Structural and "Loyalty"
Determinants of Late Local News Ratings
INTRODUCTION
Local television news was once considered a necessary nuisance for many
station programmers. It was seen as an economic drain on resources,
but
an obligation due to F.C.C. requirements to serve the public
interest.
Today, television news is typically held in higher esteem because it
is
usually a money-maker. Up to 40% of a station's revenues may come
from
news operations.[1] Yet recently, some stations have decided that
they would
be better off financially without an eleven p.m./ ten p.m.
newscast.[2]
That comes in an era in which F.C.C. requirements have been much
relaxed
and local news costs are high.
While local newscasts are of obvious concern to station operations, they
are also an integral part of many people's lives. These days the
average
American watches approximately four hours of television daily.[3]
In a
study of San Diego residents' news viewing patterns, Wulfemeyer[4]
found
over 69% of his respondents reported watching local news at least five
times per week. Almost 42% reported that television is their major
source of information about local events and issues.
Yet, even with the clear importance that local newscasts play in the
part of station operations and for the public as a whole, relatively
little is reported in the academic press about how people watch the
news.
This study is an initial investigation of news viewing patterns of the
late evening news. A number of factors are examined that appear
to
have some bearing on late news viewership of a particular station:
time
zone (ten p.m. newscasts versus eleven p.m newscasts), channel
number, a
station's early news rating, the station's lead-in show (the rating
on
the show which precedes the late news), the lead-out show, and the
total
number of people in a market viewing television during news time.
Like other television programming, late television news viewing patterns
are based on a number of factors, including those under
investigation
here. Webster & Wakshlag[5] proposed a model that predicated audience
viewing on six major factors: the structure of program options
available
to the viewers (the way shows are programmed on one or several
channels),
the viewers' preferences for particular types of programs, the
viewers'
availability (whether or not they are watching TV in the first
place),
the viewers' needs, whether or not viewers are in group settings, and
the
viewers' awareness of program options available at a particular
time.
Similarly, Cooper has proposed a structural model of syndicated
programming which examined the synergistic impact of several "structural"
factors on viewing patterns: inheritance effects, channel type,
market
concentration, and cable penetration.[6] In a later study
specifically
focusing on news audiences, Webster and Newton argued that several
structural factors affect news audiences: long-term market
characteristics and transitory program scheduling
characteristics.[7]
While structural factors such as the audience rating of lead-in shows
are a predicted determinant of news programming, so are individual
preferences and needs as might be argued by uses and gratifications
researchers. One such preference may be an audience loyalty to one
particular newscast versus other newscast alternatives. Indeed, even
structuralists Webster and Wakshlag[8] argue that channel loyalty is
one of
the key determinants of audience flow patterns from one program to
another.
HYPOTHESES
A "lead-in" effect suggests that the audience watching a show on one
station has a greater tendency to watch the following show on that
station than do other viewers. This may be due to a number of reasons:
sedentary viewers (too "lazy" to change the channel, however this
may be
diminishing due to the increase in remote control viewing devices),
the
lack of preference for any other show, viewership as a secondary
activity
(those who are not paying full attention to the screen), etc.
There appears to be ample evidence that a lead-in effect and inheritance
effects in general, do indeed exist. In an analysis of CBS network
programming for a 23-year period, Tiedge & Ksobiech[9] determined that a
strong lead-in show produced an advantage of four share points for the
subsequent show. In a study of early evening local news across U.S.
markets, Webster & Newton[10] demonstrated a strong relationship
between the
lead-in rating and the local news rating, a correlation of .75.
These
researchers determined that three factors accounted for over 80% of
the
variance between network affiliates' early news ratings: a
station's
network news rating, the lead-in to the local news, and the size of
the
available news audience. In an analysis of late evening news
performance for the three network affiliates in the Dallas-Ft. Worth
area, Boemer[11] found that lead-in shows were significantly
correlated to
late evening news with an overall correlation of .58.
It is anticipated such a positive correlation also exists across
markets. Similarly, while a lead-in show can benefit the ratings of the
late news show, the late news show can serve as a lead-in to shows
which
follow it.
HYPOTHESIS 1: A station's late news audience rating should be positively
correlated with the audience rating of the (lead-in) show
immediately
preceding the newscast and with the audience rating of the (lead-out)
show immediately following the late news.
Audience loyalty is a concept that is approached in different ways by
different researchers. For example, researchers for the PBS
television
network conceived of audience loyalty as viewer patterns to stick
with a
program rather than switch channels or turn off the set. A high
rate of
switching off a program by viewers would demonstrate less viewer
loyalty
to the program[12].
Another way to conceive of audience loyalty is by considering repeat
viewers--that is by gauging those listeners who tune into a subsequent
episode of the same program. One would assume that "loyal" viewers
would
return to watch more programs. British researchers studied whether
early
evening news viewers would remain loyal to news programming and
watch
later news programming or whether they would become news "saturated."
The 1967 study of British television viewing housewives concluded
that
"There was no inhibition about watching more than one news bulletin.
In
fact, the above results imply that of people who watched a six p.m.
bulletin (ITV or BBC), about 75% would also watch one at nine p.m."[13]
There are good reasons to assume news loyalty to one particular station
might exist. Some viewers may be more apt to watch a low numbered
VHF
station than a high numbered UHF station. [14] The news director for
the
Chicago ABC affiliate conceded that part of his station's news
audience
was attributable to a better reception than his competitors'
signal.[15]
Some news viewer loyalty is undoubtedly due to the news personalities
on
some news stations. "Q-scores" are used to rate news anchors on
likability, etc. Eastman, et. al. argues that "the affiliate with the
strongest news team usually dominates its market." [16]
There is evidence to support the notion of news viewer loyalty to a
particular station. In his study of San Diego resident news
viewership,
Wulfemeyer reported "almost 30% of the respondents reported they
watched
one television station for local news rather than another because of
the
'quality' or 'format' of the newscast."[17] Note that fits with a
uses and
gratifications approach that suggests news viewers choose programs
that
best meet their needs.
Based on notions of viewer loyalty, one would assume that early evening
news shows that perform well in terms of viewership would correlate
positively with the late evening news shows on the same channel. Hence,
I argue here for both channel loyalty among viewers and program
loyalty
(although it is conceded many of the news "characters" may change
from
the early newscast to the late newscast.) Late evening news viewer
loyalty is here conceptualized as watching a station's newscast based on
preference for a particular station's newscast as evidenced by early
evening newscast popularity.
HYPOTHESIS 2: A station's late news audience rating should be positively
correlated with its early evening news audience rating.
Thus far, it has been argued that a late news audience rating should be
positively correlated with both audience ratings for its lead in
program
and for its corresponding early evening newscast. Which of the two
audience ratings figures account for more variance in the ratings of
late
evening newscasts? It is expected that the lead-in should account
for
more explanatory power than would audience loyalty as evidenced by
the
correlation between the audience ratings for the early evening local
news
and the late evening news. In other words, the typical viewer
would be
more apt to tune to the newscast on the channel of his/her lead-in
program than to the newscast of the channel watched for the early evening
local news, should the two stations be different.
Logic suggests that despite newscast loyalty, some viewers would be apt
to prefer watching a different late evening newscast than the one
watched
in the early evening. That is due to a desire to watch something
new
rather than news story versions they had seen earlier. Webster and
Lichty argue that television viewers tend to be less channel loyal than
radio listeners. In addition, they suggest there is little network
program loyalty. "When we look at the audience for broadcast network
television then, there is very little evidence of what might be called
program type loyalty."[18]
In addition, arguments suggesting that television viewing is a
relatively passive behavior[19] would lead one to believe that inheritance
effects would take precedence over loyalty effects. McDonald and
Reese
have argued that viewing selectivity (as would be evidenced by
program
loyalty) mark an active audience.[20]
HYPOTHESIS 3: The positive correlation between a station's late news
audience rating and its lead-in show should be higher than the
positive
correlation between a station's late news audience rating and its
early
evening news audience rating.
HYPOTHESIS 4: There should be no significant difference between
stations' late news audience ratings based on network affiliation.
There is also evidence that channel assignment can have some bearing on
how strong a station's ratings are. Researchers have demonstrated a
weak, but statistically significant relationship between early news
ratings and channel assignment.[21] Channels with lower numbers
(typically
VHF stations) outperformed channels with higher numbers (UHF.) One
might
assume these differences to be still existent, but less with greater
cable penetration today, which puts UHF and VHF stations on an equal
foo
ting in their reception in cable homes.
Another factor that should have a strong bearing on a station's news
ratings is simply the total number of people watching television at
the
time of the news.[22] This is measured as the PUT level (total
number of
persons in a market using television at a particular time.) One would
logically assume that the larger the number of people watching
television, the larger would be a station's news rating (which is a p
ercentage of the total number of people in a market watching the news.)
HYPOTHESIS 5: A station's late news audience rating should be negatively
correlated with its channel assignment and positively correlated
with its
PUT level.
The number of independent stations which counterprogram in a market
should also have some bearing on the news ratings of a network
affiliate.[23]
As competition increases, the number of viewers for a specific news
show should decrease.
HYPOTHESIS 6: A station's late news audience rating should be negatively
correlated with the number of independent stations in a market.
The time of day for which the late news airs is also likely to have a
bearing on how strong are its ratings. Peak television viewing for
US
adults is between nine and ten p.m. and thereafter, falls off quite
rapidly.[24] Therefore it is anticipated that late news shows that air
at
ten p.m. will outperform those that air at 11 p.m.
HYPOTHESIS 7: Late news audience ratings should be higher in the central
and mountain time zones (which typically air at ten on network
affiliates) than those which are in the eastern and pacific time zones.
METHOD
An available set of Nielsen ratings data from the four week average for
Wednesday evenings in February 1989 were examined for ABC, CBS and
NBC
stations throughout the U.S. All Nielsen Designated Market areas in
which there were all three major network owned or affiliated stations
were used in the analysis (N=138). Due to incomplete data, nine cases
were deleted for full analysis. One station from each network was
randomly selected from each market to ensure an equal representation of
each network. Because of natural dispersion of network stations in
the
U.S., there were more stations from the eastern and central time
zones
than from the mountain or pacific regions. Designated Market Area
ranks
ranged from one to 199.
This was a correlational analysis of the relationship between the
dependent variable (late news ratings) and a variety of independent
variables. One way analyses of variance were conducted to assess
whether there were statistical differences between networks and late news
ratings, and between time zones and late news ratings. A step-wise
regression analysis was conducted following the correlation findings.
In this study, the ratings data used was for all persons aged two or
older.
Late news ratings were based on an average between the rating for the
first 15 minutes of the program and the second 15 minutes. Lead-out
ratings were based on just the first 15 minutes following the late
local
news. Other ratings were based on the half-hour total.
RESULTS
Table l presents the correlational matrix of late local news ratings and
the predicted correlates.
Table 1
Pearson Correlation[25] Matrix
[26] LATERTG LEADIN LEADOUT EARLYRTG CHANNEL PUT INDIES
LATERTG 1.000
LEADIN 0.653 1.000
LEADOUT 0.786 0.561 1.000
EARLYRTG 0.484 0.317 NS 1.000
CHANNEL -0.324 NS -0.284** -0.235**
1.000
PUT 0.541 0.360 0.654
NS
NS 1.000
INDIES NS NS 0.198* NS
-0.244 NS 1.000
[27] Unless labeled otherwise, all correlations significant at p. <0.001.
NS (Not significant)
*indicates p <.05
**indicates p <.01
[28] LATERTG-late news rating, LEADIN-lead-in show rating, LEADOUT-lead-out show
rating,
EARLYRTG-early news rating, CHANNEL-channel number, PUT-number of
persons using
television in the market, INDIES-number of independent stations in the
market.
BARTLETT CHI-SQUARE STATISTIC: 375.616 DF= 21 PROB= 0.000
--------------------------------------------------------------
As predicted there are strong correlations between the late news rating
and both the lead-in and lead-out programs. (The correlation
between the
late news ratings and the lead-in program are a bit deflated
compared to
the correlation between late news rating and lead out program due to
using the full half-hour rating for the lead-in show compared to the
rating of just the first quarter-hour of the lead-out show. The
lead-in
program ratings are not listed in quarter-hour increments in the
ratings
books.)
Also as predicted, there was a strong correlation between a station's
early news rating and its late news rating. As predicted this
relationship was not as strong as the effect of the lead-in show.
The expectation that lower channel assignments would garner higher
ratings was confirmed. Also confirmed was the expectation that ratings
would be strongly correlated with PUT levels during the news time.
Contrary to expectations, there was no statistically significant
relationship between a station's late news ratings and the number of
independent stations in the market. (This is likely due to relatively
high levels of cable penetration in most markets which leads to a
uniformity of viewing options between different markets.)
Also, contrary to expectation, there were statistically significant
differences on late news ratings between stations with different
network
ownership or affiliation.
Table 2
Late News Ratings of Differing Network Affiliates
NETWORK = 1 (ABC)
N OF CASES 43
MEAN 5.430
STANDARD DEV 3.285
NETWORK = 2 (CBS)
N OF CASES 43
MEAN 7.407
STANDARD DEV 3.686
NETWORK = 3 (NBC)
N OF CASES 46
MEAN 6.630
STANDARD DEV 3.241
BARTLETT TEST FOR HOMOGENEITY OF GROUP VARIANCES
CHI-SQUARE = 0.865 DF= 2 PROBABILITY = 0.649
ANALYSIS OF VARIANCE
SOURCE SUM OF SQUARES DF MEAN SQUARE F PROBABILITY
BETWEEN GROUPS 85.356 2 42.678 3.679 0.028
WITHIN GROUPS 1496.636 129 11.602
-------------------------------------------------------------------------------
A one-way analysis of variance, (Table 2) shows a statistically
significant difference between the highest rated network (CBS) and the
lowest rated network (ABC). Even when lead-in effect was controlled
for
(making lead-in rating a covariate) there were clear statistical
differences between networks.
TABLE 3
Late News Ratings of Differing Time Zones
TZ = 1 (Eastern)
N OF CASES 51
MEAN 5.588
STANDARD DEV 2.347
TZ = 2 (Central)
N OF CASES 52
MEAN 8.183
STANDARD DEV 3.840
TZ = 3 (Mountain)
N OF CASES 11
MEAN 7.773
STANDARD DEV 3.036
TZ = 4 (Pacific)
N OF CASES 16
MEAN 2.719
STANDARD DEV 1.032
SUMMARY STATISTICS FOR LATERTG
BARTLETT TEST FOR HOMOGENEITY OF GROUP VARIANCES
CHI-SQUARE = 31.462 DF= 4 PROBABILITY = 0.000
ANALYSIS OF VARIANCE
SOURCE SUM OF SQUARES DF MEAN SQUARE F PROBABILITY
BETWEEN GROUPS 446.334 4 111.583 12.478 0.000
WITHIN GROUPS 1135.659 127 8.942
-------------------------------------------------------------------------------
As anticipated, a one-way analysis of variance (Table 3) suggests there
are indeed different ratings levels to be associated with different
time
zones and times in which the late news aired.
Those shows in the central and mountain time zones with ten p.m. news
shows had higher ratings than those in the eastern and pacific time
zones
with eleven p.m. news shows. (See Diagram 1).
Diagram 1
[--- Pict Graphic Goes Here ---]
A step-wise regression was performed using four of the correlated
variables to provide a predictive model for late news rating performance.
Table 4 suggests that four variables account for about 80% of the
variance in late news ratings.
------------------------------------------------------------------------
TABLE 4
Determinants of Late Local News Ratings
STEPWISE REGRESSION WITH ALPHA-TO-ENTER=0.150 AND ALPHA-TO-REMOVE=0.150
STEP= 1 ENTER LEADOUT R= 0.786 RSQUARE= 0.617
STEP= 2 ENTER EARLYRTG R= 0.866 RSQUARE= 0.750
STEP= 3 ENTER LEADIN R= 0.881 RSQUARE= 0.776
STEP= 4 ENTER PUT R= 0.890 RSQUARE= 0.793
Of the four variables the lead-out show accounted for the most variance,
followed by the early news rating, the lead in show, and the PUT
level at
the time of the late newscast.
When the lead-out show is not counted as part of the regression
equation, Table 5 suggests that four factors can explain about 70% of the
variance in the late news ratings.
------------------------------------------------------------------------
TABLE 5
Determinants of Late Local News Ratings
(Excluding Lead-out Rating)
STEPWISE REGRESSION WITH ALPHA-TO-ENTER=0.150 AND ALPHA-TO-REMOVE=0.150
STEP= 1 ENTER LEADIN R= 0.653 RSQUARE= 0.427
STEP= 2 ENTER PUT R= 0.731 RSQUARE= 0.534
STEP= 3 ENTER EARLYRTG R= 0.826 RSQUARE= 0.682
STEP= 4 ENTER CHANNEL R= 0.836 RSQUARE= 0.699
------------------------------------------------------------------------
Of the four variables, the lead-in show accounted for the most
variance, followed by the PUT level, the channel's early news rating, and
the channel number.
DISCUSSION
The results suggest that structural factors indeed account for the large
proportion of variation in late local news ratings. The strongest
correlations are between late news ratings and the shows which follow and
precede the news. Such a relationship points to the key
interdependence
between local affiliates and networks. A strong performance by the
local
news can have a big bearing on the show which follows. A strong
performance by network or syndicated programming can also have a big
bearing on the late local news program. On the night of Johnny
Carson's
final appearance on the Tonight Show, the NBC station's anchor in
Chicago
paid tribute to what he termed the "Carson factor."[29] He
suggested that
NBC local news shows have benefited over the past 30 years from the
lead-out effect of the popular late-night show.
The strong relationship between late news shows and their lead-in and
lead-out shows also adds credence for the notion of a relatively
passive
audience viewing pattern despite the advent of remote control
devices
that allow for easy zapping. It also puts into question basic
premises
of economic models of audience behavior, information processing
models,
and uses and gratifications approaches that suggest an active
audience
member choosing the news show which best meets his or her needs.
Instead, there appear to be clear audience flow effects from one show to
the next.
There is some evidence of audience loyalty in terms of the significant
positive correlation between early news and late news ratings. This
lends support for the notion of a "news leader" in many communities.
This is also perhaps, partly related to network affiliation, noting
that
contrary to expectations, significant differences do occur between
CBS,
NBC and ABC stations. An important follow-up analysis would follow
the
individual viewer as a unit of analysis to determine if individuals
demonstrate station loyalty--watching the early and late news on the
same
channel, or whether the individual who is able to watch both
newscasts
seeks diversity in news programming, or whether s/he will choose to
watch
news at one time period and entertainment programming at another.
The significant differences between time zones in which ten p.m.
newscasts appear to significantly outperform eleven p.m. newscasts
suggests news stations that indeed garner a high proportion of a
station's earnings in the eastern and pacific time zones might consider
the value of a change in timing of prime-time and news programming.
Indeed there is evidence that suggests that a trend may be underway in
which Americans are adjusting their daily schedules earlier and that
television programmers, especially in the eastern and pacific time
zones,
are reacting to that change.[30]
The regression equations which suggest that the bulk of a station's news
ratings can be predicted using general structural and program
loyalty
factors suggest that these kinds of variables are far more important
from
a predictive standpoint than are specific concerns regarding content
or
anchorpersons' Q scores.
75-word Abstract
In this paper, the author establishes a strong link between late TV news
ratings and structural factors such as lead-in and lead-out effect,
the
total number of persons using television, and channel position.
The
author also establishes a link between late TV news ratings and
loyalty
factors as measured by early TV news ratings and network
affiliation.
Finally, the author establishes a strong link between late news
rating
and time-zone.
150-word Abstract
In this paper, the author establishes a strong link between late TV
news ratings and structural factors such as lead-in and lead-out
effect,
the total number of persons using television, and channel position.
Newscasts preceded or followed by strong shows will typically garner
significantly higher ratings.
The author also establishes a link between late TV news ratings and
loyalty factors as measured by early TV news ratings and network
affiliation. Late newscasts on stations with strong ratings on their
early news shows will typically fare significantly better in the
ratings
than those with weaker earlier ratings.
Finally, the author establishes a strong link between late news rating
and time-zone. The author argues the significantly lower ratings
for
newscasts on the two coasts is likely due to the later time in which
the
newscasts are aired (11 p.m. versus 10 p.m.).
Structural and "Loyalty"
Determinants of Late Local News Ratings
by
Dwight DeWerth-Pallmeyer
Assistant Professor
Department of Speech
Utica College of Syracuse University
1600 Burrstone Road
Utica, NY 13502-4892
(H) 315-733-9130
(O) 315-792-3086
For Submission to
Radio-Television Journalism Division
AEJMC
In c/o
Marcia Z. Guckes
Media Communications Department
Webster University, St. Louis, MO 63110
March 15, 1994
[1] Boemer, Marilyn Lawrence, "Correlating Lead-in Sho
w Ratings with Local Television
News Ratings," Journal
of Broadcasting & Electronic Media 32:89-94, (1987).
[2] "Knoxville T
V Drops its Late Night News," Broadcasting, (May 13, 1991, p 36).
[3] W
ebster, James G. and Lichty, Lawrence W., Ratings Analysis: Theory and
Practice
, (Hillsdale, N.J.: Lawrence Erlbaum Associa
tes, Publishers, 1991.)
[4] Wulfemeyer, K. Tim, "The Interests and Pref
erences of Audience for Local
Television News," Journa
lism Quarterly 59-60:323-328, (1983).
[5] Webster, James G., and Waksh
lag, Jacob J., "A Theory of Television Program
Choice,
" Communication Research 10:430-446, (1983).
[6] Cooper, Roger, "An Ex
panded, Integrated Model for Determining Audience Exposure to
Televisio
n," Journal of Broadcasting & Electronic Media
37:401-418, (1993) p. 403
.
[7] Webster, James G., and Newton, Gregory D., "Structural Determinan
ts of the
Television News Audience," Journal of Broadc
asting & Electronic Media 32:381-389,
(1988) p. 382.
[8] Webster and Wakshlag.
[9] Tiedge, James T. and Ksobiech, Kenneth
J., "The Sandwich Programming Strategy: A
Case of Au
dience Flow," Journalism Quarterly 65:376-383, (1988).
[10] Webster an
d Newton.
[11] Boemer.
[12] Eastman, Susan Tyler, Head, Sydney W. Kl
ein, Lewis, Broadcast/Cable Programming:
Strategies and Practices, 3rd
ed., (Belmont, C.A.: Wadsworth Publishing Co., 1989).
[13] Goodhardt,
G.J., Ehrenberg, A.S.C., Collins, M.A., The Television Audience:
Patterns of Viewing, 2nd ed., (Aldershot, U.K., 1987), p. 48.
[14] Webster and Newton.
[15] WLS Television News Director, Intervie
w, Fall (1988).
[16] Eastman, p. 219.
[17] Wulfemeyer, p.325.
[18]
17Webster, and Lichty.
[19] Neuman, W. Russell, The Future of the Mass
Audience, Cambridge, U.K.,
(Cambridge University Pres
s, 1991).
[20] McDonald, Daniel G. and Reese, Stephen D., "Television
News and Audience
Selectivity," Journalism Quarterly
24:763-768.
[21] 20Webster and Newton.
[22] 21Webster and Wakshlag.
[23] 22Owen, B.M., Beebe, J.H., & Manning, W.G., Jr. Television Econom
ics.
Lexington, MA.: (Lexington Books, 1974); Webst
er, James G., "Cable Television's Impact
on Audiences for Local News,"
Journalism Quarterly, 61: 419-422, (1984); Webster,
J
ames G., "Audience Behavior in the New Media Environment," Journal of Com
munication,
36(3): 77-91, (1986); Webster and Newton.
[24] 23Barwise, Patrick and Ehrenberg, Andrew, Television and its Audi
ence, (London:
Sage Publications Ltd., 1988).
[25] W
MAQ-TV 10p.m. newscast, (May 22, 1992).
25"More people Not Ready for Pr
ime Time," The Washington Post, (January 18, 1992,
p.A
l).
[26] WMAQ-TV 10p.m. newscast, (May 22, 1992).
[27] ore people Not R
eady for Prime Time," The Washington Post, (January 18, 1992,
p.Al).
[28] oemer, Marilyn Lawrence, "Correlating Lead-in Show R
atings with Local Television
News Ratings," Journal of
Broadcasting & Electronic
[29] Media 32:89-94, (1987).
_"Knoxville
TV Drops its Late Night News," Broadcasting, (May 13, 1991, p 36).
[30]
_Webster, James G. and Lichty, Lawrence W., Ratings Analysis: Theory an
d
Practice, (Hill
|
