Demand approach
Running head: A DEMAND APPROACH TO MEASURING DIVERSITY
The 90's experience: A demand approach to measuring diversity in broadcast
television program options for African Americans
James A. Ramos
(Student)
Michigan State University
Abstract
Addressing the limitations of traditional supply side methods used to study the
issue of diversity in broadcast television programming, this paper proposes a
demand side approach employing television ratings. Testing this alternative
methodology, it was hypothesized that diversity in programming choices for
African American households increased on commercial broadcast television as
competition in the industry increased through the growth of Fox and the arrival
of new networks in the 1990's. Results appeared insufficient to make definite
conclusions due to the study's use of a limited time frame and a small program
ranking list. However the limited support found for the hypotheses suggest that
the proposed demand side method may remain viable. Thus the paper calls for
further and deeper analysis using the demand side method.
Introduction
Collectively the three major television broadcast networks, ABC, CBS, and NBC,
enjoyed a 90% share of the primetime audience until the mid 1980's (Dizzard,
1994). This figure was ravaged as programming alternatives increased through
the rise of cable, the increase in independent stations, the proliferation of
video cassette recorders, and the entrance of rival broadcast networks
(Albarron, Pilcher, Steele, & Weis, 1991). The oligopolistic market
concentration that once fostered a spirit of cooperation and allowed the three
major networks to produce programs appealing to the largest possible audience
has decreased significantly. By 1995, these networks had an aggregate 57% share
of the primetime audience, a loss of over 30% (McClellan, 1995, April 10).
These changes in market structure have led the television industry into a state
of heightened competition that dictates the networks differentiate their
products to pursue more specialized demographics and program formats (Thomas &
Litman, 1991).
As a result, there should be a noticeable increase in program diversity,
beginning in the mid 1980's, and especially prevalent in the 1990's. It was
during this period that national cable penetration exceeded 65% (Marketer's
Guide to Media, 1993), video cassette recorders could be found in 50% of
American homes, and two more aspiring commercial broadcast networks arrived,
United Paramount Network and Warner Brothers (Collette & Litman, 1997).
A rich literature has examined diversity in television programming as it
relates to changes in the market structure and performance of the television
industry (Albarron et al., 1991; Dominick & Pearce, 1976; Lin, 1995; Litman,
1979; Wakshlag & Adams, 1985). Though these studies used varied instruments for
measuring TV diversity, their methods were all based on placing program choices
into unique categories according to genres/programming formats. This technique
produced results that were consistent across studies and that supported
theoretical assumptions about the state of the television industry. However,
two of the most recent inquiries into this area have yielded conflicting results
(Albarron et al., 1991; Lin, 1995;), leading one to question the assumptions of
this method. In an attempt to address this issue, this study will examine
diversity in television during the 1990's using a new measure that focuses on
the consumer utility (demand) aspect of programming, rather than on the supply
of unique content categories.
History and literature
Dominick & Pearce (1976) produced the seminal article studying diversity in
televison programming as it relates to market structure and industry
performance. They examined the fall weekly primetime schedules of ABC, CBS, and
NBC from 1953 through 1974, placing each program into 14 unique categories. The
categories then were analyzed for trends using several different measures of
diversity. Among these measures, the most resilient were those that looked at
the "diversity", the percent concentration of the top three categories, and
"homogeneity", an index comparing the percent of time devoted to each category
between networks. The study showed that the networks increasingly were
producing similar programming while simultaneously decreasing the number of
significant categories over the 22 years period. Thus they found that primetime
programming had become both more homogeneous and less diverse.
Dominick & Pearce (1976) correlated these facts with rising network profits
over the same period. The increased profits may have been the result of the
industry's maturing into a state of equilibrium typical of an oligopoly, wherein
the firms, becoming mutual interdependent, form a tacit "spirit of cooperation."
The increase in spirit decreased costs, resulting in increased profits. They
concluded that increased profits limited and discouraged diversity.
Litman (1979) sought to confirm the conclusions of Dominick & Pearce (1976), by
looking at the period surrounding the sudden first place primetime ratings
position of ABC in the fall 1975, which disturbed the system's equilibrium
because ABC had been the historic third place runner. In terms of diversity,
Litman proposed that the disturbed market conditions should result in the
networks, specifically CBS and NBC, taking competitive postures and
"experimenting" more with programming formats in order to reestablish previous
positions. Testing this hypothesis, he took the fall primetime schedules for
the three networks from 1974 through 1978, and coded each program into nine
unique categories. His analysis, using a Herfindahl-Hirshman Index, showed that
there was a dramatic increase in programming formats after 1975 (i.e., less
concentration), thus diversity had increased as predicted.
Litman (1979) noted, however, that measuring the number of format categories
only confirmed diversity "vertically", within the whole schedule, and that a
complete analysis of diversity required examining the choices audience members
had at a particular time, what he called "horizontal" diversity. Litman found
that there was a significant increase in the amount of horizontal diversity
after 1975, which served to further his argument that the networks had entered a
stage of heightened competition. Adding this dimension to the approach of
Dominick & Pearce (1976) allowed for a more robust analysis.
Wakshlag & Adams (1985) reexamined diversity over the periods covered by
Dominick & Pearce (1976) and Litman (1979) under the false perception that the
two studies had produced conflicting results. In fact, Litman strengthened
Dominick and Pearce's findings by approached the issue from the opposite
direction, showing that diversity would increase as the market structure
produced more rivalry. Wakshlag & Adams looked at the fall primetime network
schedule from 1950 through 1982, placing programs into 37 unique categories, a
dramatic increase over Dominick & Pearce's 14 and Litman's nine. They then used
a "classic formula for variety (entropy)...[that] takes into account both the
number of different categories offered and the concentration of materials within
those categories" (Wakshlag & Adams, 1985, p. 26), which is similar to a
Herfindahl-Hirshman Index used by Litman (1992). While particulars within the
trend differed, the general trend supported the earlier findings of both
Dominick & Pearce (1976) and Litman (1979).
The launch of the Fox Broadcasting Company (Fox) in 1986, and its subsequent
expansion to five nights a week programming by 1991, led Albarron et al. (1991)
to question how this new entrant had changed the dynamics of the industry and
affected diversity. The study used a census of regularly scheduled programs
during the television season (September through May) from 1983 through 1990 that
were broadcast on the three major networks and Fox. Programs were placed into
21 categories, and analysis was conducted using the Dominick & Pearce measures
for homogeneity, which Albarron et al. referred to "instability", and diversity.
"From 1985-1988, the index increased, ranging between 18-22. The diversity
index rose in 1989 and 1990, indicating more program categories available to the
audience" (Albarron et al., 1991). These figures are better understood in the
context of the period's heterogeneity numbers, which showed a marked increase
after 1986. Thus after 1986 the percent of total programming held by the top
three formats decreased, while the number of different formats used by the
networks increased. Albarron et al. attributed these trends to Fox's use of new
formats and to the industry's reaction to this.
The findings of Albarron et al. (1991) were contradicted by Lin (1995), who in
a similar study examined the impact of the arrival of new broadcasting and cable
programmers on the major three broadcast networks as it relates to diversity.[1]
Lin focused her analysis on just the three networks, removing Fox from her
sample since it did not offer weekly programming until 1988 and then only on a
portion of its primetime schedule and because it lacked affiliates in several
major markets during the study's time frame. She hypothesized that, "increasing
competition forced the networks to seek or maintain audiences by experimenting
with new program formats. This, in turn, should lead to an increase in the
diversity of program formats by the end of the decade" (Lin, p. 20, 1995). Lin
conducted a census of regularly scheduled primetime programs during the season
(September through May) from 1980 through 1989, coded them into 25 unique
categories, and measured diversity using a Herfindahl-Hirshman Index. While she
found fluctuation in the index from year to year, she concluded that, "it is
difficult to discern a clear pattern of change across the decade. This leaves
the study hypothesis, predicting an increase in diversity throughout the 1980s,
without support" (Lin, 1995, p. 22).
Despite her findings, it was reasonable for Lin (1995) to hypothesize a
relationship between new entrants, heightened competition, and an increase in
programming diversity. Litman (1979) demonstrated similar behavior among the
networks after ABC's sudden assent to the number one position in 1976. Lin's
(1995) study, however, appears to have suffered from its assumption that the
networks create diversity between programming formats, instead from within. The
broader implication of her study's results is that they call attention to the
shortcomings of measuring diversity using a genre categorization method.
New methodological direction
Criticism of using a genre categorization method is not new. Owen (1977) noted
that assigning programs to fixed categories assumed perfect substitution within
the group (lack of depth) and no substitution between (lack of breadth).[2]
Further, he questioned the notion that consumers could not gain some utility
from a substitute within a category.
Owen (1977) went so far as to doubt the principal of linking performance and
public policy to media programming diversity in the first place. He reasoned
that the consumer welfare derived from programming diversity only could be
evidenced truly through a measure of consumer surplus, the differential between
what the market would charge and what consumers would be willing to pay for a
product/service. According to Owen then, there is no means to measure consumer
surplus and therefore to estimating the welfare consumers derive from
programming diversity. This is because measuring consumer surplus necessitates
a monetary pricing mechanism and commercially driven television lacks a consumer
pricing model. Owen also believed that using audience number estimates was an
insufficient means to measuring diversity, since these figures lack detail as to
the "intensity" with which consumers prefer certain products/services over
others.
While agreeing with Owen's (1977) first two criticisms of diversity measures,
this study does not subscribe to the last. Owen (1977) is correct in asserting
that there is no monetary consumer pricing mechanism in the broadcast or cable
television industry, with the exception of the premium pay-cable channels (e.g.,
HBO and Showtime). It is impossible therefore to estimate consumer welfare
through diversity using a measure of consumer surplus. However, our inability
to accurately measure a phenomenon, does not negate its social significance.
Diversity remains thus an issue in the television industry related to structure
and public policy, and even if it is to be settled by market forces, maintaining
a measure of its progress may still act as an aid in assessing questions of
industry performance.
In answer to the first two criticisms of the genre categorization system used
in earlier studies, this study proposes that audience numbers may prove to be a
more useful tool in estimating diversity than had been believed previously.
Consider that despite the lack of monetary fees exacted on audience members to
view TV programming, consumers are subject to natural costs when they consume
"free television". That is they pay for programming with the time[3] and
dedication[4] they devote to watching television in place of doing some other
activity and to viewing one show over another. Thus the act of choosing to
watch a particular program consistently can be viewed as a consumer utility that
reflects the intensity of consumers' preferences.[5] Measuring this utility is
a function of the Nielsen TV ratings system. The TV ratings system by virtue of
its design reflects consumers' preferences through its measuring of audience
sizes at the showing of every episode of every TV program during virtually all
dayparts. Additionally, the TV ratings system has the added advantage of being
able to segment audience viewing figures by demographic characteristics (e.g.,
sex, age, or race).
Given that the TV ratings system measures consumer utility over the course of
time and with demographic information, it presents itself as a mechanism for
understanding consumer utility as a reflection of programming diversity. To
accomplish this one would contrast the top ranking programs among total US
households over the course of a time period (e.g., a season or year) against the
top ranking programs over the same time period for a particular audience segment
(e.g., women, Hispanics, or African-Americans). It is proposed that the greater
the heterogeneity between the shows that appear on the two lists, the greater
the strength to infer that the particular market segment's tastes are being met
with diverse programming choices. As the particular tastes of specific
demographic segments are shown be met through diverse programming, one may in
turn conclude that the programming content of the market as a whole is diverse.
A measure of this type taken over several years will reveal the trend in
diversity and thus allow for a perspective that also can consider changes in the
market structure.
This method is based on the assumption that audience segments are inclined to
prefer programming content that contains themes relevant to them, and thus will
vary in their viewing choices from total US households. This assumption is not
without empirical support. Mundorf & Brownell (1990) studying the media habits
of young adults, around age 20, and elderly adults, around age 72, found that
when it came to television viewing the two groups preferred to watch characters
of their own age group and sex. They did note an exception; young females
preferred to watch young male characters. More recently, Harwood (1997),
proposed a new theoretical perspective on media programming choice through the
integration of media uses and gratifications theory and social identity theory.
Social identity gratification theory, as Harwood referred to it, holds that,
"individuals seek out particular messages which support their social identities
(i.e., provide positive social comparisons with outgroups), and avoid messages
which do not support their social identities" (Harwood, 1997, p. 204). In
support of this idea, Harwood demonstrated that young adults and older adults
prefer to view characters of their own age.
The method proposed above answers the strongest criticisms against format
categorization schemes, since it allows any show from any genre type to be
contrasted against any show from any genre type. Using TV ratings figures in
this manner, therefore, permits the researcher to go further than merely
observing what was presented for selection by category (supply side), and
approach the issue of diversity knowing what programs were actually selected, by
whom, and with what level of dedication (demand side).[6] Observing for
variance from this perspective, it is proposed, will allow for a more complete
analysis of diversity in programming. Thus a study in this vein is warranted as
it would seek an added understanding of diversity as a performance function in
relation to market structure.
In order to explore diversity from the demand side, this study looks at
minority tastes, specifically those of African Americans, and attempts to assess
if these specialized preferences were met by the content of commercial broadcast
programmers. As competition in the industry increased through the growth of Fox
and the arrival of new networks in the 1990's it should be expected that
diversity in programming choices for African American households increased on
commercial broadcast television. Thus it is hypothesized:
H1: The difference between the top rated primetime shows among African
American households and total households grew during the period from
1990
through 1996.
Further, these needs were more likely fulfilled by the newer broadcast networks
(i.e., Fox, UPN, and WB), who it might be expected were attempting to establish
themselves in niche markets. Thus it also is hypothesized:
H2: As the difference between the top rated primetime shows among African
American households and total households grew during the period from
1990
through 1996, the networks providing the African American programming
will
increasingly be from the newer networks (i.e., Fox, UPN, and WB).
Method
In order to test these hypotheses, this study examined the top 10 regularly
scheduled primetime television programs among African Americans in relation the
top 10 shows among total households. Descriptive statistics were used to
correlate relationships between market structure and performance.
Selection rationale
Primetime was selected because it was the daypart used in previous diversity
studies. Additionally, primetime was best suited for this analysis because it
continues to garnish the highest total household ratings of any daypart. Thus,
it contains the pride programming on which the networks build their reputations,
and leads one to consider with what great care programming choices for primetime
are made.
African Americans were selected because of their central focus in the diversity
debate. The availability of data dictated that this study be limited to an
examination against total households and of only the top 10 programs. Materials
drawn from trade and popular press only ever spoke of the top 10 shows and only
against total households. Though the data for this study was gathered from a
number of different secondary sources, all rankings were based on Nielsen
ratings, almost always quoting from the same annual analysis of black television
viewing habits.
Time frame
While it would have been preferable to look at the differences between the top
10 programs for each group beginning in 1986, the year Fox launched, the
secondary source nature of obtaining TV ratings figures made this unfeasible.
Available figures begin in 1990. Therefore, the African American households and
total households were contrasted in a longitudinal analysis that began with
figures from 1990 and continued through 1996. Despite great effort to obtain
them, rankings for 1991 were not available at the time of this study. This
year, therefore, was omitted from the analysis. It was believed, however, that
despite these conspicuous gaps in the track of years, viable trends still would
be discernible, particularly as Fox did not launch a full primetime schedule
until after 1991.
Measures
vertical diversity. Vertical diversity is a measure of the programming choices
available to viewers across the whole of the television schedule. This study
uses the top 10 rankings of programs among African American households and total
households as a sample of the choices available to audience members with the
knowledge that these shows were selected most often. Vertical diversity thus
was measured using scale created by subtracting the number of programs found in
both categories and subtracting it from the 10 (i.e., the total). The closer a
year's measure is to 10, the more diversity was represented in the demand of
consumers.
horizontal diversity. Horizontal diversity is a measure of the viewing options
available to the audience at any one particular time. Litman (1979) refined
this concept[7] as a means to understanding if increases in vertical diversity
were translating into real choice for the audience. According to the measure,
diversity is limited, when "block" programming delivers essentially the same
sorts of shows to fulfill the dictates of "audience flow" strategies.
In this vein of reasoning, this study employed a measure of horizontal
diversity, looking at the two sets of rankings and counting the number of
programs that were aired opposite each other. These numbers were then analyzed
for trends. An increase in these numbers over time would suggest that African
Americans were obtaining an added sense of diversity, since they were choosing
smaller more niche oriented programming over the most popular (mass) shows.
This measure also speaks of African Americans' dedication to targeted
programming.
exhibitors. A final analysis examined the distribution of African Americans'
top 10 programs to discern which networks were most involved with delivering
this new diversity in programming, and to see how this has developed and changed
though the 1990's. Networks were separated into two groups, 3 major networks
(i.e., ABC, CBS, NBC) and 3 newest networks (i.e., Fox, UPN, WB).
Results
Vertical diversity
Diversity increased from 1990 to 1996, though only slightly. Figure 1 shows
the complete diversity scale. In 1990 three programs could be found in both
categories, but by 1996 this decreased to only one. There were no shows shared
between the categories in 1992 or 1993. NFL Monday Night Football was the only
program shared between 1994-1996. One will note that duplication already was
low between the two categories in 1990 with NBC providing eight of the top 10
shows among African American households. This indicates that some diversity
existed even before Fox launched its full primetime schedule in 1991 with NBC as
its primary source. Table 1 provides a detailed by year comparison between the
categories, including show names and networks.
Horizontal diversity
No programs ran opposite each other between the two categories in 1990, 1992,
or 1993. However, from 1994D1996 three shows were aired in opposing time slots
between the categories. In all cases the number one and two ranked programs
among African Americans ran in Thursday positions against a strong NBC Thursday
night line-up, whose shows dominate the top ten list for total households.
Clearly African Americans had choice that included some of the "hottest" shows
on television, but still chose other programs. Table 1 provides the day and
time of each program in the categories by year for comparison.
Exhibitors
In 1990 90% of the top ten TV programs in African American homes were shown on
one of the three major networks. By 1996 this number had decreased to 30%.
With the exception of 1994, this decline had been steady. Figure 2 provides a
graphic of the full trend.
A closer examination of these figures is provided by Table 2. As one will
note, Fox was the sole member of the "3 newest networks" category until 1996,
when it was joined by UPN and WB. Fox maintained a prime position in providing
choice to African Americans from 1992 on. While NBC had dominated the African
American category in 1990 (8 of 10 shows), after a steady decline, it failed to
capture a position in 1996. It should be noted that during the years reviewed
CBS did not have a single program in the African American top 10.
Discussion
It was first hypothesized that diversity in programming choices for African
American households increased on commercial broadcast television as competition
in the industry increased throughout the growth of Fox and the arrival of new
networks in the 1990's. Results are not conclusive. In terms of vertical
diversity, the diversity scale did increase over the seven years. However, the
relatively low duplication in programs between the categories in 1990, prior to
Fox's full schedule entry into the industry, suggest that some diversity was
being provided prior to the arrival of Fox as a major competitor. However there
does appear some support for this hypothesis in terms of horizontal diversity.
The success of three counter programmed showed from 1994-1996 against NBC's
strong Thursday night line-up is a firm indication that African Americans
experienced choice, and thus diversity, in their programming options. More
counter programming would have lent this measure more credibility.
It also was hypothesized that diversity would be fulfilled in large part by the
newest networks (i.e., Fox, UPN, and WB). The Fox schedule's prime position in
the African American rankings lends support for this hypothesis. However, NBC
exhibited eight of the top ten African American shows in 1990, further
suggesting that there was some level of diversity prior this study's time frame.
Clearly this study's core weakness is it's limited time frame (breadth) and
rankings (depth). A more complete analysis is necessary to determine the full
extent of programming diversity, and to give a clearer contextual frame of
demand in the 1990's,. This would require looking at the issue during the
1980's, and, if at all possible, before. Additionally, rankings beyond the top
10 would be necessary.
Given these limitations, the viability of the proposed demand side method also
must remain in question, pending future testing using lengthier more in-depth
analyses. However, there clearly is reason to believe that it will work,
particularly from the measure of horizontal diversity.
Finally, it should be emphasized that this demand side method to measuring
diversity is offered as supplement and not a substitution to a supply side
perspective. The most complete picture on programming diversity only can be
offered when the two are viewed in tangent as complements exploring the same
question from different perspectives.
References
Albarron, A. B., Pilcher, A., Steele, D., & Weis, J. (1991). Trends in network
prime-time programming, 1983-1990. Feedback, Fall, 2-5.
Collette, L., & Litman, B. R. (1997). The peculiar economics of new broadcast
network entry: The case of United Paramount and Warner Brothers. Journal of
Media Economics, 10, 4, 23-37.
Dizzard, W. (1994). Old media, new media. New York: Longman.
Dominick, J. R., & Pearce, M. C. (1976). Trends in Network prime time
programming, 1953-74. Journal of communication, Winter, 70-80.
Greenberg, B. S., & Collette, L. (1997). The changing faces on TV: A
demographic analysis of network television's new seasons, 1966-1992. Journal of
Broadcasting & Electronic Media, 41, 1-13.
Harwood, J. (1997). Viewing age: Lifespan identity and television viewing
choices. Journal of Broadcasting & Electronic Media, 41, 2, 203-213.
Levin, H. J. (1971). The invisible resource: Use and regulation of the radio
spectrum. Baltimore, MD: Johns Hopkins Press.
Litman, B. R. (1979). The television networks, competition and program
diversity. Journal of broadcasting, 23(4), 393-410.
Litman, B. R. (1992) Economic aspects of program quality: The case for
diversity. Studies of broadcasting, February, 121-156.
Litman, B. R. (1993). The changing role of the television networks. In A.
Alexander, J. Owen, & R. Carveth (Eds.), Media economics: Theory and practice
(pp. 225-244). Hillsdale, NJ: Erlbaum.
Litman, B. R., Hasegawa, K., Shrikhande, S., & Barbatsis, G. (1994). Measuring
diversity in US television programming. Studies in broadcasting, 30, 131-153.
Marketer's guide to media. (1993). New York: Adweek.
McClellan, S. (1995, April 10). Big three post record share slide. Broadcast &
Cable, 125, 15, 8.
Mundorf, N. & Brownell, W. (1990). Media preferences of older and younger
adults. The Gerontologist, 30(5), 685-691.
Owen, B. M. (1977). Regulating diversity: The case of radio formats. Journal of
Broadcasting, Summer, 305-319.
Thomas, L. & Litman, B. R. (1991). Fox broadcasting company, why now? An
economic study of the rise of the fourth broadcast "network." Journal of
Broadcasting & Electronic Media, 35, 139-157.
Wakshlag, J., & Adams, W. J. (1985). Trends in program variety and the prime
time access rule. Journal of broadcasting and electronic media, 29(1), 23-34.
Cite Greenberg, characters are different, more variety, audience with more
choice should fractionalize/segment. Thus there is reason to suspect that there
should be an increase in programming diversity.
Footnote: In the debate over television's diversity,
XXXXXXXXXXXXXXXX analyzed a week of programming from national commercial
broadcast television, public television, basic cable, and premium cable. This
week served as their sample.
.Of the 19 "narrative" categories, commercial broadcasters proved to be amongst
the least concentrated according Herfindahl-Hirshman Indices. Litman et al.
proposed commercial television was the most diverse (least concentrated) because
these networks use a widest-appeal strategy to programming that collectively
proved significant. Thus cable did not appear to offer more diversity, but
rather more of the same. (Further analysis of this article is needed).
Diversity is an extremely complex variable. Put this entrance in, somehow
acknowledging the fact.
[1] It should be noted that Litman, Hasegawa, Shrikhande, & Barbatsis (1994)
found commercial broadcast television to be more diverse than public television
or cable. Litman et al. proposed commercial television was the most diverse
because these networks used a widest-appeal strategy to programming that
collectively generated more choice. Thus cable did not appear to offer greater
diversity, but rather more of the same programming.
[2] Litman (1993) in a more recent piece hypothesized that diversity is a
commodity subject to the laws of supply and demand, and that diversity was
composed of both breadth and depth.
[3] See Litman (1993) for a discussion regarding this.
[4] Dedication, meaning the process by which individuals return to particular
programs episode after episode.
[5] Please note that this is contrary to what Owen (1977) supposed.
[6] See Litman's (1993) discussion on diversity as a commodity.
[7] Litman (1979) notes Levin (1971) as the originator of the idea of
horizontal diversity.
|
