Crime and Agenda-Setting, 1988-1995:
The Relationships Among the President, the Press, and the Public
Patrick M. Jablonski, Ph. D.
The University of Central Florida
School of Communication
P.O. Box 161344
Orlando, Florida 32816-1344
(407) 823-2840 (Office)
(407) 823-6360 (Fax)
[log in to unmask]
William. J. Gonzenbach, Ph. D.
The University of Alabama
The College of Communication
P.O. Box 870172
Tuscaloosa, Alabama 35487
(205) 348-7668 (Office)
(205) 348-9257 (Fax)
A paper submitted to the 1996 AEJMC Convention in Anaheim, California
Running head: Crime and Agenda-Setting, 1988-1995
Prior research has established that the media disproportionately
emphasize violent crimes, resulting in greater public crime fears. Despite a
decrease in the rate of violent crime, the percentage of the public identifying
crime as the nation's most important problem literally exploded during the
second half of 1993. This study examines the relationship between the mass
media, the president, public opinion regarding the crime issue in the United
States from 1988 to 1995. Cross-lagged correlations are used in an attempt to
assess the agenda relationships among the press, the president, and the public.
Most important problem survey results from multiple organizations are aggregated
into a time series of 96 monthly time points to measure the public agenda. The
media agenda is developed from an analysis of the monthly frequency of crime
stories in The New York Times. The presidential agenda is developed from a
similar analysis of the Public Papers of the Presidents. Results indicate that
public opinion about crime influences the press agenda. The press agenda, in
turn, influences the presidential crime agenda. The president appears to also
influence the public agenda, but at a 5-month lag.
Crime and Agenda-Setting, 1988-1995:
The Relationships Among the President, the Press, and the Public
Despite a decrease in the rate of violent crime (Whitman, 1996), the
percentage of the public identifying crime as the nation's most important
problem literally exploded during the second half of 1993. Indeed, by the end
of 1993, crime had supplanted all other problems as it became the public's
number one concern (Morin, 1994). Brown and Pasternak (1994, p. A1) note that
the ascendance of crime has little basis in statistical reality: "Fear flows
with its own infernal logic. The upsurge is not driven by a sudden rising tide
of crime; the rate of violent episodes, although much higher than 3 decades
ago, is down slightly from peaks in the 1980s. The difference is in the
perception of crime's pattern."
Policy-makers at federal and state levels were likewise concerned
with crime by the end of 1993. Crime legislation, such as the Brady Gun Control
Law, the Omnibus Crime Act, and a variety of "three strikes and you're out" laws
are all evidence of the importance of the crime issue. The increased concern
about the crime issue, starting in 1993, was not limited to the public and
political elites, however. Media coverage of crime has similarly expanded since
the early 1990s (Whitman, 1996).
The rise of crime on the American agenda provides a unique
opportunity to investigate the relationship between public opinion, the press,
and the president. The purpose of this study is to determine what forces
influence, and were influenced by, crime-related public opinion from 1988 to the
end of 1995. Specifically, this study investigates the agenda relationships
among the president, the press, and the public regarding crime.
Crime: The Media and Public Opinion
A number of studies have found that the amount of crime news is
driven by forces other than the level of crime. In an early study, Davis (1952)
found that crime rates do not correlate with newspaper coverage, while public
opinion more closely mirrors the press, a finding later supported by Sheley and
Ashkins' (1981) investigation of crime trends. Further, there are indications
that such disproportionate coverage can lead to an exaggerated fear of crime
(Payne & Payne, 1970). Hindelang (1974) noted that media coverage patterns
result in the perception that violent crimes are the most common, a conclusion
echoed by Erskine (1974a, 1974b, 1974c, 1974d). A number of studies echo this
claim. Fishman (1978) explicated how news reporters created a "crime wave" by
highlighting a few crimes against the elderly in New York City. News coverage
of crime can be fairly distorted. Jones (1976) reported that murders received
90 times the coverage of other major offenses in St. Louis newspapers.
Similarly, Graber (1979) found that the most violent crimes are exaggerated at
the expense of other offenses. Moreover, crime news coverage was found to be
driven by forces other than social significance. The most obvious reason for
covering violent crime is consumer demand or fascination (Katz, 1987).
Ultimately, the power of media coverage was underlined by Einsiedel, Salomone,
and Schneider's (1984) finding that the fear of victimization is better
predicted by media exposure than by actual experience with crime.
The paradox of perceptions differing from reality is not a new
phenomenon. The dramatic rise of crime on the national agenda was greatly
influenced by a complex mix of events, factors, and various societal components.
Agenda-setting research may give an indication of how crime, a virtual non-issue
in early 1993, had become a major issue by late 1993. The increased media
coverage of crime, coupled with presidential emphasis, public opinion and the
public's fear of crime have all shaped the life of the crime issue. As Smith
(1987) argues, the media and other societal forces mutually influence each other
over time, defining the scope of major issues.
Agenda-setting traces its foundation back to Walter Lippmann (1922),
who was among the first to argue that the issues which concern the public are in
large part determined by the media. Cohen (1963) subsequently argued that the
media, while not successful in telling us what to think, are extremely
successful in telling us what to think about. Agenda-setting research has
flourished in the years since the publication of McCombs and Shaw's (1972)
seminal study. According to the agenda-setting hypothesis, the amount of public
concern for an issue is in part a function of media emphasis of that issue.
Over 200 agenda-setting articles have been identified in the social science
literature since 1972 (for a review, see Rogers, Dearing, and Bregman, 1993).
Prior research suggests that public concern about a particular issue
does not operate in a vacuum on the political landscape. Instead, public
opinion may have a hand in driving presidential attention, which in turn affects
media emphasis. The inverse of these processes may be operating as well.
Indeed, all three of these societal components interact at some level regarding
major issues. Gonzenbach (1992) found that, when framed in the context of the
drug issue, the agendas of the press, the president, and the public reciprocally
affect each other. The rise and fall of issues on the American political
landscape demonstrates the complexity of the processes involved in the study of
Given that one major component involved in determining the nature and
strength of the relationship between the press and public opinion is time,
cross-sectional "snapshots" of this process are suspect in their accuracy.
Erbring, Goldenberg, and Miller (1980), argue that time series analysis in
agenda-setting is necessary in order to develop more realistic and accurate
studies by avoiding the narrow focus of cross-sectional designs. There are
other advantages to a time series approach to agenda-setting: First, media
effects on public opinion are cumulative. By analyzing data over time,
researchers might be able to measure this phenomenon. Second, since it is
impossible to control for confounding variables in the typical agenda-setting
study, multiple measures over time give researchers more evidence to evaluate
changes in one of the agendas being studied.
Kepplinger, Donsbach, Brosius, and Staab (1989) suggested that time
series analysis can assist in investigating the ageless question of if the press
drives or follows public opinion. Indeed, Watt and van den Berg (1978) provide
the theoretical underpinnings of such an approach to the relationship between
public opinion and the mass media. Kepplinger et al. (1989) rely on Watt & van
den Berg (1978) in identifying and extending four theories about this
1) The mass media create public opinion, hence,
the correlation between media content at one time will be
strongly correlated with public opinion at a later time than
2) The mass media mirror public opinion, meaning
that the correlation between media and public agendas at the
should be stronger than any other correlation.
3) Public opinion creates media coverage,
implying that the correlation between public opinion at one
be more strongly correlated with public opinion at a later
any other correlation.
4) There is no significant correlation between
media coverage and public opinion.
Gonzenbach (1992) applied these four theoretical approaches to the
relationships among the agendas of the president, the public, and the media.
Meanwhile, Rogers, Dearing, and Chang (1991) utilized time series models to
investigate the inter-relationships of real-world cues, the science agenda, the
media agenda, the polling agenda, and the policy agenda for the public issue of
AIDS. The present study draws upon this tradition by investigating the
relationship between the agendas of the press, the president, public opinion,
and real-world cues over time within the context of the crime issue.
Public Opinion and the Mass Media
There are a number of indications that public opinion and the mass
media mutually influence each other. Erbring et al. (1980) found that the
effects of the public agenda on the media agenda gradually develop over a long
time. However, the media agenda impacts public opinion more quickly. Although
Behr and Iyengar's (1985) experimental manipulation indicated that public
concern regarding unemployment and energy did not drive media coverage, public
attention may have modestly increased coverage of inflation. Similarly, Rogers
and Dearing (1987) found that there is a "two-way mutually dependent
relationship between the public agenda and the media agenda" (p. 571).
Kepplinger et al. (1989) also demonstrate that press coverage about German
Chancellor Helmut Kohl preceded public opinion. Most important, this study
found that media coverage also followed public opinion, albeit with a differing
set of opinions. Gonzenbach (1992) also found a relationship of mutual
influence between the press and the public for the drug issue.
The President and Public Opinion
Public opinion is not solely affected by the mass media. Shaw and
McCombs (1989) argue that political elites, interest groups and major
organizations also affect the public agenda. Behr and Iyengar (1985) note that
the president is the key American political figure and receives the plurality of
media and public attention. As with the interaction of the press and the
public, the president likewise affects and is affected by the press and public
agendas. Converse (1987) claims that public opinion affects the president and
policy-making. The opposite effect can be true as well. The president can
alter the public agenda by advocating certain legislative or moral positions in
oral, written, and symbolic communications (Chapel, 1976). Moreover, the
president's prominence allows the president an opportunity to regularly
communicate a specific agenda (Behr & Iyengar, 1985) which may have a
significant effect on public opinion (Gilbert, 1981).
The President and the Mass Media
In addition to the public, the president similarly interacts with the
press. Press coverage is very often affected by the president's explicated
agenda. Emerson (1987) has argued that the Iran Contra scandal presented a
direct embarrassment to Reagan's war on terrorism. Consequently, press coverage
of terrorism abated with decreased presidential emphasis on the issue
(Jablonski, 1992). Hence, the media can play a significant role in driving
policy making and policy makers can similarly affect the media.
Agenda-building studies, moreover, indicate that media coverage is
largely determined by what issues political elites and the public deem important
(Protess & McCombs, 1991). A bidirectional relationship may exist between the
agendas of the media and the president (Denton & Hahn, 1986). Put simply, the
media can influence the presidential agenda while the president can influence
the media agenda. The media provide the president with exposure to the public
and vice versa (Graber, 1982; Edwards & Wayne, 1985). In addition, the media
follow every move a president makes, adding to the considerable amount of news
coverage devoted to the White House (Orman, 1990). This also enlarges the
typical audience for any one presidential message (Denton & Hahn, 1986).
Furthermore, a president may transmit a list of priorities to the public and
Congress at the start of an administration through the mass media (Light, 1991).
This is an important function, especially if one feels that the role of the
president is to shape national attitudes into a coherent policy (Powell, 1986).
The president, as the chief political figure in America, also
provides news stories. The president may therefore play a major role in
influencing the news media agenda (Gilberg et al., 1980; Lang & Lang, 1981;
Weaver & Elliott, 1985; Turk, 1986; Robinson, 1990). Behr and Iyengar (1985)
note that the president is the key American political figure and receives the
plurality of media and public attention. Levels of press coverage are very
often affected by the president's explicated agenda. Presidents can create news
by holding press conferences and photo opportunities, among other ceremonies
(Ansolabehere, Behr, & Iyengar, 1993). Presidents use pseudo-events and other
communication opportunities to infuse the news with "self-serving commercials"
(Denton & Hahn, 1986, p. 275). Indeed, Behr and Iyengar (1985) found that a
single presidential speech on a given issue can result in as many as ten
stories. These researchers note that when Reagan addressed the economy, lead
stories about the economy increased in the media.
The relationship between political elites and the media has also been
studied in lower levels of public office. Cook et al. (1983) and Protess et al.
(1987) indicate that Chicago area policy elites had modest degrees of attitude
change after being exposed to investigative reports about problems within that
city. Pritchard (1986) reports that prosecutorial decisions are significantly
predicted by the amount of media attention devoted to a certain case. Finally,
Schmid (1992) demonstrates that the president (and executive policy) can be
directly affected by mass media coverage of terrorism. Hence, the media can
play a significant role in driving policymaking while policymakers can similarly
affect the media.
The agenda-building process can work in the other direction as well.
Cobb and Elder (1972) posit that the mass media play a huge role in elevating
issues to the presidential agenda. Therefore, the mass media may directly
affect the presidential agenda. Weaver's (1990) attempt to define the direction
of influence in the press-politician relationship concludes that the linkage's
characteristics are largely case-specific. In some situations, the press have
the upper hand, while in other situations, the politicians drive the agenda.
The relationships described so far do not exist independent of each
other. Indeed, as prior research suggests, the press, the president, and the
public may influence each other's agendas regarding important issues
(Gonzenbach, 1992; Rogers, Dearing, and Chang, 1991). The triangulation of the
press, the president, and public opinion over time can provide the researcher
with a more robust macro-level picture of agenda-setting. As Kosicki (1993)
concludes, "Agenda-setting is one small part of a larger process of
understanding the very complex interrelationships among media organizations,
public opinion, and public policy-making" (p. 117).
This work draws most specifically upon the time series investigations
of agenda-setting by Gonzenbach (1992), Rogers et al. (1991) and Kepplinger et
al. (1989). This study seeks to address three basic questions about the roles
played by the press, the president, and public opinion with regard to the crime
1) Do the mass media influence public opinion or does
public opinion influence the media's agenda ?
2) Does the president influence public opinion, or does
public opinion influence the president's agenda?
3) Does the president's agenda influence the media's
agenda or does the media agenda influence the president's
The study comprises three agenda measures for each month from January
1988 until December 1995. The public opinion measure is the American public's
monthly opinion of crime as the nation's most important problem. The media
agenda consists of articles in The New York Times related to violent crime. The
presidential public relations agenda is operationalized as the monthly frequency
of public statements about crime by the president.
The study combines multiple, similarly worded "most important
problem" survey questions from 14 different public opinion survey organizations
in order to measure the public agenda (for a breakdown, see Table 1). The
percentage of crime being mentioned as the most important problem was tracked
from January 1988 until December 1995. 115 monthly polls are used to determine
public opinion about the crime issue for 61 of the 96 months of the study, while
35 months are estimated by interpolation. Each survey was conducted via
telephone, was based on a national sample of adults, and utilized an open-ended
Of the 115 surveys, 19 accepted multiple responses, while 96 did not.
The results of the 19 multiple response surveys were scaled by dividing the
percentage response by the total percent of the survey. For example, the August
1992 Gallup poll indicates that 7 percent of those surveyed identified crime as
the nation's most important problem. All of the responses in this particular
survey totaled 182 percent. Therefore, the adjustment was calculated by
dividing 182 into 7 (the unadjusted survey percentage). In this example, the
adjusted percentage is 3.8 percent.
The sample sizes for the 115 surveys range from 509 to 5,791 with a
mean of 1324.6 and a standard deviation of 670.9. The mean for the mid-point of
when the individual polls were conducted during each month was 14.5 with a
standard deviation of 8.6 days. Certain months contained more than one most
important problem survey. The mean of the multiple surveys was taken for this
type of month.
Use of Most Important Problem Polls by Organization
Organization Number of Polls Used
CBS/New York Times 23
ABC/Washington Post 14
Los Angeles Times 13
CNN/USA Today 7
Princeton Survey Research Associates 7
The Wirthlin Group 6
Washington Post 5
Associated Press/Media General 2
Institute for Communication Research 1
New York Times 1
Public Opinion Strategies 1
The New York Times
This study utilizes a content analysis of the media crime agenda in
The New York Times. The New York Times was used because it is the elite
newspaper in the United States and is fairly robust to the proximity effects of
major issues. The unit of analysis was individual articles or stories about
violent crime. The monthly frequency of crime stories and articles was used
primarily because this method was more efficient in terms of time and money than
calculating column inches or minutes of coverage. The use of frequency counts
has been found to be a fairly reliable measure of media content, and is
comparable to the tabulation of column inches (Stone and McCombs, 1981).
The New York Times file within the Lexis/Nexis data service was used.
The Lexis/Nexis search string was, "Crime and national desk or editorial desk
and homicide or murder or violent or violence or rape or robbery and not news
summary and date aft December 1987 and date bef 1996." Only articles in the
national news and editorial sections of The New York Times were recorded.
Articles from local sections (the Connecticut desk, the Metro desk, etc.) were
omitted from the study. In addition, to avoid duplication of articles, news
summary pieces were excluded from the analysis since they merely represent
articles extant in that day's edition. Each story was analyzed to ascertain the
article's relationship to the crime issue in the United States.
Similarly, the presidential agenda is measured via the public
relations efforts of the presidents. The Public Papers of the Presidents was
utilized to determine the monthly frequency of mentions of the crime issue by
the sitting president. This resource is also included in the Lexis/Nexis
service. The search string utilized was, "Crime and date aft December 1987."
The retrieved documents were analyzed for their relevance to the research topic.
Finally, the use of frequency counts was decided upon for many of the same
reasons as in the investigation of the media agenda.
This study employs Autoregressive Integrated Moving Averages (ARIMA)
time series analysis as popularized by Box and Jenkins (1976). In the present
study, ARIMA modeling is used to identify each of the univariate time series.
Cross-lagged correlations were then calculated for all permutations of the three
variables: press-public opinion; president-public opinion; president-press
(McCleary and Hay, 1980). 6-month lags were calculated in each direction for
the cross-lagged correlations. While Winter and Eyal (1981) determined that a
4-6 week span is optimal, Stone and McCombs (1981) found that the process of
adapting media changes in the public agenda may take between 2 and 6 months.
Shoemaker et al. (1989) reported that the optimal effects were several weeks and
3 months preceding a poll. Meanwhile, Kepplinger et al. (1989) found optimal
effects 3 months prior to the poll. Finally, Gonzenbach (1992) reported that
the relationship between the press and public opinion is "a two-way street" (p.
141). In other words, support was found for the immediate agenda-setting
effects of the media as reported by Shoemaker et al. (1989) and Erbring et al.,
(1980) in addition to the long-term (6-month) correlations found by Kepplinger
et al. (1989) and Brosius and Kepplinger (1990). Hence, the media affect public
opinion and vice-versa, though to different degrees and perhaps at different
points in time. ARIMA time series is an appropriate method of analyzing the
four univariate time series in this study since there are over 50 monthly
observations (Montgomery, Johnson & Gardiner, 1990). In addition, this method
is an accurate time series analytical tool since it calculates the first- and
second-order autoregression within each series. In other words, for a given
series, the error terms or residuals at different points in time should not be
correlated. When this occurs, the estimated regression model fits the data very
well, since the model underestimates the true variability of the residuals. The
result is a deceptively inaccurate model, which in turn leads to inaccurate
testing of hypotheses. The Durbin-Watson d-statistic can be used to test for
first-order autocorrelation (Freund & Littell, 1991). However, the time series
being studied might be a second-order autoregressive process (SPSS, 1988).
ARIMA analysis is used to determine and model the errors which may make a
time-series largely unpredictable. In effect, ARIMA increases the accuracy of
the forecast model, resulting in more accurate hypotheses testing.
While much of the criticism of ARIMA modeling is appropriate for
forecasting procedures, such as the large number of time points and the
difficulty in refitting the model with the acquisition of new data, this type of
analysis is well-suited to studies which use historical data and examine only a
few series (Montgomery, Johnson, & Gardiner, 1990). There are, however,
drawbacks to this approach. Since this study draws bivariate comparisons
between the series, the simultaneous effects of third variables cannot be
determined, much less controlled (Fan, 1988).
The data indicate that crime stays relatively stable in prominence
across measure until July, 1993. Press coverage, presidential attention, and
public opinion all show signs of increase at this time. Notably, the increases
are dramatic for the public and press agendas, while the presidential agenda
measure rises more gradually.
Crime was not viewed as a major problem from 1988 until mid-1993 as
evidenced by the responses to "most important problem" survey questions. Crime
fears peak at around 5 percent in the first half of each year studied until
1993. The effect of certain incidents, such as the Los Angeles. riots in May,
1992 may explain this phenomenon (see Figure 1). Notably, crime does not reach
the 6 percent level until May, 1993. During this time, other issues were
certainly important, such as health care, gays in the military, the budget, and
the situations in Bosnia and Somalia. However, crime concerns expand rapidly in
October, peaking in December with 20 percent. Several press accounts
attribute the October-November-December expansion to several high-profile
crimes, such as the kidnapping and murder of Polly Klaas in California and the
mass murder on the Long Island Railway in December, 1993.
While concern for crime remained higher than usual through 1995, that
concern appears to be a decreasing trend since the peak months during 1993 and
1994. Other issues have attenuated the supremacy of crime on the nation's
important problem agenda. Specifically, concern regarding the federal budget
deficit, the economy, and unemployment increase throughout 1995. Granted, crime
is still identified as the most important problem by a plurality of respondents,
but other issues appear to have gained ground by the end of 1995.
The New York Times
Crime articles in The New York Times resemble an increasing linear
trend, overall, between 1988 and 1995. Moreover, monthly frequencies of
articles about crime before 1993 have intermittent peaks. Several of the cycles
seen in the public opinion measure seem to exist for The New York Times' crime
coverage. However, press coverage of crime reaches unprecedented levels
starting in August of 1993. Several events occur at this time: the Rodney King
and Reginald Denny beating trials in Los Angeles and the murders of foreign
tourists in Florida. While press coverage abates in September, it expands once
again in the last 3 months of 1993, the result of still more events. Press
coverage remains at the same basic level during 1994, and, arguably, for much of
1995. However, the high degrees of coverage between August 1993 and August 1994
do not occur during 1995. This may have some relationship to the similar
decrease in concern for crime in the public agenda at this time. In addition,
1994 may have had more prominent crime stories than 1995. There were a number
of events in 1994 which were covered by the media, including the Susan Smith
infant drownings in South Carolina, a number of homicides committed by pre-teens
in Chicago, and the Nicole Simpson-Ronald Goldman murder investigation.
Presidential attention to crime, as measured by the Public Papers of
the Presidents, is interesting in that mentions of crime increase in October of
each election year analyzed by the study (1988, 1990, 1992, 1994). The highest
frequency of presidential crime emphasis occurs in October 1992 while the lowest
total occurs 2 months later. In the latter half of 1993, however,
presidential attention increases in August, but falls in September. Several
issues, such as health care and the budget battle, served to draw attention away
from crime. Crime seemed to be back on the presidential public relations agenda
in October and November, 1994, partly a result of the debate and subsequent
mid-November vote on the Brady gun control bill. However, presidential mentions
of crime decrease throughout 1995. Again, this trend appears to mirror what is
found with the other two agendas: Attention to the crime issue peaks in 1993,
but decreases by the end of 1995.
Univariate Series Analysis
The nature of the three time series studied was in part explicated
with the ARIMA analysis. Table 2 includes the ARIMA model terms for each
series, along with the estimates of how well each model fits the corresponding
data. The public opinion time series is an integrated series, needing only to
be differenced (0,1,0). The model for The New York Times indicated a moderate
first-order autoregressive process and a differenced, seasonal moving average
process: (1,0,0) (0,1,1). The seasonality of this series was 12 months.
Although traditional Box-Jenkins ARIMA analysis would dictate dropping the
seasonal moving average component to the New York Times series (in light of its
.719 probability), the Akaike Information Criterion (AIC) and the Schwartz
Bayesian Criterion (SBC) indicated that the model fit the data very well. The
SPSS Trends manual notes that a researcher may choose to ignore a probability
estimate if the AIC and the SBC indicate the model can be accepted (SPSS, 1988).
Finally, the ARIMA model for the presidential time series indicated a strong
first order autoregressive process: (1,0,0).
ARIMA Analysis of Univariate Series
Series Model Coefficients t-ratio p
Public Opinion (0,1,0)
New York Times (1,0,0) (0,1,1) 12 f1 = .39 4.38 .000
q12 = .99 0.36 .719
President (1,0,0) f1 = .54 6.18 .000
ARIMA Cross-Correlation Analysis
Once the ARIMA model for each univariate series is determined, it is
necessary to examine the bivariate relationships among the agendas of the press,
the public, and the president. The bivariate analysis takes each modeled series
and examines the cross-lagged correlations for each possible paired combination
(Gonzenbach, 1992; Mark, 1979; Vandaele, 1983). Table 3 depicts the lagged
correlations for each of the three possible relationships in this study. With
regard to crime, public opinion preceded coverage in The New York Times, with
significant correlations at 2 and 5 months. The cross-correlation analysis did
not indicate the existence of a reversal of this process. Hence, public opinion
about crime plays a significant role in affecting the crime agenda of The New
York Times, but not vice-versa. Perhaps the press reacts to the news desires of
the reader by tailoring coverage to the market. On the other hand, the press
may be ferreting out newsworthy stories independent of a marketing concern.
Overall, however, the results suggest that the relationship between the press
and the public appears to be relatively one-sided as far as the crime issue is
Public opinion and the presidential public relations agenda regarding
the crime issue were significantly correlated at lag +5 only. This may very
well be a spurious effect since it comes, quite literally, from out of the blue:
No hint at such an effect at surrounding lags exists. An important note should
be made, however. Gonzenbach's (1992) similar examination of the drug issue
reported a nearly identical result in both degree and time. Hence, the
president may indeed be affected by the public or the press after a period of
time, but the dynamics behind such a relationship, if it indeed exists, defy
easy explanation. The presidential and public agendas regarding crime do not
appear to drive each other, as far as monthly time-lags are concerned. Perhaps
altering the lag periods to smaller units would uncover the nature of the
relationship. However, this is impossible with current public opinion measures.
The relationship between the press and the president was significant
at lag +2. This indicates that the president tends to follow the agenda of the
press. The result for lag +2 may be used as evidence of the president following
the press at a later date. This result appears to be a good example of
agenda-building. The president seems to react to the media agenda regarding
crime. No other correlations are significant.
The cross-correlation analysis presented here implies an intriguing
agenda-building/agenda-setting process for the crime issue in the United States
between 1988 and 1995. This study indicates that the press does follow public
opinion. A traditional agenda-setting effect was not detected in this study.
The lack of effects from The New York Times on crime-related public agenda may
be explainable due to the fact that fear of crime is one of the more obtrusive
issues in contemporary society. The public does not necessarily need to rely on
news to be concerned with crime. Direct experience with crime and non-news
media programming may be enough to spur concern with crime.
The effect of public opinion on the media does not end in a vacuum,
however. This study presents evidence that the press also serves an
agenda-building function by driving the importance of crime on the presidential
agenda. Indeed, the president appears to follow The New York Times crime agenda
at 2-month intervals. These two results imply an intriguing relationship, with
the public's concern about crime spurring media coverage, which in turn drives
the presidential agenda (Figure 2). This study fails to isolate what drives
public opinion about crime, however. Nevertheless, a number of sources cite
increased crime coverage on television news. Indeed, a recent study by the
Center for Media and Public Affairs indicates that crime coverage on television
news has quadrupled in the 1990s, despite an overall decrease in the actual
crime rate (Whitman, 1996). Perhaps public attention to the crime issue is
driven by the media after all, just not The New York Times. While the
relationship depicted in Figure 2 is tantalizing, the cross-correlation results
fail to indicate that the public agenda is significantly correlated with the
presidential agenda over time, except for the (questionable) result at lag +5.
Setting the Crime Agenda Between 1988 and 1995
Public Opinion -----> The New York Times -----> Presidential Agenda
The three research questions posed by this study called for an
investigation into each possible relationship among the crime agendas of the
president, the press, and the public, specifically: 1) do the mass media
influence public opinion or does public opinion influence the media agenda?; 2)
does the president influence public opinion, or does public opinion influence
the president's agenda?; 3) does the president's agenda influence the media
agenda or does the media agenda influence the president's agenda? This
investigation indicates that public opinion about crime appears to drive The New
York Times' crime coverage. The present analysis also fails to find a
significant relationship between the agendas of the public and the president
regarding the crime issue. Finally, it appears that the presidential crime
agenda is driven by the press. Notably, this study fails to find evidence of a
"two-way street" regarding the crime issues for any of the relationships among
the three agendas.
Future research should address a number of the limitations
encountered by this study. Obviously, the nature of correlational data preclude
making causal statements about the agenda relationships reported here. The
effect of confounding variables is relatively unknown. Indeed, there is good
reason to suspect that other forms of mass communication may have an effect on
public opinion. This study cannot control for those effects. In addition, the
bivariate cross-correlation analysis similarly cannot control for the combined
effects of agendas. In other words, the combined effects of the presidential
and press agendas may result in a different picture of influence on the public
Future research should also utilize real world measures of the crime
problem not to add to the long line of findings demonstrating the discrepancy
between crime news and crime reality, but instead to factor in the degree to
which the measures used in this study are "event driven." The concern of the
public about crime may be a series of reactions to high-profile crimes, such as
the Polly Klaas case, the O.J. Simpson murder trial, the Susan Smith case, and
the Menendez brothers trial. In addition, the public may take their cues from a
direct experience with crime, fear of crime, or with depictions of crime in the
The effect of television news coverage of crime should also be
investigated using similar time series modeling procedures utilized in this
study. The inclusion of a television news measure in such a study might
indicate that the public indeed takes their cues from the electronic media,
while the print media may be influenced by public attitudes and perceptions
about the crime issue. It seems logical that the effects of crime coverage in
the newspaper would not be as strong as similar coverage depicted on television.
Notably, O'Keefe and Reid-Nash (1987) identified television news as having a
greater fear effect than newspaper coverage. Perhaps a study employing a
television news measure would show a stronger agenda-setting effect regarding
the crime issue in addition to the apparent agenda-building effect detected
The increase in public concern about crime in the last half of 1993
came at a time when most measures of the problem indicated that crime was
actually decreasing. While public concern for crime has receded from the high
levels of 1993 and 1994, the percentages of respondents identifying crime as the
most important problem is still higher than in the early 1990s. Crime has not
fallen off of the public agenda, although it does not appear to be as important
by the end of 1995 as it was in 1993. This study may be more important for what
it says about newspaper coverage of the crime issue than for what it implies
about public opinion processes. Perhaps crime coverage in elite newspapers is
geared more to responding to the social issue of crime, rather than to the
individual instances of crime themselves. In this view, increased public
concern for crime may receive more media scrutiny, in turn driving the
presidential agenda. Fear of crime does indeed defy logical, objective measures
of the crime problem. Answering the question of what drives that fear remains a
poignant issue for future research.
Ansolabehere, S., Behr, R., & Iyengar, S. (1993). The media
game: American politics in the television age. New York: Macmillan.
Behr, R. L. & Iyengar, S. (1985). Television news, real-world
cues, and changes in the public agenda. Public Opinion Quarterly, 49, 38-57.
Box, G. E. P. & Jenkins, G. M. (1976). Time-series analysis:
Forecasting and control. San Francisco: Holden-Day.
Brosius, H-B. & Kepplinger, H. M. (1990). The agenda-setting
function of television news: Static and dynamic views. Communication Research,
Brown, S. & Pasternak, J. (1994, February 13). A nation with
peril on its mind. Los Angeles Times, p. A1.
Chapel, G. W. (1976). Speechwriting in the Nixon administration.
Journal of Communication, 26, 165-172.
Cobb, R. W. & Elder, C. D. (1972). Participation in American
politics: The dynamics of agenda-building. Baltimore: Johns Hopkins
Cohen, B. (1963). The press and foreign policy. Princeton:
Princeton University Press.
Converse, P. (1987). Changing conceptions of public opinion in
the political process. Public Opinion Quarterly, 51, s12-s24.
Cook, F. L., Tyler, T. R., Goetz, E. G., Gordon, M. T., Protess,
D., Leff, D. R., & Molotch, H. L. (1983). Media and agenda setting: Effects on
the public, interest group leaders, policy makers and policy. Public Opinion
Quarterly, 47, 16-35.
Davis, F. J. (1952). Crime news in Colorado newspapers.
American Journal of Sociology, 57, 325-330.
Denton, R. E. & Hahn, D. F. (1986). Presidential communication:
Description and analysis. Westport, CT: Praeger.
Edwards, G. C. & Wayne, S. J. (1985). Presidential leadership:
Politics and policymaking. New York: St. Martin's Press.
Einsiedel, E. F., Salomone, K. L., & Schneider, F. P. (1984).
Crime: Effects of media exposure and personal experience on issue salience.
Journalism Quarterly, 61, 131-136.
Emerson, S. (1987, February 16). Arms for hostages deal:
Dogging the money trail. U.S. News and World Report, 102, p. 21.
Erbring, L., Goldenberg, E. N., & Miller, A. H. (1980).
Front-page news and real-world cues. American Journal of Political Science, 24,
Erskine, H. (1974a). The polls: Causes of crime. Public
Opinion Quarterly, 38, 288-298.
Erskine, H. (1974b). The polls: Control of crime and violence.
Public Opinion Quarterly, 38, 490-502.
Erskine, H. (1974c). The polls: Fear of violence and crime.
Public Opinion Quarterly, 38, 131-145.
Erskine, H. (1974d). The polls: Politics and law and order.
Public Opinion Quarterly, 38, 623-634.
Eyal, C. H., Winter, J. P., & DeGeorge, W. F. (1981). The
concept of time frame in agenda-setting. Mass Communication Review Yearbook 2,
Fan, D. P. (1988). Predictions of public opinion from the mass
media: Computer content analysis and mathematical modeling. Westport, CT:
Fishman, M. (1978). Crime waves as ideology. Social Problems,
Freund, R. J. & Littell, R. C. (1991). SAS system for
regression. Cary, NC: SAS Institute, Inc.
Gilberg, S., Eyal, C., McCombs, M. E., & Nicholas, D. (1980).
The State of the Union Address and the press agenda. Journalism Quarterly, 57,
Gilbert, R. E. (1981). Television and presidential power.
Journal of Social Political and Economic Studies, 6, 75-93.
Gonzenbach, W. J. (1992). A time-series analysis of the drug
issue, 1985-1990: The press, the president and public opinion. International
Journal of Public Opinion Research, 4, 126-147.
Graber, D. A. (1979). Is crime news coverage excessive? Journal
of Communication, 29, 81-92.
Graber, D. A. (1982). Introduction: Perspectives on
presidential linkage. In D. A. Graber (Ed.), The president & the public.
Philadelphia: Institute for the Study of Human Issues, pp. 1-14.
Hindelang, M. J. (1974). Public opinion regarding crime,
criminal justices, and related topics. Journal of Research in Crime and
Delinquency, 11, 101-116.
Jablonski, P. M. (1992). Terrorism and the United States
government: The effects of official definition and agenda-setting on media
coverage. Unpublished Master's Thesis. The Florida State University,
Jones, E. T. (1976). The press as metropolitan monitor. Public
Opinion Quarterly, 40, 239-244.
Katz, J. (1987). What makes crime 'news'? Media, Culture, and
Society, 9, 47-75.
Kepplinger, H. M., Donsbach, W., Brosius, H., and Staab, J. F.
(1989). Media tone and public opinion: A longitudinal study of media coverage
and public opinion on Chancellor Kohl. International Journal of Public Opinion
Research, 1, 326-342.
Kosicki, G. M. (1993). Problems and Opportunities in
Agenda-Setting Research. Journal of Communication, 43, 100-127.
Lang, G. E. & Lang, K. (1981). Watergate: An exploration of the
agenda-building process. In G. C. Wilhoit and Harold de Bock (Eds.), Mass
Communication Review Yearbook 2, pp. 447-468. Newbury Park, CA: Sage.
Light, P. C. (1991). The president's agenda. Baltimore: Johns
Hopkins University Press.
Lippmann, W. (1922). Public Opinion. New York: Macmillan.
Mark, M. M. (1979). Inferring cause from passive observation.
In T. D. Cook & D. T. Campbell (Eds.), Quasi-experimentation: Design & analysis
issues for field settings. Boston: Houghton Mifflin Company.
McCleary, R. & Hay, R. A. (1980). Applied time series analysis
for the social sciences. Beverly Hills, CA: Sage.
McCombs, M. E. & Shaw, D. L. (1972). The agenda setting function
of mass media. Public Opinion Quarterly, 36, 176-187.
Montgomery, D. C., Johnson, L. A., & Gardiner, J. S. (1990).
Forecasting and Time Series Analysis. New York: McGraw-Hill.
Morin, R. (1994, January 30). Crime time: The fear, the facts.
The Washington Post, p. A1.
O'Keefe, G. J. & Reid-Nash, K. (1987). Crime news and real-world
blues. Communication Research, 14, 147-163.
Orman, J. (1990). Presidential accountability: New and
recurring problems. Westport, CT: Greenwood Press.
Payne, D. E. & Payne, K. P. (1970). Newspapers and crime in
Detroit. Journalism Quarterly, 47, 233-238.
Powell, J. (1986). The relationship of the president and the
press. In K. W. Thompson, (Ed.), Presidents, prime ministers, and the press
(pp. 19-38). Lanham, MD: University Press of America.
Pritchard, D. (1986). Homicide and bargained justice: The
agenda-setting effect of crime news on prosecutors. Public Opinion Quarterly,
Protess, D. L., Cook, F. L., Curtin, T. R., Gordon, M. T., Leff,
D. L., McCombs, M. E., & Miller, P. (1987). The impact of investigative
reporting on public opinion and policy making. Public Opinion Quarterly, 51,
Robinson, C. J. (1990). Mass media and the U. S. presidency. In
J. Downing, A. Mohammadi, and A. Sreberny-Mohammadi (Eds.), Questioning the
media: A critical introduction (pp. 88-99). Newbury Park, CA: Sage.
Rogers, E. M., & Dearing, J. W. (1987). Agenda-setting research:
Where it has been , where is it going? In James A. Anderson (Ed.),
Communication Yearbook 11. Newbury Park, CA: Sage, pp. 555-94.
Rogers, E. M., Dearing, J. W., & Bregman, D. (1993). The anatomy
of agenda-setting research. Journal of Communication, 43, 68-84.
Rogers, E. M., Dearing, J. W., & Chang, S. (1991). AIDS in the
1980s: The agenda-setting process of a public issue. Journalism Monographs,
Schmid, A. P. (1992). Terrorism and the media: Freedom of
information vs. freedom from intimidation. In Lawrence Howard (Ed.), Terrorism:
Roots, impact, responses. New York, NY: Praeger. pp. 95-117.
Shaw, D. L. & McCombs, M. E. (1989). Dealing with illicit drugs:
The power-and limits-of mass media agenda setting. In Pamela J. Shoemaker
(Ed.), Communication campaigns about drugs: Government, media and the public.
Hillsdale, NJ: Lawrence Erlbaum Associates, pp. 67-80.
Sheley, J. F. & Ashkins, C. D. (1981). Crime, crime news, and
crime views. Public Opinion Quarterly, 45, 492-506.
Shoemaker, P. J., Wanta, W., & Leggett, D. (1989). Drug
coverage and public opinion, 1972-1986. In Pamela J. Shoemaker (Ed.),
Communication campaigns about drugs: Government, media and the public.
Hillsdale, NJ: Lawrence Erlbaum Associates, pp. 67-80.
Smith, K. A. (1987). Effects of Newspaper Coverage on Community
Issue Concerns and Local Government. Communication Research, 14, 379-395.
SPSS, Inc. (1988). SPSS Trends. Chicago: SPSS.
Stone, G. C. & McCombs, M. E. (1981). Tracing the time lag in
agenda-setting. Journalism Quarterly, 58, 151-155.
Turk, J. V. (1986). Information subsidies and media content: A
study of public relations influence on the news. Journalism Monographs, 100.
Vandaele, W. (1983). Applied time series and Box-Jenkins models.
Orlando, FL: Academic Press.
Watt, J. H. & van den Berg, S. A. (1978). Time series analysis
of alternative media effects theories. In B. D. Reuben (Ed.), Communication
Yearbook 2 , New Brunswick, NJ: Transaction Books, pp. 215-224.
Weaver, D. H. (1990). Setting political priorities: What role
for the press? Political Communication and Persuasion, 7, 201-211.
Weaver, D. H. & Elliott, S. N. (1985). Who sets the agenda for
the media? A study of local agenda-building. Journalism Quarterly, 62, 87-94.
Whitman, D. (1996, March 4). A bad case of the blues. U.S. News
& World Report, 120, pg. 54.
Winter, J. P. & Eyal, C. H. (1981). Agenda setting for the civil
rights issue. Public Opinion Quarterly, 45, 376-383.
 This may be due to the fact that George Bush was a lame duck
president in December, 1992 and had reduced his public appearances and
statements during the transition to the Clinton Administration.
Cross-lagged Correlations of Public Opinion, Presidential Agenda and
New York Times Coverage
Precedes Synchronous Follows
NY Times .00 .18* -.16 .11 .19* -.02 .16
.06 -.01 .16 -.01 -.06 -.03
President -.02 -.10 .05 -.10 -.14 .11 .09 -.02
.17 -.10 -.07 .27* .09
Lag -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4
Precedes Synchronous Follows
NY Times .09 -.08 .07 .10 -.05 .14 .01 -.10 .22* .09
.08 .13 -.06
Lag -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4