). Ferguson & Perse (2000)
argued that this is also the case in the Web. Some Web sites are
visited by most users, whereas most Web sites are overlooked. Thus
channel repertoires might lead to audience concentration. Research
question 3 deals with the hypothetically accepted relationship
between channel repertoires and audience concentration.
Research Question3: Do Web repertoires in an individual respondent
level lead to audience concentration in an aggregate level?
Methods
Data collection
A survey of 162 college students from the three U.S. states
(Illinois, Michigan, Indiana) was conducted in April and May 2001 by
email. Four colleges each state were selected and five majors each
school were selected by the multi-stage cluster sampling method. Then
the survey was conducted to students who agreed to participate in the
survey every 6 days four times starting from Friday by the
diary-keeping method. What Web sites they visited in 8p.m.-1a.m and
how long they stayed on the Web during the survey days (4 days).
The unit of a Web site basically followed Adamic & Huberman (1999)
considering 'b1.com' and 'b2.com' as the different ones, and
'b.com/c1' and 'b.com/c2' as the same one. However, 'a1.b.com' and
'a2.b.com' were considered as the different ones in this study unlike
in Adamic & Huberman(1999).
Analysis Methods
A method of measuring Web repertoire must be decided to see if Web
users have their own repertoires in using the Web in Research
Question1. Because channel repertoire is based on repeat viewing, it
is a similar notion to audience loyalty. Channel repertoire is an
indicator of audience loyalty, which is usually defined as repeat
viewing, but they are not synonymous. If we look at the methods of
measuring the two concepts, the differences would be clear.
Barwise(1986), Webster & Wang(1992) and Sherman(1995) defined repeat
viewing as an average frequency that viewers watched episodes of a TV
program. For audience loyalty on the Web, Nielsen//NetRatings Inc.
uses "visits per person" which is defined as total visits divided by
unique audience. "Visits per person" also means an average frequency
of visits. If an audience member visits a Web site very frequently
and others visit it only one time, the average frequency of visits is
higher than 1. Thus, "visits per person" represents an audience
behavioral pattern of a Web site in an aggregate level. Unlike repeat
visits, Web repertoires concern an audience behavior in an individual level.
This study introduces the notion of relative Web repertories, which
measures how large portion of the Web sites an audience member
visited is one's Web repertoire. Thus, there is no guarantee that Web
sites included in more people's repertoires have higher repeat visits
(= visits per person) than ones included in fewer people's
repertoires. Every unintentional visit is counted for repeat visits,
but it may not be counted for Web repertoires.
The list of Web sites that respondents had visited during the
four-day-survey period was made on the respondent base. How many Web
sites respondents visited is counted. Two or more visits of a Web
site in a day were counted 1. Then the percentage of the number of
Web sites that an audience member visited in all four-survey-days out
of the total number of Web sites that one visited during the survey
period is calculated. It is the relative Web repertoire as expressed here.
Relative repertoire size = (The number of Web sites that an audience
member visited in all four-survey-days * 100) / The total number of
Web sites that one visited during the survey period.
The percentage of the number of Web sites that an audience member
visited three or two times out of the total number of Web sites that
one visited during the survey period is also calculated. First,
comparisons among different degrees of repertoires (four, three and
two visits) are presented. Next, to get convincing results this study
uses only the case for four time visits to analyze the relationship
between Web repertoires and audience concentration. The high relative
repertoire size means that the percentage of Web sites included in
one's repertoire out of the total number of Web sites that one visited is high.
For Research Question 2, the Pearson correlations between the
absolute number of Web sites that a respondent visited and the
visiting time per person, and between the relative number of Web
sites and the visiting time per person are calculated respectively.
For Research Question 3, the cumulative distribution of unique
audience on the cumulative number of Web sites that respondents
visited four times (that is, Web sites included in their repertoires)
is presented. That can be expressed by the Lorenz curve. The Lorenz
curve plots cumulative percentages of the industry size against
cumulative percentages of firms, starting with the smallest. The
Lorenz curve expresses concentration as the ratio of the area between
the Lorenz curve and a 45_ line to the area of the triangle under the
45_ line, which is the Gini coefficient[1], a measure of the relative
size of firms. The Gini coefficient has a maximum of 1 and a minimum
of 0 (Aitchison & Brown, 1963; Gini, 1936; Hay & Morris, 1979;
Lorenz, 1905; Neuman, 1991). Gini coefficient of 1 indicates perfect
inequality, while o indicates perfect equality.
The advantage of this index is that it reflects all firms in the
industry and the overall concentration pattern of the industry,
unlike concentration ratios (Cowell, 1977). Moreover, Lorenz curves
make graphical comparison across media possible.
Results
The relative size of Web repertoires
<Table1> shows the percentages of respondents each relative
repertoire interval. 21 respondents (12.96%) have the relative
repertoire size of 10-20% if a Web site is said to be in one's
repertoire when one visits it at least one time each survey day.
64.82% of respondents fall into the relative size of 20-50% in the
same condition (4 visits). 51.23% of respondents have the relative
repertoire size of 60-80% if a Web site is said to be in one's
repertoire when one visits it two times during the survey period. No
one falls into the relative size of 0-30% in the same condition (2
visits). Hence, the results show that most respondents have Web
repertoires although the size varies according to individuals.
----------------------------------------------
Table1
-----------------------------------------------
The relationship between the duration of use and repertoires
<Table 2> presents the relationship between time per person and the
repertoire size expressed in absolute numbers, and <Table 3> presents
the relationship between time per person and the relative repertoire
size. When the repertoire in <Table 2> and <Table 3> is defined as a
group of Web sites visited by an audience member at least one time
each survey day, the correlation coefficients are .673 and -.388
respectively. The relationship between time per person and the
relative repertoire size is negative whereas the relationship between
time per person and the absolute repertoire size is positive. The
results in <Table 2> and <Table 3> mean that respondents who stay
longer on the Web tend to have a bigger repertoire and depend less on
the Web repertoire.
------------------------------------------------
Table 2
-----------------------------
------------------------------------------------
Table 3
-----------------------------
Web repertoires and audience concentration
<Table 4> reveals that the total number of unique Web sites in
respondents' repertoires is 349, when the repertoire is defined as a
group of Web sites visited by an audience member at least one time
each survey day. <Table 5> shows that a majority of respondents have
2-4 Web sites in their own repertoires. That means that they visited
2-4 Web sites almost every day.
------------------------------------------------
Table 4
-----------------------------------------------
------------------------------------------------
Table 5
-----------------------------------------------
<Table 6> presents that the first-ranked site, www.yahoo.com is
included in 38 respondents' repertoires, and search.yahoo.com and
www.hotmail.com are included in 31 respondents' repertoires. 64 out
of 349 Web sites are yahoo-related, 34 are MSN-related, 15 are
AOL-related. 204 out of 349 Web sites are included in only one
respondent's repertoire.
------------------------------------------------
Table 6
-----------------------------------------------
<Figure 1> is the Lorenz curve of <Table 6>. The Lorenz curve
concaves to the right bottom corner. That means that respondents'
repertoires concentrate on the high-ranked Web sites located on the
right of X- axis. According to the calculation for Gini coefficients
explained earlier, the Gini coefficient of the Lorenz curve is 0.417.
------------------------------------------------
Figure 1
-----------------------------------------------
Discussion
Employing the Notion of channel repertoires to the Web
Since Ferguson & Perse (2000) suggested the introduction of channel
repertoires into Web studies, not a few studies actually have been
done. This study dealt with Web repertoires by examining college
students, who are a relatively homogeneous group. The results
revealed that Internet users do not choose a Web site randomly by
having their own Web repertoires. Millions of Websites are available
to audiences. If their choice is at random, it is not possible to
choose a specific Web site more frequently. The empirical evidence of
Web repertoires implies that audiences will respond to item abundance
provided by future media in the same way.
The degree of use and repertoires
Just as Cable television viewers do, Internet users who stay on the
Web longer have a bigger repertoire in an absolute number. Meanwhile,
time per person and the relative repertoire size are negatively
correlated. Light Web users tend to visit only some specific Web
sites repeatedly. Some light Web users never visit other Web sites
than their repertoires. They tend to spend less time in seeking
something new, because searching new items or channels seem to them
wasteful. A respondent who visited a relatively small number of Web
sites actually visited the same sites each day during the survey period.
This result implies that in the multi-channel environment
light-media-users tend to use channels or items included in their own
channel repertoires, and they are likely reluctant to try to use
others. For light-media-users, it is not economic to search for an
interesting program out of hundreds of channels whenever they want to
use the media.
A mechanism of audience concentration
The association between channel repertoires and audience
concentration has been also mentioned by Ehrenberg (1988), and
Ehrenberg & Wakshlag (1987) who suggested the double jeopardy effect
indicating the correlation between repeat viewing and audience size.
The double jeopardy effect is the phenomenon that channels with low
audience loyalty defined as repeat viewing have small audience size.
However, the double jeopardy effect did not present mechanism that
repeat viewing leads to audience concentration. This study
distinguished channel repertoires from repeat viewing, and showed how
repertoires are connected to audience concentration.
The fact that search engines and portal sites are high-ranked in
unique audience size shows that Internet users need packaged
information because they have trouble in finding some items that they
want in the sea of the Internet. Some search engines and portal sites
such as Yahoo.com, MSN.com, altavista.com and aol.com became the hubs
of Web networking. A Web site can be connected to audiences easier
when it is networked to the hubs. A great percentage of Web sites are
not connected to the hubs, then audiences do not actually have the
way to reach such sites. For this reason, yahoo.com, the first
successful search engine, took the top, and was followed by other
yahoo-related Web sites and MSN-related Web sites.
The results of this study imply that audiences have repertoires
in choosing items or channels in the multi-item media environment.
Digitalized television offers a good number of channels, which will
increase in the future. Television audiences will face the similar
environment to the Web. Because they do not have enough information
about all channels, it is highly likely that they rely on the fame of
channels and their own channel repertoires. The more abundant items
are offered, the higher audience concentration on a smaller
percentage of the total items available is expected.
Some advertisers still want to reach a broad audience, and need the
advertising vehicles for the purpose. Small media firms cannot
satisfy such requirements. As a result, mergers & acquisitions (M&A)
in media industries become usual. Such Web firms as AOL, Yahoo and
MSN lead the Internet M&A market. The television industry must not be
so different.
Suggestions
This study focuses on presenting the evidence of an
association between Web repertoires and audience concentration that
has been hypothetically suggested in some past studies. This study
has some limitations. First, this study did not deal with the factors
that affect audience preferences, which should be studied in the
following works. Second, the sample needs to include a broad audience
because the results of this study are based on the examination of
college students. However, we cannot assume in this study that
college students have pretty different behavioral patterns in using
the Web from others. Third, a method of measuring Web repertoires
should be more improved by conceptualizing Web repertoires in
relation to audience behavior.
<Table 1> The relative Web repertoire size
The relative Web repertoire size (%)
Web repertoire levels
4 visits 3 visits 2 visits
(respondents/(%)) (respondents/(%)) (respondents/(%))
90-100
80-90
70-80
60-70
50-60
40-50
30-40
20-30
10-20
0-10
Total
10 / 6.17
1 / 0.62
3 / 1.85
10 / 6.17
10 / 6.17
30 / 18.52
41 / 25.31
34 / 20.99
21 / 12.96
2 / 1.23
162 / 100
15 / 9.26
5 / 3.09
16 / 9.88
23 / 14.20
31 / 19.14
40 / 24.69
24 / 14.81
6 / 3.70
2 / 1.23
0 / 0
162 / 100
23 / 14.20
17 / 10.49
43 / 26.54
40 / 24.69
19 / 11.73
17 / 10.49
3 / 1.85
0 / 0
0 / 0
0 / 0
162 / 100
<Table 2> The Correlation coefficients between time spent using and
the absolute repertoire size
The absolute repertoire size
Repertoire levels 4 visits 3 visits 2 visits
Average time spent
using .673 .783 .830
* p < .01 (two-tailed)
<Table 3> The Correlation coefficients between time spent using and
the relative repertoire size
The relative repertoire size
Repertoire levels 4 visits 3 visits 2 visits
Average time spent using -.388 -.440 -.301
* p < .01 (two-tailed).
<Table 4> The total number of unique Web sites in respondents'
repertoires (4 visits)
The total number of Web sites
Minimum
Maximum
Mean
S.D.
349
1
38
2.7
4.5353
<Table 5> The absolute repertoire size (4 visits)
Minimum
Maximum
Median
Mode
Mean
1
35
4
3
5.86
<Table 6> The ranking of Web sites in terms of the repertoire size
(4 visits) and the number of respondents
Ranking
Web sites (4 visits)
Respondents
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
www.yahoo.com
search.yahoo.com
www.hotmail.net
www.msn.com
search.msn.com
www.cnn.com
www.amazon.com
mail.yahoo.com
www.napster.com
www.ebay.com
www.geocities.com
www.lycos.com
gooles.yahoo.com
www.aol.com
www.goto.com
dir.yahoo.com
home.netscape.com
38
31
31
29
27
25
23
21
18
17
14
11
11
10
9
8
8
18-24
25-30
31-46
47-62
62-83
84-145
146-349
7
6
5
4
3
2
1
<Figure 1> The Lorenz curve of Web repertoires
[--- ??? Graphic Goes Here ---]
References
Adamic, L.A., & Huberman, B.A. (1999). The nature of markets in the
World Wide Web. Palo Alto, CA: Xerox Palo Alto Research Center. Mimeo.
Aitchison, J., & Brown, J. A. C. (1963). The lognormal distribution.
Cambridge, UK: Cambridge University Press.
Barwise, T.P. (1986). Repeat viewing of prime time TV series. Journal
of Advertising Research, 26(4), 9-14.
Chan-Olmsted, S. M. (1996). Market competition for cable television:
reexamining its horizontal mergers and industry concentration.
Journal of Media Economics, 9(2), 25-41.
Cowell, F. A. (1977). Measuring inequality. New York: Halsted Press.
Ehrenberg, A.S.C. (1988). Repeat-buying: Facts, theory and
applications. New York: Oxford University Press.
Ehrenberg, A.S.C. , & Wakshlag, J. (1987). Repeat-viewing with people
meters. Journal of Advertising research, 27(1), 9-13.
Ferguson, D. A., & Perse, E. M. (1993). Media and audience influences
on channel repertoire. Journal of Broadcasting & Electronic Media, 37, 31-47.
Ferguson, D. A., & Perse, E. M. (2000). The World Wide Web as a
functional alternative to television. Journal of Broadcasting &
Electronic Media, 44, 155-174.
Gini, C. (1936). On the measure of concentration with special
reference to income and wealth. Summary of proceedings from the
Cowles Commission Research Conference on Economics and Statistics, 73-80.
Hay, D. A., & Morris, D. J. (1979). Industrial economics: Theory and
evidence. Oxford: Oxford University Press.
Heeter, C. (1985). Program selection with abundance of choice: A
process model. Human Communication Research, 12, 125-152.
Heeter, C., & Greenberg, B. (Eds). (1988). Cableviewing. Norwood, NJ: Ablex.
Lin, C.A. (1994). Audience fragmentation in a competitive video
marketplace. Journal of Advertising Research, 34(6), 1-17.
Lorenz, M. O. (1905). Methods of measuring the concentration of
wealth. American Statistical Association, 9, 209-219.
McQuail, D. (1992). Media performance: Mass communication and the
public interest. Newbury park, CA: Sage.
Napoli, P. M. (2001). Foundations of communications policy:
Principles and process in the regulation of electronic media.
Cresskill, NJ: Hampton.
Napoli, P. M. (1999). Deconstructing the diversity principle. Journal
of Communication, 49(4), 7-34.
Neuendorf, K.A., Atkin, D.J., & Jeffres, L.W. (2001).
Reconceptualizing channel repertoires in the urban cable environment.
Journal of broadcasting & Electronic Media, 45(3), 464-482.
Neuman, W. R. (1991). The future of the mass audience. New York:
Cambridge University Press.
Owen, B. M., & Wildman, S. S. (1992). Video economics. Cambridge, MA:
Harvard University Press.
Perse, E.M., Ferguson, D.A., & McLeod, D.M. (1994). Cultivation in
the newer media environment. Communication Research, 21, 79-104.
Picard, R. G. (1988). Measures of concentration in the daily
newspaper industry. The Journal of Media Economics, 1(1), 61-74.
Sherman, S.M. (1995). Determinants of repeat viewing to primetime
public television programming. Journal of Broadcasting & Electronic
Media, 39, 472-481.
Waterman, D. (1991). A new look at media chain and groups: 1977-1989.
Journal of Broadcasting & Electronic Media, 35, 167-178.
Webster, J. G., & Lin, S. (2002). The Internet audience: Web use as
mass behavior. Journal of Broadcasting & Electronic Media, 46, 1-12.
Webster, J. G., & Phalen, P. F. (1997). The mass audience:
Rediscovering the dominant model. Mahwah, NJ: Lawrence Erlbaum Associates.
Webster, J.G., & Wang, T. (1992). Structural determinants of exposure
to television: The case of repeat viewing. Journal of Broadcasting &
Electronic Media, 36. 125-136.
Yim, J. (2002). Audience concentration in the media. Unpublished
doctoral dissertation. Evanston, IL: Northwestern University.
Yim, J. (2003). Audience concentration in the media: Cross-media
comparisons and the introduction of the uncertainty measure.
Communication Monographs, 70(2), 114-128.
Youn, S. (1994). Program type preference and program choice in a
multichannel situation. Journal of Broadcasting & Electronic Media,
38, 465-475.
[1]
where n=The total number of Web sites visited 4 times. y= the
average number of visitors across Web sites visited 4 times, yk=the
number of visitors of each Web site, k= 1,2,3,…, n.
|
