SEASONALITY AND TALK RADIO AUDIENCES
Jeffrey S. Wilkinson, Assistant Professor
Department of Broadcasting
The University of Tennessee at Knoxville
Knoxville, TN 37996-0333
(615) 974-4291 (fax: 974-2814)
Internet: [log in to unmask]
August E. Grant, Associate Professor
Department of Radio/TV/Film
The University of Texas at Austin
Austin, TX 78712
(512) 471-6640 (fax: 471-4077)
Internet: [log in to unmask]
Paper submitted to the Media Management and Economics division of the
Association of Educators in Journalism and Mass Communication.
Seasonality and Talk Radio Audiences
It is important for advertisers and radio station managers to understand
the cyclical nature of audiences. This study examined talk radio
in all U.S. markets that are measured four times per year over a
five-year-period in order to identify some trends and evidence of
seasonality. The findings suggest that some dayparts are more seasonal than
others, some seasons benefit certain dayparts over others, and talk radio
is influenced by and influential with elections under some conditions.
SEASONALITY AND TALK RADIO AUDIENCES
The only constant about radio audiences is their inconsistency. Audiences
fluctuate with changes in program, listener preferences, and even the
seasons of the year. A great deal of research has been devoted to how
ratings change as a result of changes in station behavior or listener
preference, but almost no research has addressed the seasonality of radio
ratings. This seasonality is especially important, because as a cycle,
regular variations should be expected in radio ratings. The purpose of
paper is to identify some of the underlying elements contributing to
The problem with any such exploration is the tendency to treat radio as a
homogenous medium, when it actually is comprised of a multitude of
serving a variety of roles and goals for listeners. Rather than examining
all radio formats, the focus of this paper is upon Talk Radio, because
is designed to fulfill the broadest range of goals, simultaneously
informing and entertaining listeners.
Seasonality and Cycles
The variability in media consumption patterns reflects and is consistent
with those found in nature and society (Zerubavel, 1981). For example,
television viewing patterns have been found to be both seasonal and
cyclical (Barnett, Chang, Fink, & Richards, 1991). Some of the seasonal
factors that influence television viewing include weather (Gould,
& Chapman, 1984) and number of hours of daylight (Gensch and Shaman,
An understanding of the cyclical nature of audiences is important for
advertisers, station managers, and researchers because broadcasting
companies are especially sensitive to cycles. Davis and Lee (1980) have
reported that cycles exist for advertising sales, and sales are the
lifeblood of the media--particularly broadcasting companies. Most if not
all operating revenues for these organizations come from ad sales
Because broadcast media are so dependent on advertising, and advertising
is based on audience ratings, both advertisers and managers must know
something about regular fluctuations in audience consumption patterns as
reflected by ratings data. People have both weekly patterns of media
well as daily patterns. Two complimentary cycles are evidenced in daily
television and radio use. For example, overall radio listening is
highest in the mornings, while television viewing is generally highest at
Identification of degree of seasonality in radio ratings, as well as of
factors that increase or moderate the effect, is important to both
broadcasters and advertisers. In order to interpret ratings, broadcasters
need to be aware of the extent of regular variations in listening
for their station and their competitors.
Understanding seasonality is probably most important for advertisers, who
rely upon ratings data in the purchase of commercial time. Since
all ratings data are measures of past performance by stations,
are forced to assume that current station performance can best be
by past performance. Identifying the degree of seasonality for specific
formats will help advertisers make more informed decisions, and
their ability to predict the audience reached through a specific buy.
Radio is a vibrant medium
Radio remains the most common and mobile form of electronic communication.
There are well over 11,000 radio stations licensed in the U.S., listened
to every day by 77% of the population. Furthermore, approximately 95%
all cars have radios in them, undoubtedly a major contributor to the
estimated three-and-one-half hours daily listening time (Radio Advertising
Bureau, 1993, cited in Bartlett, 1993).
The radio industry can reap tremendous benefits from understanding cycles.
Although eclipsed by television in terms of attention and dollars, radio
is alive, vibrant, and growing. Recent figures show that 1994 overall
radio ad sales were up 11-12% over 1993 figures (Radio & Records,
6, 1995). According to Duncan's American Radio, total radio station
revenues may have surpassed the $10 billion-dollar mark for the first time
ever. The Radio Advertising Bureau has further noted that the industry
enjoyed over two years of solid growth and sales increases
Cable, November 7, 1994), and many believe that this growth will
There are some who believe that radio sales are more sensitive to
fluctuations in the economy (Broadcasting & Cable, October 24, 1994), but
little research has explored changes in audience beyond dayparts. The
notion that "radio audiences are largest in the morning, smaller at
mid-day, larger in the afternoon, and disappear at night" is well-known and
supported by ratings figures.
But if cycles have been identified for television viewing (Barnett et al,
1991), it is logical that such cycles also exist for radio listening.
these cycles also exist for radio, then it is important for
and researchers to identify and understand them. This information
used to help advertisers schedule messages, and even lead to
whether some message approaches work better than others, depending on what
part of the slope the overall audience listening cycle is located. All of
these issues are the rationale for investigating the seasonality of
The Rise of Talk Radio
Radio was forced to adapt when television came on the societal scene 40+
years ago. In order to survive, radio stations turned to niche
and the notion of "the format" was born. Today there are dozens of
some of them virtually indistinguishable from the other.
Radio listening patterns shifted dramatically in the late 70's and early
80's. AM radio listenership plummetted, as listeners migrated to the
band. Despite roughly equal numbers for commercial AM and FM stations
(roughly 5,000), the audience share for AM radio is at best one-fourth of
the total listening audience (Keith 1993/1994).
Not all radio stations have been created equal. Bates (1995) found that
daytime-only AM stations are handicapped in terms of station value.
handicap is estimated to reduce the value of those stations by
one-half that of comparable full-time AM stations. This effect was even
more pronounced in large-market AM/FM combos, where it appears that
daytime-only frequency contributes nothing to the overall value to the
Recently there have been some encouraging signs for AM stations.
Listenership levels appear to be no longer dropping, and in fact, may be
increasing. Talk Radio is leading the recovery for AM, and many
point to the phenomenon of Rush Limbaugh as a prime factor (Radio Only,
1994). Limbaugh is credited with turning many marginal operations into
successes, and has been found to have brought a substantial number of
listeners to the AM band (xxxxxxxxxxxx, 1995).
The success of the Talk Radio format (which includes the hybrid News/Talk
as well) on AM makes it of special interest to researchers. Talk Radio
found a niche that links listeners to government, information, and
in ways that music formats have not or cannot. Talk Radio is
personal (Tramer & Jeffres, 1983), and involves its listeners to a
extent than other formats (Warner & Buchman, 1993).
Talk Radio is also inextricably linked to political speech. In fact, one
of the socially-defined seasonal cycles for Talk Radio stations may be
their coverage of political activity. For example, talk show host Rush
Limbaugh generated enough controversy for Congress to discuss the
reinstatement of the Fairness Doctrine. More recently, talk show hosts
broadcast their programs from Capitol offices provided by House Speaker
Newt Gingrich (Radio & Records, January 6, 1995).
Of all the AM formats, Talk Radio may be arguably the most successful. In
many of the largest markets, the highest-rated AM station offers a
News/Talk format (Keith, 1993/1994). In fact, this type of format is
frequently the ONLY AM station ranked in the top-10 in terms of audience
ratings. The successes of the AM Talk Radio format and the role of
seasonality and cycles in listening are the focus of this study.
The hypotheses are based upon the notion that there are fluctuations in
radio audiences as there are in television audiences. This seasonality
be considered by managers and consultants because it directly impacts on
estimated station worth and value.
Although radio is a mobile medium and probably not dependent upon
environmental factors in the same way television viewing is, Media System
Dependency Theory suggests that the result will be the same.
this theory, people will turn to the media for help during times of
ambiguity and threat (Ball-Rokeach, 1985). Talk radio, by virtue of its
political and opinion-driven content, probably fulfills this function
better than other types of radio formats.
Media system dependency also suggests that dependency relationships at one
level of analysis will affect those at other levels. In this case, our
focus is on the relationship of the individual listener to the talk
medium. In considering this relationship, we must also consider the
macro-level dependency relationships, specifically between talk radio and
the political system. Ball-Rokeach (1985) specifies that the media
is dependent upon the political system for legitimation, judicial and
lative protection, and for "political decision makers' capacities to
generate conflict and drama that is grist for the media mill."
Therefore, summer months are typically a "slower" time in terms of
political, economic, and even media activity. At the same time there is
less activity among these institutions, personal activity also tends
increased leisure time which corresponds to greater available daylight
time. More vacations are taken during the summer, which further breaks
regular listening or viewing routines. All of these personal, social,
natural factors result in decreased dependency on media, and
reduced media use during summer months. Therefore, our first hypothesis
addresses this issue:
H1: The seasonality in talk radio audiences follows the same pattern of
television viewing in that warm weather listening (Spring and Summer)
be least while cold weather (Fall and Winter) listening will be
A second consideration regarding talk radio audiences is the variable of
market size. A functionalist approach would suggest that "big city"
radio listening may serve a different purpose from that of "small
talk radio. Media System Dependency Theory hints that the relatively
unstable nature of the large, supposedly impersonal metro makes the media
(and talk radio) more important to its consumers than the alleged more
personal nature of living in a small town. If Talk Radio is more
in the big city, then listeners will use greater effort to listen on a
regular basis, resulting in greater seasonality.
However, one may suggest the same argument could hold true for smaller
markets. This is based on the notion that, as the number of stations
market goes down, format duplication disappears and the station
different role in the community. In the larger markets, there may be
than one talk, news, or newstalk station competing with dozens of
signals. But in the smaller markets, there may be only one station
this programming because Talk Radio is more resource intensive (expensive)
than programming music. Since there is no direct competition for that
audience, the small market Talk Radio station better serves as a local
barometer for goings-on in the community. One may argue that this would
result in higher dependence and therefore, greater seasonality based on
regular listening cycles. Despite the competing notions of whether
small markets are subject to greater seasonality, this paper chooses to
side with the media system dependency model rather than "conventional
dom" on this matter, suggesting that a second hypothesis would be:
H2: Larger markets (those with more stations and presumably, more than one
Talk or News/Talk station) will show MORE seasonality than smaller markets
(with fewer stations and therefore fewer Talk or News/Talk stations). The
larger the area population, the greater the change from season to
Within the larger framework of seasonality in ratings, listening patterns
should also exhibit both daily and weekly cycles. For example, because
listening is greatest in the morning, this time period should be least
sensitive to events and social changes because of the greater numbers
involved. Morning listening may be perceived as ritualized listening, and
should therefore should remain relatively stable over the course of
Therefore, the next set of hypotheses focus on specific dayparts:
H3a: Morning drive times will show least seasonality compared to mid-day,
afternoon, or evening periods because morning listening is habitual
Another listening period that should show seasonality is evening
listening. Talk radio listeners who listen at night have probably
supplanted television viewing with this type of leisure activity,
H3b: Evening listening will show greatest seasonality.
A final area of interest would be to examine cyclical infuences from
another sphere of social life. Every two years there is a major political
election season, and every four years Americans elect a new President.
Because Talk Radio has become an influential and popular outlet for
political speech, it would make sense that these stations would become
more popular during election periods. As mentioned earlier, the Rush
Limbaugh program has brought new listeners to AM talk radio, and also
helped revive interest in American politics (Lewis, 1993). Because this
program airs during mid-days in most markets, it suggests a link
the mid-day time slot and recent election-period ratings. This leads
H4a: Talk Radio ratings overall will be significantly higher during an
election season (which tends to occur during Fall of every other year)
during the "off" years (odd-numbered years).
H4b: Talk Radio ratings during mid-days will be the highest during the
Fall season of election years.
The method of assessing radio listening hinges on secondary analysis of
ratings data. Ratings information was obtained from the Arbitron
Archive at the University of Georgia and from Arbitron's own archives.
Ratings data from all U.S. radio markets that are rated four times each
year were collected (N=94). This resulted in a data set comprised of
ratings for each of those markets from Spring 1988 through Spring 1993
latest date for which data were available at the time of data collection).
It is logical to believe that cyclical patterns in listening would be
identifiable over a five year period.
In addition to the ratings data, the variable of MSA (Metro Survey Area)
was collected from the 1994 Broadcasting & Cable Yearbook. The MSA
were based on 1993 Arbitron data.
After this variable was matched to the ratings data, SPSS-PC+ software was
used to test each hypothesis.
Seasonality was conceptually defined as the degree of variation between
the highest and lowest ratings period of the year. This variable was
operationalized by computing the mean rating for each daypart across the
five years, then computing the difference between the highest and
rated of the four ratings periods (fall, winter, spring, and summer).
The first hypothesis proposed that seasonality in talk radio audiences
will follow the same pattern as television viewing in that Spring and
Summer ratings will tend to be lowest, while Fall and Winter ratings will
tend to be highest. This hypothesis was supported. The maximum or
ratings for each individual station occurred during Fall or Winter
three-fourths of the time (73%), while Spring or Summer produced the
highest ratings only one-fourth of the time.
The second hypothesis suggested that the larger the market, the greater
the change from season to season. This hypothesis was supported. MSA
population as a surrogate measure of market (where more people indicates
smaller market size) was negatively related to seasonality (r=-.29,
Further research was conducted to explore the relationship between market
size and seasonality within each daypart. Strong seasonal effects were
observed during the daytime (AM drive, r=-.30, p<.005; Midday, r=-.35,
.001; PM drive, r=-.24, p<.05), but market size was not related to
seasonality of evening and weekend ratings.
The third set of hypotheses suggested seasonality predictions for various
time periods. Hypothesis 3a predicted that morning drive time will
least seasonality of the three day periods (morning drive, mid-day,
afternoon drive). This hypothesis was not supported. Although morning
was significantly less than midday and evening time periods, morning and
afternoon drive times were not significantly different.
Hypothesis 3b predicted that evenings would show greatest seasonality.
This hypothesis was supported. Evenings showed the greatest variability
all the time periods analyzed (see Table 1).
The fourth area of interest was talk radio listening during elections. It
was hypothesized that overall ratings for talk radio would be
higher during election years than non-election years. This hypothesis was
The final hypothesis focused on the relationship between election seasons
and the recent Rush Limbaugh phenomenon. Specifically, that Limbaugh
the mid-day time period for talk radio would show significantly higher
ratings during recent election seasons. This hypothesis was supported.
overall means were significantly higher during election years for the
mid-day time periods (t=2.57, df=92, p<.01).
This study sought to identify whether seasonality does extend to talk
radio. The answer seems to be a qualified "yes" with several
for advertisers and station managers. This study found support for
hypothesis that seasonality and cycles exist with radio--specifically
Radio--audiences. Market size is significantly correlated to larger
figures and greater seasonality. Also, weekends and evenings tend to show
more seasonality than the daylight time periods. This analysis was
to find clear evidence of cyclicality during the morning, midday or
afternoon dayparts, and this bears further scrutiny.
Examining arbitron ratings in the top 94 markets over a five-year-period
does reveal some important things about talk radio listening in
First of all, the daytime radio listener is different from evening or
weekend listening. Daytime listening appears to be linked to social and
political trends and events, while nighttime and weekend listening is
ritualized. This suggests that evening and weekend talk radio
the medium in the same cyclical patterns as television viewers use tele
But talk radio listening during the day is something entirely different
from these "off-hour" dayparts. Apparently, talk radio during the day
functions more as a barometer in people's daily lives, and is therefore
subject to unpredictable fluctuations. These fluctuations stem from
involvement with our world, so political, social, and economic events
trigger changes in our listening behaviors en masse. With this in
seems this study took a series of ratings "snapshots" over a five-year pe
riod to identify some key periods and trends in America (see graph in
Throughout 1988 and up to the Summer of 1989, daytime talk radio was mired
in a downward slope as fewer people were tuning in. The nadir for talk
radio stations was the time roughly between Summer 1989 and Summer
across the country, overall ratings dipped to even lower levels than the
year before. Judging from the downward slope, 1988 was not a banner
for talk radio either.
But nationally, Talk Radio showed a dramatic turnaround in the Winter
1991. This may be explained by two factors, one political and one
The Gulf War riveted people's attention and evoked extreme feelings of
partiotism and concern. Because of the ability of talk stations to
listeners (by entertaining AND informing at the same time), they
outlets for the community to express its sentiments about the war. This
political event translated into listeners (which in turn translated in
higher ratings and therefore advertising dollars) much the same as CNN
viewership skyrocketed during this period.
But the Gulf War may have simply been what triggered Talk Radio's
comeback. While CNN viewership gains disappeared after the war (Zoglin,
1992), Talk Radio ratings across the country continued to bounce back
even higher levels. This new acceptance and interest in talk radio
spurred on by controversial hosts like Rush Limbaugh.
It is suggested here that, although Limbaugh premiered in 1988, it wasn't
until the Gulf War attracted so many people to "tune in" (as
theory would predict) that many people (especially conservatives)
their spokesperson. Once they tuned in, the behavior became a habit,
Limbaugh successfully tapped into what these individuals desired in a
show. The past two years do not show an overall seasonality or
pattern for talk radio, in part because new listeners have been brought
the AM band through Limbaugh and others. This would explain why the
slope continues upward rather than a relatively stable wave of ups and
The implications of this study are numerous, and further examination of
other radio formats rating data is advised. Because evening and
listening is ritualized media behavior, management should arrange their
programming so that it matches audience needs. This is good news for
syndicators and producers of shows with a track record. Evenings and
weekends are not the best time for experimenting with new material,
these times show great seasonality. While some change and experimentation
is certainly unavoidable, programs that have become established with
listener will be the most effective, because it's become habitual
consumption. This brings a predictability to ratings that advertisers
long sought. For example, if a campaign must be launched during a time
audiences are traditionally smaller (such as spring or summer), this
knowledge allows the advertiser to set more realistic goals for the
effectiveness of their radio buy.
Another implication of this study suggests that ratings services should
consider weighting of ratings to take seasonality into account. This
effectively "level the playing field" in that advertisers could
that differences in ratings periods can be independent of the station
programming. What further study should attempt to do is identify how the
ritualized listeners are different from the more "active" daytime
listeners. Although the daytime audience tends to be much greater, they are
not very seasonal and use the medium for surveillance rather than
entertainment and leisure. This should suggest varying strategies for
advertisers to appeal to the different needs of each audience group.
In conclusion, it appears that talk radio is alive and well, and thriving.
By finding evidence of different patterns of seasonality among various
dayparts, this study suggests that radio audience fragmentation not
occurs between formats, but with a given format as well. By better
identifying the consumers of a given radio format such as talk radio and
the functions it fulfills, station managers and advertisers may better
those listener needs.
Andreasen, M. (1985). Listener recall for call-in versus
structured interview radio formats. Journal of Broadcasting &
Electronic Media. 29(4), 421-430.
Ball-Rokeach, S. (1985). The origins of individual media system
dependency: A sociological framework. Communication Research,
Barnett, G.A., Chang, H.J., Fink, E.L., & Richards, W.D. (1991).
Seasonality in television viewing. Communication Research.
Bates, B.J. (1995). What's a station worth? Models for
determining radio station value. The Journal of Media
Economics, 8(1), 13-23.
Bartlett, D. (1993). News radio--More than masters of disaster.
Media Studies Journal, 7(3), 37-49.
Davis, D.K., & Lee, J.W. (1980). Time series analysis models for
communication research. In P.R. Monge & J.N. Cappella (Eds.),
Multivariate techniques in human communication research (pp.
429-454). New York: Academic Press.
Gensch, D., & Shaman, P. (1980). Models of competitive
television ratings. Journal of Marketing Research, 17, 307-315.
Gimme gimme: EFM is asking stations to pony up more cash. (1994,
July). Radio Only. p.24.
Gould, P., Johnson, J., & Chapman, G. (1984). The structure of
television. London: Pion.
Keith, M.C. (1993). Whither (or wither?) AM? Media Studies
Journal, 7(3), 105-111.
Keith, M.C. (1993/1994). AM radio: The status and the struggle.
Journal of Radio Studies, 2, 1-10.
Lavine, J.M., & Wackman, D.B. (1988). Managing media
organizations: Effective leadership of the media. New York:
Lewis, T. (1993). Triumph of the idol--Rush Limbaugh and a hot
medium. Media Studies Journal, 7(3), 51-61.
Petrozzello, D. (October 24, 1994). Broadcasters see revenue
growth in radio's future. Broadcasting & Cable, pp.37-38.
Petrozzello, D. (November 7, 1994). Radio Groups prosper in
third quarter. Broadcasting & Cable, p.58.
Radio Revenue Reports Rosy. January 6, 1995. Radio & Records,
Rehm, D. (1993). Talking over America's electronic backyard
fence. Media Studies Journal, 7(3), 63-70.
Tramer, H., & Jeffres, L. (1983). Talk radio -- Forum and
companion. Journal of Broadcasting & Electronic Media, 27(3),
Warner, C., and Buchman, J. (1993). Broadcast and cable
selling. Belmont CA: Wadsworth Publishing Company.
Zerubavel, E. (1981). Hidden rhythms: Schedules and calendars in
social life. Chicago: University of Chicago Press.
Zoglin, R. (1992, Jan 6). How a handful of news executives make
decisions felt round the world. Time, 139(1), pp.30-32.
Variable Mean SD t-value df 1-tail Sig
Evening 2.16 1.899
AM drive 1.26 .926
4.96 93 p<.001
Evening 2.16 1.899
Mid-day 1.69 .978
2.31 93 p<.05
Evening 2.16 1.899
PM drive 1.12 .877
5.62 93 p<.001
Evening 2.16 1.899
Weekend 1.36 1.126
4.74 93 p<.001