Content-Type: text/html

     Fearing the "Mean World"
 
 
 
 
 
 
 
 
 
 
 
Fearing the "Mean World":
Exploring the Victim-Offender Relationship's
Influence on Fear of Violent Crime
 
 
 
James McQuivey
S.I. Newhouse School of Public Communications
Syracuse University
 
 
 
215 University Place
Syracuse, NY  13215
315 443-1922 phone
315 443-3946 fax
<a href="/cgi-bin/wa?LOGON=A3%3Dind9710B%26L%3DAEJMC%26E%3D7BIT%26P%3D345206%26B%3D--%26T%3DTEXT%252FPLAIN%3B%2520charset%3DUS-ASCII" >[log in to unmask]</a>
 
 
 
 
 
PLEASE CONSIDER THIS PAPER FOR YOUR GRADUATE STUDENT PAPER AWARDS
 
     Fearing the "Mean World"
 Abstract
Current cultivation research is expanded to include a distinction among
different types of violent crime and their potential cultivation impact.  It is
theorized that there are cognitive reasons for a difference in how "mean" people
feel the world is when exposed to crime-filled media content.  It is
hypothesized that the crime type people are most significantly impacted by is
stranger-perpetrated violent crime which is more likely to arouse fear than
acquaintance-perpetrated crime.
 
 Fearing the "Mean World": Exploring the victim-offender relationship's
influence on fear of violent crime
Introduction
     In this decade, violence has become a concern for so many people that it
was ranked by a majority of respondents in a 1994 Gallup Poll as the number one
problem facing our country (Gallup, 1995).  Crime is a real problem, especially
to its victims.  But is crime as serious as our fear of it would indicate?
Research on this point has consistently shown that people are unable to estimate
the extent of crime in society and that the variations in such estimates are
independent of the actual crime rate itself (Gilliam, Iyengar, Simon & Wright,
1996).
     Closer examination of the Gallup data shows that until 1994, crime was
almost never listed as the number one problem in the country by more than 10% of
the population.  In 1994 the numbers jumped to 37%, then to a record 52% only to
settle back in at 27% in 1995.  (For comparison, the only other issue to hold
more than 50% of the public's concern in Gallup polling was the drug issue in
1989.)  During this same time period, crime--violent and nonviolent--has
actually fallen across the country.  For example, the National Crime
Victimization Survey (NCVS) of the Bureau of Justice Statistics reported in 1992
that violent crime had fallen 9% from a peak in 1981 (US Department of Justice,
1994, p. 1).  The Uniform Crime Report of the FBI confirms this finding,
reporting that the 1994 overall crime rate was 8% lower than the 1990 crime rate
(US Department of Justice, 1995).  This is curiously the same year that the
Gallup Poll reported the highest rates of concern for crime.
     If actual crime rates do not influence perception of crime, then what does?
Media researchers have long proposed that the mass media influence our
perception of crime.  But close inspection of this influence has yielded varied
results and produced a multitude of conditions under which certain effects may
or may not occur.  One area that has not been investigated is the difference
between perceptions of crime committed by strangers and crime committed by
acquaintances of victims.  The NCVS reports that in 53% of violent crimes the
offender is a stranger to the victim (US Department of Justice, 1996, p. 7).
That means almost half of violent crimes involve offenders who know or are
related to their victims.  This fact implies a qualitative difference in the
nature of such crime.  When people are asked to estimate the frequency of
violent crime, do they account for this qualitative difference or do they merely
apply a mental image of a certain type of crime, whether stranger-perpetrated or
acquaintance-perpetrated?  Is one type of crime more likely to be misperceived
and feared than the other?  Does exposure to news coverage or violent crime
affect how people perceive the two different types of crime?  This paper
undertakes just such an investigation, examining the media's influence on fear
of stranger-perpetrated crime.
Theory
     Media effects researchers have been concerned with the question of violent
crime and violent crime reporting since the earliest days of the discipline.  In
the entertainment media, the concern has been with the effects of violent
portrayals on the viewer or reader (Lowery & DeFleur, 1988).  In the news media,
there has been concern over the exaggerated coverage violent crime receives
(Johnstone, Hawkins, & Michener, 1995).  The famed line, "If it bleeds, it
leads," is a well-known maxim for what determines newsworthiness of crime.  If
crime reporting overplays the dangers of violent crime, these researchers reason
that the public may have an exaggerated fear of crime.
Cultivation Theory
     The best known field of research that has studied the effects of violent
television programming on fear of crime has been cultivation theory, which first
posited in the '70s that exposure to television could cultivate in viewers an
exaggerated view of their own chances of being victims of crime (Gerbner &
Gross, 1976).  Initial findings were very positive, only to be challenged later
on methodological grounds (Hirsch, 1980; 1981).  In general, the original
success of cultivation theory was diminished as more research established the
myriad of controls that had to be examined before a cultivation effect could
reliably be measured (Potter, 1994).  The results of such efforts, though still
championed by advocates of the theory, have proven to be very small (Gerbner,
Gross, Morgan & Signorielli, 1994).
     Cognitive vs. affective.  An early attempt to improve upon cultivation
research is the distinction between different levels at which cultivation can
theoretically take place (Hawkins, Pingree, & Adler, 1987).  At the cognitive
level, it is theorized that exposure to mediated violence could lead viewers to
consciously  overestimate the rate of violent crime or the prevalence of law
enforcement officials.  This is differentiated from the affective level, where
it is claimed that exposure to mediated violence could lead viewers to fear
crime and overestimate their own chances of being a victim of crime.  Research
into this differentiation has shown that television exposure does often
cultivate exaggerated perceptions at the cognitive level, but not at the
affective level (Hawkins & Pingree, 1990).
     Extended cultivation.  Other researchers have extended cultivation research
by narrowing the essentially mass audience perspective of the original theory
(Hawkins & Pingree, 1981).  This body of extended cultivation research has
sought to link specific program viewing to a cultivation-like effect, although
Gerbner and his colleagues do not recognize the extended cultivation research as
a true test of cultivation (Gerbner, 1990).  As part of this extended
cultivation research, news programming has often been isolated as a potential
contributor to a mean-world syndrome effect (McLeod, 1995).  This is because
news content is "real" where fictional portrayals are understood by the audience
to be "unreal."
News Content
     The question of whether news content may influence perception of violent
crime inevitably leads to the question of where news content comes from.  If
news coverage is to be blamed for distorting the public's perception of violent
crime, there must be a theoretical explanation for why such coverage is
emphasized.  Shoemaker (1987) summarized the relevant theories of news content.
Those most relevant to crime reporting posit that news content is a function of
newsgathering routines, is a product of the socialization patterns and norms of
the reporters, is the result of the decisions made by editorial "gatekeepers,"
or is shaped by the societal forces that exert pressure on the news
organizations to produce certain types of content.  Within these theories,
researchers have claimed that crime reporting overplays the most deviant violent
acts because they provide good visuals for television or print coverage; because
they involve weak victims such as women, children, and the elderly; or because
they involve celebrities (Johnstone, Hawkins, & Michener, 1995).
     The value of news content theories depends on the extent to which news
coverage actually influences people.  One apparent anomaly which weakens that
assumption is the finding of a recent study that violent crime is still
considered a serious concern for people even though recent press articles and
stories had focused on the decreasing crime rate (McLeod, 1995).  How is it
possible that violent news content could create a negative effect on the public,
but positive news about crime has no effect?  Is there a theoretical
justification for how individuals process news information that accounts for
this apparent contradiction?
Cognitive Processing
     In order to explain how people are affected by crime news, a cognitive
processing model would have to explain why individuals would overperceive news
about violent crime yet disregard positive news about a declining crime rate.
One model proposed by Wyer and Srull (1980) for use in social cognition includes
functional processors in the cognitive system that assign incoming information
to their ultimate destination within the system.  As applied to communication
studies by Reeves, Chaffee, and Tims (1982), these processors include a
preencoder, an executive unit, an encoder/organizer, an integrator, and a
response selector.  As new information enters the system, it is filtered by the
preencoder into the most relevant bits.  The executive unit routes the flow of
different information bits into different internal processing units where the
encoder/organizer compares the bits to knowledge already processed and
accumulated, ultimately pairing the information with those concepts or schemata
that are judged to be relevant to the new information.  The integrator
identifies the potential impact of the information on the existing framework and
encourages the response selector to choose an appropriate output mode.
     Upon entering this system, news information might be given a higher
priority by the preencoder than entertainment information because it is
perceived as "real."  This higher priority would enhance the cognitive processes
all the way down the line, encouraging more specific processing at all stages.
At the encoder/organizer stage, a judgment must be made concerning the relevance
of the new information.  The integrator would translate the emphasis placed on
the news information into an indicator of seriousness, encouraging the response
selector to increase general fear of violent crime in the system so that the
individual can avoid danger.  Specific responses may include locking car doors
while driving down city streets, purchasing an alarm system or carrying mace.
Interestingly, the result of such activities would be an increase in salience of
crime for the system, further signaling the preencoder to be on the watch for
information about violent crime.
     It should be cautioned that an elaborately constructed model like this is
elegant and well-reasoned, but assumes that all thought processes are nearly
rational or at least ordered.  Nor is it based on an understanding of the
biological processes involved in information processing in the brain.  Recent
research that has suggested an important role for the biological design of the
brain itself was conducted by Shoemaker (1996).  Based on a theoretical
discussion of cultural and biological evolution Shoemaker suggested there may be
a biological predisposition for news-seeking in the brain itself.  Our current
taste for news about our environment is related to the early hominid's desire to
know the location of potential prey as well as potential predators.  Strictly
applying evolutionary concepts, this means that those individuals who constantly
monitored their environments for change survived as a result and passed that
behavior to their children.  Such conditioning, implemented over time has
instilled in us today a cultural predisposition for the environment-monitoring
function which the news media currently fulfill.
     Such a profound biological operant should be measurable not only through
observing behavior, but should be logically consistent with the structure of the
human brain as we understand it.  Decades of painstaking neuroscientific
research combined with recent technological advances in Positron Emission
Tomography (PET) have provided us a very detailed understanding of how many
parts of the brain work.  At the cellular level, research into the neuron--the
essential unit of brain and nervous system functioning--has yielded much of
value (Kandell, Schwarz & Jessell, 1991).  For example, we now understand that
brain cells communicate to each other through intricate connections, with each
cell synapsing on (connecting with) up to 10,000 other cells.  These connections
are not formed randomly, but follow specific growth patterns which are first
genetically determined but that are then either encouraged by the mental
processes of the individual or discouraged.  In other words, there may be a
genetic basis for certain information processing patterns.  However, once
formed, neuronal connections will atrophy and often die if the mental processes
of the individual don't "exercise" that pattern of connections.  This indicates
that the presence of certain patterns of neuron firings (whether those involved
in muscle coordination or in conscious thought) will encourage the growth of
that pattern.  The processing of information about violent crime could be such a
pattern.  This could partially explain why positive information about crime is
ignored, since the pattern has grown around retaining and processing negative
information.
     At the brain level, there are further indications that information about
violent crime would receive special treatment by the brain.  In the brain,
researchers have identified three different evolutionary structures (Kandell,
Schwarz, & Jessell, 1991).  The first section, the brain stem, is common to
almost all creatures.  It is theorized that this section of the brain evolved
first.  The second section is called the limbic, or reptilian, brain.  Only
vertebrate species have developed this section of the brain.  The most recent
evolutionary contribution to the brain is the cortex.  The human cortex is
highly developed and produces the sensations we call consciousness.  In it are
contained millions of neurons which dedicate themselves to processing,
integrating, and evaluating sensory input.  However, much of what we actually do
is determined not at the cortical level, but below it at the limbic and brain
stem level.  For example, the brain stem is responsible for autonomic processes
which we don't think about often.  This includes heart beat, respiration, and
glandular secretions.  Up one step at the limbic level are contained most of the
emotional processes that we experience.  At the center of this limbic brain are
the thalamus and the hypothalamus, two substructures which filter nearly all
inputs to the cortex.  These two substructures are responsible for
managing--among other things--two of our most basic emotions: fear and sexual
arousal.
     As sensory input proceeds through the limbic brain, it is evaluated (much
like the preencoding function of the cognitive model described above) by the
hypothalamus for relevance to our emotional state.  Inputs that signal potential
danger result in the release of hormones and other secretions that alter the
physiological state of the body.  The information is passed further up the brain
into the cortex where the sensations of fear or worry are felt in the cortex.
In situations of extreme danger, this fear is used by the body to keep the mind
alert and aware.  In situations of remote danger, this fear could be used by the
mind to tag selected pieces of information and store them for future use.  For
example, knowledge about a particular park where a violent assault took place
may be stored by the brain to be recalled whenever the person walks through that
park.  This can be regarded as a defense mechanism designed to ensure that the
individual will be alert to any signs of danger in the future.
Theoretical Synthesis
     What are the implications of this brain research on this paper?  Revisiting
the basis for cultivation theory, it shows that there may be a biological basis
(whether perpetuated culturally or genetically) for the media's ability to
cultivate fearful reactions in people.  However, it also exposes some subtle
weaknesses in the cultivation concept.  First, only those messages which contain
legitimate potential threats to the safety of the individual are likely to
produce a fear reaction with lasting effects.  This could exclude fictional
portrayals.  While the limbic brain can't distinguish between fantasy and
reality, the cortex can, thus shutting down the lower system's fear reaction and
eventually stopping it from occurring under similar circumstances in the future.
This is similar to the concept of desensitization used in media violence and
pornography studies (Eysenck & Nias, 1978).
     The implications are not so clear for the difference between affective and
cognitive cultivation effects.  Based on the model described above, one would
assume that an affective response would be most readily produced, given that the
limbic brain influences the fear emotion directly.  However, one element of the
system is that it loops:  that is, outputs from one iteration become inputs in
the next.  Thus, affective outputs become inputs to the cognitive system,
encouraging the individual to process the reaction consciously.  It is possible
that this could result in greater cognitive perception of crime as well.  If
cognitive processing allows the individual to rationally combat the fear, the
affective reaction may diminish over time, leaving the cognitively elaborated
construct in tact.  This paper proposes that both reactions will be evident in
heavy news media consumers since they are constantly involved in this
affective-cognitive cycle at any one given moment.  Any similar effect for light
news media consumers will be small.
     This raises a further question based on the discussion of offender-victim
relationships presented in the introduction above.  Is it possible that certain
types of crime are more likely to trigger an intense affective reaction?  For
example, if a violent murder is likely to produce a fear reaction, is it as
likely to do so if the violent murder is committed by the husband of the victim
in their own home?  Remember the purpose of the fear reaction is to signal
potential danger to the individual.  For whom does such a spousal murder signal
potential danger?  Only people whose life circumstances are similar in some way
to the victim's.  If this reasoning is correct, only certain people will be
likely to experience the affective or cognitive effect of exposure to violent
news media where the relationship between the victim and the offender is close.
     On the other hand, in the case of violent crimes committed by
strangers--serial killings, the Oklahoma bombing, and some rapes--the limbic
system is likely to respond to such information simply because the victims are
generally seen as unsuspecting.  Much like the saying, "when you least expect it
. . . expect it," this reasoning is irrational but self-perpetuating because
try
as they might, individuals cannot cognitively eliminate the risk of such a
random-seeming occurrence.  Thus, this paper proposes that the fear reaction
proposed above with respect to news exposure will actually be in reaction to
stranger-perpetrated crime, not acquaintance-perpetrated crime.
Hypotheses
     Consistent with the simple cultivation perspective discussed above,
hypotheses one and two will test the viability of the cultivation concept with
respect to the stranger-acquaintance split.  The first hypothesis assesses the
effect of mean television exposure on the cognitive perception (awareness) of
stranger-perpetrated violent crime as follows:
     HI:        The more TV people watch, the more they will be aware of
          stranger-perpetrated violent crime.
     The second hypothesis assesses the effect of mean television exposure on
the affective perception (fear) of stranger-perpetrated violent crime as
follows:
     H2:        The more TV people watch, the more they will fear
          stranger-perpetrated crime.
     Hypotheses three and four test the television news portion of extended
cultivation theory explained above for both the cognitive and affective
perceptions of stranger-perpetrated violent crime.  They are as follows:
     H3:        The more television news people are exposed to, the more they
          will be aware of stranger-perpetrated crime.
     H4:        The more television news people are exposed to, the more they
          will fear stranger-perpetrated crime.
     The logic behind the extended cultivaiton hypotheses does not only apply to
television news however.  If it can be argued that the "real" nature of
television news can create awareness or fear of stranger-perpetrated crime, the
same logic should also apply to other news media.  That leads to the creation of
the next two hypotheses:
     H5:        The more news media people are exposed to, the more they will be
          aware of stranger-perpetrated crime.
     H6:        The more news media people are exposed to, the more they will
          fear stranger-perpetrated crime.
     Control variables that could influence people's perception of violent
crime--whether the perpetrator is a stranger or an acquaintance--will be applied
to all these variables, including: income, education, age, gender, personal
experience with violent crime, and salience of the crime issue.
Method
     In late September and early October of 1996 a 14-day telephone survey of
adults living in a major metropolitan area in the northeast United States was
conducted by forty-four graduate students in a large university in the region.
The survey instrument was designed by eight graduate students with the aid of
their supervisor.  The resulting omnibus survey contained 120 questions and was
designed to collect metropolitan-area adults' opinions on a wide variety of
topics.  Based on a pretest of 25 completed surveys drawn from a subset of the
random sample used to generate the final sampling frame, the instrument was
refined before entering the field in late September.
     In order to generate a random sample of the metropolitan area's adults, a
CD-ROM telephone directory (SelectPhone, 1997) was used to develop the sampling
frame for the study.  All telephone numbers within the metropolitan calling area
were selected from the CD-ROM database.  After eliminating business and
duplicate telephone numbers, a list of 197,000 telephone households was
identified as the study's sampling frame.
     These telephone numbers were then randomly ordered based on a computer
generated list of random numbers and divided into groups of fifty telephone
numbers to create 23 replicates of 50 numbers each for a total of 1,150
telephone numbers which were used in the study.  Further randomization within
each household was achieved through the Kish method, which was used to randomly
identify male and female members of the household to be included in the study.
     The present study's author was one of the eight researchers who helped
design the instrument.  As part of this study, the survey asked respondents to
answer questions about fear of crime and victimization, knowledge and awareness
of different crime types, and experience with violent crime.
     Exposure to television was operationally defined as the product of the
responses to the following questions:  1) In an average week, how many days do
you watch television? 2) On an average day, how much time do you watch
television?  Similarly, exposure to television news was defined as the product
of the responses to two questions:  1) In an average week, how many days do you
watch television? 2) On an average day, how much time do you spend watching
television news?
     Consistent with the theoretical discussion surrounding hypotheses five and
six, exposure to news media was operationally defined as the product of the
responses to the following questions:  1) In an average week, how many days do
you watch television? 2) On an average day, how much time do you spend watching
television news? plus the product of the responses to the following: 1) In an
average week, how many days do you listen to the radio? 2) On an average day,
how much time do you spend listening to radio news? plus the product of the
responses to the following questions: 1) In an average week, how many days do
you read a daily newspaper? 2) On an average day, how much time do you spend
reading a newspaper?
     Operationalizing awareness of stranger-perpetrated crime was done on two
levels.  On the first level, awareness of stranger-perpetrated crime was defined
as the response to the Likert-type question:  In most violent crimes, the
victims don't know their attackers, with potential responses ranging from
stronly agree (coded as 5) to strongly disagree (coded as 1).  Using this
definition, a score of 5 indicates strong awareness (perceived, not actual) of
stranger-perpetrated crime.
     The same concept was simultaneously measured through a more complex
mechanism where respondents were asked to estimate how many murder victims out
of 100 had known their killers[1].  The same question was asked for rape victims
and assault victims.  The responses to the three questions were summed to
produce an additive index with values ranging from 0 to 300 (alpha=.77). Higher
scores indicate less awareness of stranger-perpetrated crime.
     Fear of stranger-perpetrated crime was operationalized by combining scores
for two Likert-type questions into a scale.  The two questions are as follows:
1) When I leave my home, I worry that I could be mugged or attacked by
strangers, and 2) I'm afraid to walk alone at night.  Both questions range from
strongly agree (coded as 5) to strongly disagree (coded as 1).  The resulting
scale will have possible ranges from 2 to 10 where ten will indicate greatest
fear of stranger-perpetrated crime (alpha=.71).
     For control variables, salience of crime is operationalized through the
Likert-type question: Crime is the number one problem facing our country, with
resopnses ranging from strongly agree (coded as 5) to strongly disagree (coded
as 1).  Personal experience with violent crime is operationalized as the
response to the question, "Have you ever been the victim of a violent crime? "
For those who responded affirmatively, the follow-up question, "Did you know the
person who attacked you?" was asked.
     Based on the above operationalizations, all six hypotheses will be tested
using pearson's correlation coefficient.  To allow for inclusion of multiple
control variables, the same relationships will be further tested by hierarchical
regression analysis.
Results
     Table 1 contains the means and standard deviations of the TV and news media
exposure variables.  Table 2 shows the means and standard deviations for
indicators of fear and awareness of stranger-perpetrated crime as well as age.
Note that the mean exposure to television during an average week is roughly 15
hours, of which, more than a third, or 5+ hours, were spent with television
news.  Mean exposure to all news media is approximately 12 hours a week.  Table
3 shows percentages for categorical variables.  Of the respondents, 21% report
that they have been victims of violent crime themselves.  Of them, 62% knew the
perpetrator of that crime.
     Table 4 reports the results of the various correlations between media
exposure variables.    As can be seen, exposure to radio news does not correlate
significantly with the other media, whereas all other relationships are
significant.  Table 5 shows the bivariate tests of all six hypotheses.  Of the
six, hypotheses 1-4 are supported by the Pearson correlation coefficients.
Exposure to television in general leads to both fear and awareness of
stranger-perpetrated crime, and exposure to television news results in increased
fear and awareness.  However, the correlation between exposure to TV and
exposure to TV news is so large that it is questionable whether hypotheses three
and four receive support because of the nature of television news coverage or
because exposure to television news is itself dependent on exposure to
television programming in general.  If TV news content really does produce
different effects than TV in general, hypotheses five and six should receive
some support.  Based on the correlation coefficients, these hypotheses did not.
From this it cannot be stated that news media have a significant impact on fear
or awareness of stranger-perpetrated crime.
     In order to explore this situation further, hierarchical regressions were
conducted.  Tables 6 and 7 show the hierarchical regressions for the awareness
and fear indexes.  In each case, the impact of demographic variables, the
respondent's past experience with violent crime, and the respondent's attitude
towards the importance of crime as a national issue are controlled for in a
series of independently tested regression blocks.  The media exposure variables
included in the analysis are TV exposure and all news media exposure.  TV news
exposure was omitted because it correlated so highly with TV exposure that it is
unlikely to be measuring anything unique.  Thus, these regressions represent
attempts to retest hypotheses one, two, five, and six.
     Based on the relationships shown in Table 6, it is apparent that only one
variable has any impact on awareness of stranger-perpetrated crime:  whether or
not the respondent was the victim of a violent crime.  Curiously, the beta
coefficient is negative, indicating that those who were victims of violent crime
tended to estimate that more violent crime was committed by acquaintances rather
than by strangers.  This could be reflective of the fact that 62% of these
people were acquainted with their attackers.  Thus, personal experience with
violent crime is the only significant determinant of awareness of
stranger-perpetrated crime.
     Examining Table 7, it is obvious that something very different is happening
here.  Three variables are significantly related to fear of stranger-perpetrated
crime:  respondent's sex, respondent's attitude towards crime as a national
issue, and TV exposure.  From these figures, it appears that women are more
likely to fear stranger-perpetrated crime than men, people who believe crime is
the "number one" problem in our country are also more prone to fear, and people
who watch more TV are more likely to fear stranger-perpetrated crime as measured
here.  Note, however, that the R-square changes are very small for all the
variable blocks other than the demographic one, which has an R-square change of
.13 (p<.001).  For the TV exposure variable, while its beta coefficient was .14
(p<.05), the R-square change attributable to the media block was only .016
(p<.05).  From Tables 6 and 7, we again find no support for hypotheses five and
six which stated that exposure to news media in general would cause increased
fear or awareness of stranger-perpetrated crime.
Discussion
     The purpose of this study was to expand the current field of cultivation
research to include a distinction between different types of violent crime and
their potential cultivation impact.  It is theorized in the paper that there are
cognitive reasons for a difference in how "mean" people feel the world is when
exposed to crime-filled media content.  It is hypothesized that the crime type
people are most significantly impacted by is stranger-perpetrated violent crime,
rather than acquaintance-perpetrated crime, holding that portrayals of the
former are more likely to arouse fear reactions in the individual than the
latter.
     In order to maintain consistency with traditional cultivation research,
fear of stranger-perpetrated crime and awareness of stranger-perpetrated crime
are tested separately.  In the first two hypotheses, mean exposure to television
is tested against both fear and awareness of stanger-perpetrated crime.
Pearson's correlations performed on the relationships were significant at very
low levels (.19 and .13, respectively).
     The next hypotheses test the extended cultivation concept discussed
previously.  That is, that the content of television news is more likely than
the overall content of television to produce a cultivation effect.  Pearson's
correlations produced significant relationships between exposure to television
news and the fear and awareness indexes (.23 and .14, respectively).  This
indicates support for the extended cultivation hypotheses, although this result
is problematic given that television news exposure and television exposure
themselves are highly correlated at the .67 level (see Table 5).  An attempt was
made to determine whether exposure to nonnews television content was independent
of exposure to news television.  Here again, the correlation was high (see Table
4).  The one important detail to note is that the correlation between television
exposure and fear is lower than the correlation between television news exposure
and fear (.19 and .23).  This difference provides some indication that exposure
to television news may be measuring something qualitatively different from
exposure to television, but generalizing on the basis of these results is not
recommended.
     One further attempt to get at the difference between television and news
content was in hypotheses five and six.  Here, exposure to all news
media--including radio and newspaper--was tested against the fear and awareness
indexes.  The resulting correlations were not significant.  This could indicate
that, counter to the theoretical discussion presented above, the fact that news
information could be interpreted by users as more "real" than fictional
portrayals does not significantly impact the fear or awareness generated by it.
     To resolve the difficulties presented by these results, a look at the
hierarchical regressions might help.  As was discussed above, Table 6 shows that
when controlled for various variables that impact awareness of
stranger-perpetrated crime, the significance of the TV exposure variable drops
out.  The only remaining variable that is significantly related to awareness is
the person's past experience with violent crime, defined as whether or not the
individual was ever the victim of a violent crime.
     In Table 7, where fear of stranger-perpetrated crime is measured with
multiple controls, three variables come out as significant:  sex, reported
belief that crime is the "number one" problem in our country, and exposure to
television.  This is an interesting result in light of the fact that most
cultivation research findings drop out when multiple controls are applied.  It
is also interesting given that the cognitive aspects of cultivation (as in
awareness) are traditionally easier to measure than the affective aspects (as in
fear).  Here, the reverse is true.  This may be due to the fact that the mean
world index used in cultivation research was altered for the purposes of this
study to include specific references to strangers, thus trying to highlight the
impact of crime-type on the cultivation effect.
     The author believes this finding to be significant because it supports the
idea that the mean world Gerbner and his associates have sought to describe is
really a world of stranger-perpetrated violent crimes, not a world of crime in
general.  If these results hold true, they are a powerful argument for
reconsidering the theoretical basis of cultivation theory.
Considerations for Future Research
     The principle weakness of this study is the lack of a convincing method for
estimating fear of acquaintance-perpetrated crime.  Such a measure was not
included in this study because of the inherent difficulty in asking people in a
survey context if they are afraid of people they know hurting them.  To really
study the impact of perpetrator-type on the cultivation relationship, an
effective measure of that concept should be developed so that fear of the
different types of perpetrators can be compared and contrasted.
     Future research should seek to confirm these results in other geographical
areas as well.  When the sample was compared to the 1990 census data of the
metropolitan area's adults, it was found that the survey slightly
overrepresented females, upper income households, and highly educated
individuals.  Though the variations were slight and are very common for
telephone surveys, it is important to consider them when generalizing on the
basis of these results.
 
Fearing the "Mean World"
 References
 
Eysenck, H. J. & Nias, D. K. (1978). Sex, violence and the media. New
     York: Harper & Row.
Gallup Organization. (1995). Welfare, federal defecit emerge as public
     concerns: Crime and health care fade.
     http://www.gallup.com/newsletter/jan95/public_concerns.html, last searched
     8/21/96.
Gerbner, G. (1990). Epilogue: Advancing on the path of righteousness
     (maybe). In N. Signorielli, & M. Morgan (Eds.), Cultivation analysis: New
     directions in media effects research (pp. 249-262). Newbury Park, CA: Sage.
Gerbner, G., & Gross, L. (1976). Living with television: The violence
     profile. Journal of Communication, 26(2), 172-199.
Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1994). Growing up
     with television: The cultivation perspective. In Bryant, J. & Zillman, D.
     (Eds.), Media effects: Advances in theory and research (pp. 17-41).
     Hillsdale, NJ: Erlbaum.
Gilliam, F. D., Jr., Iyengar, S., Simon, A., & Wright, O. (1996). Crime in
     black and white: The violent, scary world of local news. Press/Politics, 1
     (3), 6-23.
Hawkins, R. P., & Pingree, S. (1981). Uniform messages and habitual
     viewing: Unnecessary assumptions in social reality effects. Human
     Communication Research, 7, 291-301.
Hawkins, R. P., & Pingree, S. (1990). Divergent psychological processes in
     constructing social reality from mass media content. In N. Signorielli, &
     Morgan (Eds.), Cultivation analysis: New directions in media effects
     research (pp. 35-50). Newbury Park, CA: Sage.
Hawkins, R. P., Pingree, S., & Adler, I. (1987). Searching for cognitive
     processes in the cultivation effect: Adult and adolescent samples in the
     United States and Australia. Human Communication Research, 13(4), 553-577.
Hirsch, P. (1980). The "scary world" of the non-viewer and other
     anomalies: A reanalysis of Gerbner et al.'s finding of cultivation
     analysis. Communication Research, 7, 403-456.
Hirsch, P. (1981). On not learning from one's own mistakes: A reanalysis
     of Gerber et al.'s finding of cultivation analysis, part II. Communication
     Research, 8, 3-37.
Johnstone, J. W. C., Hawkins, D. F., & Michener, A. (1995). Homicide
     reporting in Chicago dailies. Journalism Quarterly, 71(4), 860-872.
Kandel, E. R., Schwartz, J. H., Jessell, T. M. (1991). Principles of
     neural science, 3rd ed. Norwalk, CT: Appleton & Lange
Lowery, S. A., & DeFleur, M. L. (1988). Milestones in mass communication
     research: Media effects, 2nd ed. New York: Longman.
McLeod, J. M. (1995). The synthetic crisis: Media influences on
     perceptions of crime. Paper presented to the Association for Education in
     Journalism and Mass Communications, Washington, D.C., August.
Potter, W. J. (1994). Cultivation theory and research: A methodological
     critique. Journalism Monographs, 147(October).
Reeves, B., Chaffee, S. H., & Tims, A. (1982). Social cognition and mass
     communication research. In M. E. Roloff & C. R. Berger (Eds.), Social
     cognition and communication. Beverly Hills, CA: Sage Publications, Inc.
Shoemaker, P. J. (1987). Building a theory of news content: A synthesis of
     current approaches. Journalism Monographs, 103(June).
Shoemaker, P. J. (1996). Hardwired for news: Using biological and cultural
     evolution to explain the surveillance function. Journal of Communication,
     47(2), 32-47.
US Department of Justice. Bureau of Justice Statistics. (1994). Violent
     crime, NCJ-147486. WashingtonDC: BJS.
US Department of Justice. Bureau of Justice Statistice. (1996). Criminal
     victimization 1994, NCJ-158022. Washington DC: BJS.
US Department of Justice. Federal Bureau of Investigation. (1995). Uniform
     crime reporting program press release-11/19/95.
     http://www.fbi.gov/ucrpress.htm, last searched 9/13/96.
Wyer, R. S., & Srull, T. K. (1980). The processing of social stimulus
     information: A conceptual integration. In R. Hastie et al. (Eds.), Person
     memory: The cognitive basis of social perception. Hillsdale, NJ: Erlbaum.
 
Fearing the "Mean World"
 Table 1.       Means and standard deviations for TV and news media exposure
variables.
Variables
Mean
Std. Dev.
N
 
 
 
 
In an average week, how many days do you watch television?a
5.75
1.99
397
On an average day, how much time do you watch television?b
143.83
100.43
396
On an average day, how much time do you spend watching television news?b
51.89
41.51
396
Television exposure variablec
909.91
732.2
396
Television news exposure variabled
326.82
291.32
396
In an average week, how many days do you listen to the radio?a
5.36
2.36
393
On an average day, how much time do you spend listening to radio news?b
32.12
46.75
360
Radio news exposure variablee
202.34
310.4
388
In an average week, how many days do you read a daily newspaper?a
4.70
2.66
397
On an average day, how much time do you spend reading a newspaper?b
38.88
32.46
393
Newspaper exposure variablef
200.07
194.9
393
News media exposure variableg
726.96
517.6
383
______________________
a       Days out of seven
b       Coded in minutes
c       (minutes of television watched per day x number of days watched per week)
d       (minutes of television news watched per day x number of days watched per week)
e       (minutes of radio news listened to per day x number of days listened to per
  week)
f       (minutes of newspaper reading per day x number of days read per week)
g       Television news exposure variable + radio news exposure variable + newspaper
  exposure
 Table 2.       Means and standard deviations for fear and awareness of
stranger-perpetrated crime and age.
Variables
Mean
Std. Dev.
N
 
 
 
 
In most violent crimes, the victims don't know their attackersa
2.96
1.15
389
If 100 murders were committed in the United States this month, how many victims
do you think knew their killers?
54.25
25.18
367
If 100 rapes were committed in the United States this month, how many victims do
you think knew their rapist?
59.03
24.19
374
If 100 violent attacks were committed in the United States this month, how many
victims do you think would know their attacker?
51.54
24.43
376
Awareness of stranger-perpetrated crime scaleb
134.03
60.40
355
I worry I could be mugged or attacked by strangersa
2.44
1.11
395
I'm afraid to walk alone at nighta
2.85
1.28
397
Fear of stranger-perpetrated crimec
5.29
2.11
394
Crime is the number one problem facing our countrya
3.65
1.13
412
Age (in years)
46.10
16.12
392
______________________
 
a       strongly agree=5, agree=4, neutral=3, disagree=2, strongly disagree=1
b       Combined estimates, each out of 100, of how many victims of murder, rape, and
  assault knew their victims, ranging from 0 to 300, reversed so that 300
  indicates more awareness and 0 indicates less awareness, alpha=.77
c       Combination of "I worry I could be mugged or attacked by strangers" and "I'm
  afraid to walk alone at night," ranging from 2 to 10, where 10 indicates more
  fear and 2 indicates less fear, alpha=.71.
 
 Table 3. Percentages for experience with violent crime variables, gender,
education, and income.
Variables
%
 
 
Have you ever been the victim of a violent crime?
  Yes
  No
 
 
  20.8
  79.2
100.0%  (N=390)
(If yes) Did you know the person who attacked you?
  Yes
  No
 
 
  62.2
  37.8
100.0%  (N=82)
Sex
  Male
  Female
 
  41.3
  58.7
100.0%  (N=397)
Education
  Less than high school
  High school or equivalency
  Some college
  Associate's degree
  Bachelor's degree
  Graduate or professional degree
 
  10.9
  22.5
  21.3
  12.4
  21.5
  11.4
100.0%  (N=395)
Income
  Less than $20,000
  $20,000 - $29,999
  $30,000 - $39,999
  $40,000 - $49,999
  $50,000 - $59,999
  $60,000 - $69,999
  $70,000 - $79,999
  $80,000 or more
 
  16.7
  14.4
  14.1
  15.8
  11.1
    7.3
    9.4
  11.1
100.0%  (N=341)
 
 Table 4.       Pearson correlation coefficients for all media exposure variables
(exposure to TV, TV news, nonnews TV, radio, newspaper, and all news media).
Variables
2
3
4
5
6
1.  Exposure to TV*
.63c
(395)
1.00c
(395)
-.06
(387)
.20c
(392)
.40c
(382)
2.  Exposure to TV news**
-
.62c
(394)
.02
(387)
.28c
(392)
.67c
(383)
3.  Exposure to nonnews TV***
 
-
-.06
(386)
.20c
(391)
.39c
(381)
4.  Exposure to radio news****
 
 
-
.13a
(384)
.66c
(383)
5.  Exposure to newspaper*****
 
 
 
-
.61c
(383)
6.  Exposure to all news
     media******
 
 
 
 
-
 
 
 
 
 
 
a p < .05
b p < .01
c p < .001
 
 
 
 
 
 
*       (minutes of television watched per day x number of days watched per week)
**      (minutes of television news watched per day x number of days watched per
     week)
***     ([minutes of television watched per day - minutes of television news
     watched per day] x number of days watched per week)
****    (minutes of radio news listened to per day x number of days listened to
     per week)
*****   (minutes of newspaper reading per day x number of days read per week)
******  Television news exposure variable + radio news exposure variable +
     newspaper exposure
 
 
 Table 5.       Pearson correlation coefficients for independent and dependent
variables (fear and awareness of stranger-perpetrated crime by exposure to media
variables).
Variables
2
3
4
5
6
1.  Exposure to TV*
.63c
(395)
.40c
(382)
.19c
(392)
.13a
(354)
.08
(387)
2.  Exposure to TV news**
-
.67c
(383)
.23c
(392)
.14a
(355)
.07
(387)
3.  Exposure to all news media***
 
-
.10
(379)
.04
(346)
.04
(375)
4.  Fear of crime index_
 
 
-
.16b
(353)
.15b
(385)
5.  Awareness of
     stranger-perpetrated crime
     index__
 
 
 
-
.54c
(351)
6.  In most violent crime, victim's
     don't know their attackers___
 
 
 
 
-
 
 
 
 
 
 
a p < .05
b p < .01
c p < .001
 
 
 
 
 
 
*       (minutes of television watched per day x number of days watched per week)
**      (minutes of television news watched per day x number of days watched per
     week)
***     Television news exposure variable + radio news exposure variable + newspaper
     exposure
_       Combination of "I worry I could be mugged or attacked by strangers" and "I'm
     afraid to walk alone at night," ranging from 2 to 10, where 10 indicates
more
     fear and 2 indicates less fear, alpha=.71
__      Combined estimates, each out of 100, of how many victims of murder, rape,
     and assault knew their victims, ranging from 0 to 300, reversed so that 300
     indicates moreawareness and 0 indicates less awareness, alpha=.77
___     strongly agree=5, agree=4, neutral=3, disagree=2, strongly disagree=1
 
 
 
 Table 6. Hierarchical regression analysis of demographic variables, experience
with violent crime,  salience of crime, and media exposure on awareness of
stranger-perpetrated crime index*.
 
Independent variable blocks
 
Std. beta
R-square change
Total R-square
Adjusted R-square
 
 
 
 
 
1. Demographic variables
    Age
    Gender (female = 1)
    Income
 
 .084
-.006
-.008
 
 
 
.012
 
 
 
.012
 
 
 
.002
2. Have you ever been the victim of
    violent crime (yes = 1)
-.155b
.025b
.037
.024
3. Crime is the number one
    problem in our country**
 .084
.009
.046
.030
4. Media exposure variables
    TV exposure***
    All news media exposure****
 
 .125
-.042
 
 
.012
 
 
.058
 
 
.035
 
 
 
 
 
_________________
a p<.05
b p<.01
c p<.001
 
*       Combined estimates, each out of 100, of how many victims of murder, rape, and
     assault knew their victims, ranging from 0 to 300, reversed so that 300
     indicates more awareness and 0 indicates less awareness, alpha=.77
**      strongly agree=5, agree=4, neutral=3, disagree=2, strongly disagree=1
***             (minutes of television watched per day x number of days watched per week)
****    Television news exposure variable + radio news exposure variable +
  newspaper exposure
 Table 7. Hierarchical regression analysis of demographic variables, experience
with violent crime,  salience of crime, and media exposure on fear of
stranger-perpetrated crime index*.
 
Independent variable blocks
 
Std. beta
R-square change
Total R-square
Adjusted R-square
 
 
 
 
 
1. Demographic variables
    Age
    Gender (female = 1)
    Income
 
 .014
 .301c
-.029
 
 
 
.130c
 
 
 
.130
 
 
 
.121
2. Have you ever been the victim of
    violent crime (yes = 1)
 .097
.008
.138
.127
3. Crime is the number one
    problem in our country**
 .206c
.045c
.183
.170
4. Media exposure variables
    TV exposure***
    All news media exposure****
 
 .142a
-.016
 
 
.016a
 
 
.199
 
 
.181
 
 
 
 
 
_________________
a p<.05
b p<.01
c p<.001
 
*       Combination of "I worry I could be mugged or attacked by strangers" and "I'm
     afraid to walk alone at night," ranging from 2 to 10, where 10 indicates
more
     fear and 2 indicates less fear, alpha=.71
**      strongly agree=5, agree=4, neutral=3, disagree=2, strongly disagree=1
***             (minutes of television watched per day x number of days watched per week)
****    Television news exposure variable + radio news exposure variable +
  newspaper exposure
[1] Question wording was:  If 100 murders were committed in the United States
this month, how many victims do you think knew their killers?  The same question
was asked again twice, substituting rapes and violent attacks for murders, and
rapist and attacker for killers.