The media, attorneys, and other actors in the legal system define the CSI Effect as a phenomenon in which jurors who view television crime shows, particularly CSI, require forensic evidence in every trial. This requirement is reported to affect jurors’ verdicts (Tyler, 2006). If forensic evidence is present, the CSI-watching jurors may give it too much weight; if forensic evidence is absent the CSI-watching jurors will be skeptical of other common trial evidence because of the belief that forensic evidence should be available in all trials. As Cole & Dioso-Villa (2007: 441) argue, the theme derived from CSI (and subsequently adopted by jurors) is “that people lie, but science always tells the truth.”

Research on cultivation theory has examined how fictional television representations affect people’s perceptions of reality, which in turn affects their behaviors (Shanahan & Morgan, 1999). The CSI Effect, from a cultivation theory perspective, proposes that crime television programs have a real world effect on jurors’ decisions (including verdict). That is, viewing fictional television representations of crime and crime investigation can affect people’s perceptions of crime investigation in the real world, which then affects their behavior during trial. The media adopted this view of crime shows and started presenting the CSI Effect as a phenomenon beginning in 2003 (Houck, 2006).

Research and legal comment in this area has also indicated that legal professionals believe there is a CSI Effect and have changed their trial and voir dire strategies in response to this perceived threat (Watkins, 2004; Maricopa County, 2005; Tyler, 2006). In one study (2005), the majority of prosecuting attorneys believed that shows like CSI directly affected jurors’ judgments; similar results were also reported with defense and prosecuting attorneys in a subsequent survey (Watkins, 2004). Additionally, these attorneys reported changing the types of questions they ask jurors during voir dire (Watkins, 2004) and adding witnesses to their presentation of evidence (Houck, 2006) all based on their belief that the CSI Effect exists. Depending on attorneys’ perceptions and their use of voir dire questions and challenges, certain jurors may be excluded based on whether they watch CSI type shows. Research on whether the CSI Effect actually changes jurors decisions has shown somewhat equivocal results, with some studies showing no relationship between jurors’ CSI watching habits and verdict (Podlas, 2006), and others showing mixed results (Hayes-Smith & Levett, 2011; Reardon, O’Neil & Lawson, 2007; Shelton, Kin & Barak, 2006).

Even so, attorneys may change the way they interact with jurors based on their perception of the CSI Effect, which in turn may change the way jurors respond to the attorney or evidence. For example, some attorneys may warn jurors against using knowledge gained from shows like CSI in their decision making. Research has examined the effect of this type of anti-CSI warning on jurors’ decisions (York, O’Neil, Evans, 2006). According to York et al. (2006), the anti-CSI warning yielded a backfire effect, decreasing jurors’ confidences in the prosecutor’s case. However, it is also possible that such warnings or changes in voir dire strategy are unnecessary because jurors are already aware of the CSI Effect. If the general public is aware of the CSI Effect, this awareness may lessen the impact that crime shows have on public views of forensic evidence, and in turn, jurors’ trial decisions. That is, if the general public is aware that watching shows like CSI may affect their trial decisions in a negative way, it is possible that jurors may correct for any possible CSI Effect. To our knowledge, research to date has not yet examined whether the general population is aware of the CSI Effect. For the present study, we extended research examining legal professionals’ beliefs about the CSI Effect and explored community members’ beliefs in a CSI Effect. We explored peoples’ knowledge of and beliefs in the CSI Effect using an online survey, soliciting members of the community, and using both qualitative and quantitative items.

The Social Construction of the CSI Effect

According to Houck (2006), the CSI Effect started to appear in the press in 2003 with anecdotal evidence from attorneys (mostly prosecutors) and judges. For example, the vice president of the National District Attorneys Association reported to CBS News that jurors were expecting DNA tests for almost all cases, but in most cases, such evidence is not readily available (Houck, 2006). In 2007, Cole & Dioso-Villa identified 416 different news articles that used the term “CSI Effect” in either the title or the text of the article. The popularity of the CSI Effect as a news story appears to be due to localization, meaning that a news story with a particular theme can simply be recast in any area by using the local players to demonstrate the theme (Cole & Dioso-Villa, 2007). That is, every town has police investigators, prosecutors and defense attorneys, and those players can be cast into the story discussing the CSI Effect.

A recent content analysis of newspaper headlines assessed how the CSI Effect was described in the media (Patry, Stinson & Smith, 2008). Results indicated that the news media characterized the CSI Effect as a phenomenon that negatively influences juror decision-making. Specifically, they found that the media reported juries were entering the courtroom with unrealistic expectations about forensic evidence because of what they had seen on various crime dramas. Further, those unrealistic expectations were ultimately affecting jurors’ verdicts. Another examination of news stories covering the CSI Effect showed that the media’s depiction of the CSI Effect had changed from ‘a potential nuisance’ in 2002 to a ‘huge problem’ in 2006 (Cole & Dioso-Villa, 2007). However, the media made these conclusions in their news stories without the use of empirical evidence (Cole & Dioso-Villa, 2007).

Given that the media has highly publicized the existence of a CSI Effect, the CSI Effect fits well into what both Surette (2007) and Barak (1994) have discussed regarding the social construction of crime and justice by the media. That is, the media assists in the creation of the public’s knowledge regarding crime and justice in part because the average individual is unlikely to have had much contact with the criminal justice system (Surette, 2007). The media’s construction of the CSI Effect is supported by examining the headlines and themes present in the news media (Cole & Dioso-Villa, 2007; Patry et al., 2008) and is consistent with the underlying assumption of cultivation theory that proposes that the elite of society, in this case, the media, constructs social reality for the remaining general population (Shanahan & Morgan, 1999). For example, one USA Today headline read ‘CSI Effect has juries wanting more evidence’ (Willing, 2004; also see Lovgren, 2004). In this article, Willing (2004) reported on the highly publicized Robert Durst case in which the jury consultant for the defense said he purposely placed people on the jury who watched shows, such as CSI, because he believed they would find the defendant not guilty due to lack of forensic evidence. Durst was eventually found not guilty. According to the media, this verdict was at least partially attributed to the way the jury was stacked with CSI viewers.

Legal comment concerning the CSI Effect has primarily indicated that legal professionals believe it exists and change their trial or investigation strategies accordingly (Watkins, 2004; Maricopa County, 2005; Tyler, 2006; Houck, 2006; Stinson, Patry & Smith, 2007; Stevens, 2008). In one of the first assessments of the CSI Effect, fifty-three prosecuting and defense attorneys were questioned about their perceptions of how viewing crime dramas affect the jury, their pretrial preparations, and whether their interactions with the jury were affected by their belief in the CSI Effect (Watkins, 2004). All of the prosecutors and most defense attorneys believed that crime dramas created unrealistic expectations among potential jurors concerning forensic evidence. However, only nine attorneys said that they had a direct experience with a juror who had this type of unrealistic expectation (Watkins, 2004). Conversely, almost half (49%) of the participants responded that they believed between one and five acquittals had occurred because of a lack of forensic evidence (Watkins, 2004). Even so, attorneys responded that changes had occurred in their preparation for court (e.g., changing voir dire strategies and asking jurors whether they watched crime dramas, asking for more forensic evidence, emphasizing police officers’ non-requests for forensic evidence, and emphasizing forensic testing of irrelevant information).

Other studies have shown corresponding results (Maricopa County, 2005; Robbers, 2008). In one open-ended survey of trial attorneys, three themes emerged: 1) attorneys recalled specific instances of cases where they felt forensic evidence (or the lack of it) influenced case outcomes and attributed this to jurors viewing crime dramas, 2) attorneys expressed changing their job execution because of crime dramas, and 3) attorneys believed that jurors’ constructions of the criminal justice system were shaped by viewing crime dramas (Robbers, 2008). Another study surveyed all 102 prosecutors in one office and found prosecutors believed there was a CSI Effect (Maricopa County, 2005). Specifically, 45 % of prosecutors believed jurors rely too much on scientific evidence, 72 % contended that jurors perceive themselves to have forensic expertise from viewing those shows, and 80 % felt that defense attorneys play on jurors’ perceptions created by watching crime dramas to help their case (Maricopa County, 2005).

Another series of similar studies examining attorneys’ beliefs about the CSI Effect found mixed results (Patry et al., 2008; Stinson et al., 2007). First, lawyers in this study did not perceive the CSI Effect as a problem with jurors, but they reported that their clients had distorted views of the legal process because of crime dramas. Expanding beyond the opinions of lawyers, Stinson et al. (2007) and Patry et al. (2008) surveyed police investigators, medical examiners, fire/arson examiners and other legal professionals who investigate various causes of death. In this study, 94 % of respondents indicated that they believed that television crime dramas have altered the public’s perception of their profession.

Increased Demand for Forensic Evidence

Collectively, these studies demonstrate that attorneys generally believe the CSI Effect exists, and several different legal actors perceive that crime dramas alter the public’s perception of forensic evidence. Some research has shown that prosecutor’s concerns that jurors weigh forensic evidence more than other types of evidence are not unwarranted. In a survey of community members measuring opinions regarding different types of forensic evidence, researchers reported that community members preferred forensic evidence to other common types of evidence. Community members’ two major preferences were for DNA and fingerprint evidence. However, this study did not examine whether participants’ preferences for scientific evidence was linked to watching crime dramas (Patry et al., 2008).

A second survey in the same series of studies showed that participants who reported watching more crime dramas were more likely to view some forms of forensic evidence (DNA, arson, compositional and handwriting analysis) as being more reliable compared to those participants who watched less crime dramas. However, this relationship did not hold for other forms of evidence (i.e., fingerprints, toxicology, confessions, and eyewitness testimony; Patry et al., 2008; Smith et al., 2007).

As a result of the perception that forensic evidence will be an evidentiary requirement for a defendant’s guilt, one study reported an increase in the amount of evidence collected from crime scenes; for example, in scenes where a police officer would have previously collected 50 pieces of evidence, police officers may attempt to collect up to 400 (Houck, 2006). Forensic labs then become backlogged because of the increase in the amount of evidence submitted for analysis. This increased demand (based on the perception that jurors require forensic evidence to convict) creates more strain on the entire process and those involved. A report by the Bureau of Justice Statistics (BJS) stated that as of January 2001, 81 % of DNA crime laboratories reported increased numbers of forensic cases and DNA analyses backlogs (reported in Steadman, 2002). Much of the increase in the trend may be attributed to more available science; however, it may also be due to a greater demand by criminal justice professionals for this type of evidence.

Despite the increase in demand for forensic evidence, the BJS (2008) reported that in reality, crime rates are experiencing a downward trend. Violent crime and property crime rates in 2005 were at their lowest recorded levels since 1973. Homicide rates were also decreasing, and homicide investigations are the cases in which forensic-type evidence is most often utilized (BJS, 2008). With crime rates down and the demand for forensics going up, most likely, criminal justice actors are collecting more evidence for examination due to increased pressure to do so. Stevens (2008) argues that the so-called CSI Effect may have a trickle-down effect. First, jurors may demand more evidence from attorneys. Prosecutors being pressured to convict along with their perceptions of jurors will want and demand more evidence from investigators, and investigators in turn will collect more evidence. This affects labs, which are then overwhelmed with the evidence collected.

Overview of the Current Study

These studies demonstrate that legal actors, attorneys, and the media believe that the CSI Effect exists and that individuals tend to prefer the types of evidence used on the shows. Further, criminal justice professionals are changing the way they collect evidence because of these perceptions. However, to date, no one has asked jurors whether they perceive a CSI Effect (see Watkins, 2004; Maricopa County, 2005; Patry et al., 2008). Many people do not have direct contact with the legal system, which may mean they have not been exposed to the idea of the CSI Effect in voir dire or trial strategy. However, the media’s publicity about the CSI Effect could mean that community members have been exposed to the idea by reading newspapers or watching television. Further, it is possible that those who watch CSI may be more likely to have heard of the CSI Effect. That is, those individuals may be more apt to pay attention to publicity about the CSI Effect. If individuals were exposed to the CSI Effect idea through this medium, it may be likely that they would perceive that watching shows like CSI has the potential to bias juror decision making. In turn, this awareness could mean that if they were called to jury duty, they may take this proposed effect into consideration in their decision making. It is possible they would actively try to suppress this bias, which may be dependent on their motivation and ability to do so (Fleming, Wegener, & Petty, 1999; Petty & Wegener, 1993). In this study, we started exploring this possibility by examining community members’ beliefs about the CSI Effect.

Overall, this study was exploratory so there were no formal hypotheses. The overarching goals were to investigate the following questions: First, have participants heard of the CSI Effect? And, are participants who are heavy crime show viewers more likely to have heard of the CSI Effect compared to light crime show viewers? Second, what are community members’ perceptions of the CSI Effect? To answer our research questions, we conducted an online survey of community members.

Method

Participants

Participants were 191 community members solicited through a snowball sampling technique. There were two recruiting procedures (explained below). The overall sample was 65 % female, 35 % male with 83 % non-students. Participants indentified their racial/ethnic background as White/Non-Hispanic (81 %), Black, Non-Hispanic (6.3 %), Asian (2.1 %), Hispanic (2.6 %), American Indian (.5 %) and Other (2.6 %). 33 % reported being single, 52 % reported being married, 13 % reported being divorced, and 2 % reported being widowed. Total household income of 46 % of the participants was less than 50,000 dollars a year. The CSI Effect is argued to influence jurors’ decisions, so we asked participants if they had ever been called for jury duty. 59 % responded they had and 39 % responded they had not. Of those who had been called, only 21 % of them had decided a verdict in trial.

Procedure

Several sampling procedures were used. First, we posted a link and request for participation on Craigslist.com® (16 % of participants). Second, we canvassed public libraries in the area with flyers to recruit participants (73 % of participants). Public libraries were selected because computers with internet access were available for public use in each location. These two recruiting techniques were used in concert due to there not being a readily available sampling frame of the general population, and so non-probability sampling was an appropriate technique to solicit participants (Kaye & Johnson, 1999). The survey was posted on Surveymonkey.com®. After signing in, participants were asked to read the informed consent prior to proceeding with the questionnaire. 75 % of 259 participants who started the survey completed the survey.

Measures

Television and Crime Viewing Behavior

Consistent with prior research on cultivation theory (Gerbner & Gross, 1976; Shanahan & Morgan, 1999), participants were asked the following open-ended questions (each of these were used separately as general television viewing measures): (a) On an average day how many hours do you watch television; (b) Of these hours a day how many are spent watching crime drama shows; (c) how many hours, in general, do you watch television every week; (d) of these hours how many are spent watching crime drama shows; and (e) please indicate approximately how many hours a week that you view any of these (from the list above) crime dramas (Gerbner et al. 1978).

Participants were asked whether they watch several popular crime shows separated by genre category. These shows were categorized according to Shelton et al. (2006) and a full listing of the shows is available in the Appendix (forensic dramas, forensic documentaries, general crime documentaries, crime/courtroom drama, and general crime news shows). For each genre, participants indicated whether they had seen each show classified in that genre (coded as a 1 if they had seen the show and 0 if they had not). They also answered two general questions about the amount of time they spent viewing each genre in a day and in a week. Responses to these questions were summed across genres to create an overall viewing scale (one for hours in a day and one for hours in a week).

Knowledge of the CSI Effect

The participants were asked if they had heard of the CSI Effect. Participants responded to: (a) Have you heard of the phenomenon termed the CSI Effect on juror decisions? (yes or no); (b) If yes, do you know what it means? (yes or no) (c) If yes, briefly describe what you think the CSI Effect is. Participants’ responses to this last open ended item (and all other open ended items) were examined for themes and then the themes were coded using two independent coders with a high overall agreement rate (κ = .72). Next, participants responded to: (a) do you think that the CSI Effect as you defined it occurs in real life? (yes or no).

Perceptions of the CSI Effect

After answering questions about their knowledge about the CSI Effect, participants were given a definition of the CSI Effect and asked to agree or disagree with a series of statements designed to measure whether participants believed the CSI Effect existed. Each statement was measured by a 7-point Likert-type scale ranging from ‘strongly disagree’ to ‘strongly agree’, and items were re-coded so that a higher rating indicated stronger belief in the CSI Effect. Exploratory factor analyses on these items revealed three underlying constructs, and results of the separate factor analyses revealed that at least 40 % variance was explained by the constructs and all factor loadings were above a .50 threshold (Allison, 1999). These scales and items are reported in Table 1.

Table 1 Participants’ responses to perceptions of the CSI Effect scales
Full size table

We then ran bivariate analyses and multivariate regression analyses that examined television viewing behaviors and participants’ demographics in relation to whether they had heard of a CSI Effect and their perceptions of the CSI Effect (results reported below). Multivariate statistics (heteroskedasticity, multivariate normality, and multicollinearity) were tested for each model. There were no indicators of multicollinearity (all tolerance levels well above the .4 threshold) and the scatterplots for homoscedasticity were reasonably distributed.

Results

Descriptives

In responding to questions about how often they viewed television, participants reported watching an average of 17.86 hours (range 0–61) of television weekly with 4.60 of those hours spent watching crime shows (range 0–32). Participants reported watching an average of 2.99 hours daily (range = 0–15), and of these hours an average of .86 hours (range = 0–6) was spent viewing crime shows. When asked how many hours a week that they view any crime show, participants’ responses averaged 5.13 hours (range = 0–29). The distributions for all of these measures were highly skewed (all skew’s > 7.53) with the values closely surrounding the mean (all kurtosis > 6.44).

Participants also responded how many hours weekly and daily they viewed each of the specific genres of following crime shows: Forensic Dramas weekly M = 2.53 (range = 0–45) and daily M = .60 (range = 0–9); Forensic Documentaries weekly M = .76 (range = 0–30) and daily M = .23 (range = 0–5); Crime Documentaries weekly M = .92 (range = 0–10) and daily M = .27 (range = 0–9); Crime Dramas weekly M = 1.80 (range = 0–45) and daily M = .54 (range 0–12); and Crime News weekly M = 1.28 (range 0–15) and daily M = .37 (range 0–6). The distributions were also all highly skewed (all skew’s > 11.70) with many zero responses and with values close to the mean (all kurtosis > 11.89). To get a better picture of whether viewing each of these genres of shows was related to beliefs about the CSI Effect (and to address skewness), participants’ responses in each category were recoded into zeros if participants indicated not viewing any of the shows and ones if they indicated viewing at least one hour a week to create a dichotomous variable. Most participants reported viewing forensic documentaries and crime documentaries (70 %) followed by forensic dramas (59 %), crime news (58 %), and crime dramas (50 %). Even though the continuous items were skewed logging them further complicated the interpretation and did not help with skewness; therefore they were not used in final multivariate models. However, we did use the dichotomous variables for the television and crime viewing measures for the bi-variate analyses.

CSI Effect Qualitative Analysis: What Does the CSI Effect Mean to Community Members?

The overarching goal of this study was to examine whether community members were familiar with the CSI Effect. The majority of participants (70 %) indicated that they had not heard of the CSI Effect, 29 % indicated that they had. Of the 29 % of participants who indicated that they had heard of the CSI Effect, 68 % indicated that they knew what the CSI Effect meant; 88 % of the 58 participants who indicated that they knew what the CSI Effect was answered the open ended question regarding their description of the CSI Effect.

These responses were coded for four major themes. The main theme in participants’ responses was that the CSI Effect meant unrealistic expectations of evidence (67 %). For example: “people expect a lot of crime scene evidence. They think there is always some defining evidence left at the scene that will lead to the guilty party”. The second theme was that participants explained that people tend to believe that forensic science in the real world is conducted similarly to how it is conducted on crime shows (26 %). For example, one respondent explained: “when people think real life crime and justice is the same as what they see on TV”. The third theme was that participants believed that CSI-type shows are educating criminals (18 %). For example: “when people who commit a crime (or crimes) try to cover-up evidence based on what they have seen on TV”. The final theme was related to people who watch CSI believe they may have the ability to solve crimes because of knowledge they learned on these shows (8 %); for example: “[CSI] creates people who believe that they can solve crime and know more”. There were a few answers (6 %) that did not neatly fit into any theme, for example: “it is when the crime is copied from a movie script”. When asked whether they think the CSI Effect occurs in real life, 85 % of the participants answered yes and 15 % responded no.

Within the most prevalent, unrealistic expectations of evidence theme, there were five sub-themes in which participants indicated more specific definitions of the CSI Effect. All percentages are within response percentages. First, participants believed that the CSI Effect affects jurors’ expectations of forensic evidence (37 %). For example: “juries tend to expect too much from forensic testimony”. Second, people expressed that shows like CSI influence both criminal justice actors and juries (16 %). For example: “the police officers, forensic teams and the legal system are expected to be able to gather, process and use high tech means to supply evidence in court”. Third, participants expressed that the CSI Effect means that to solve a crime, criminal justice actors need forensics (10 %). For example: “people expect a lot of crime scene evidence. They think there is always some defining evidence left at the scene that will lead to the guilty party”. Fourth, people expressed that the expectations from these shows make it more difficult for state attorneys/prosecutors to win a case (8 %). For example: “as a juror you expect the presentation of convincing evidence (fingerprints, etc.) to be part of the case for conviction”.

CSI Effect Quantitative Analysis: What Does the CSI Effect Mean to Community Members?

After we provided a definition of the CSI Effect, participants responded to the items measuring their perceptions of said effect (items and responses available in Table 1). To examine participants’ responses, we examined participants mean responses to each scale and item within the scale. In addition, we examined the percentage of participants who agreed with the item (i.e., answered strongly agree, agree, and slightly agree), were neutral on the item, and disagreed with the item (i.e., answered strongly disagree, disagree, and slightly disagree).

For the trial outcomes scale, the items measured the CSI Effect as an influence on the outcomes of a trial. The means of responses to items were around the midpoint of the 7-point scale; the lowest mean was 3.24 (SD = 1.43) with most respondents indicating they believed watching crime shows influences trial outcomes on most measures. For example, when asked if, “jurors who watch crime drama shows will make a decision about a case differently than those who do not watch crime drama shows” 70.6 % agreed that this statement was true.

On average, participants’ responses were higher on the education scale with the highest item mean being 5.35 (SD = 1.28). Thus, participants reportedly believed that crime shows generally educate the public. For example, 57.5 % of participants agreed with the item “I think that shows such as CSI give people more faith in the criminal justice system”.

The third and final scale examined participants’ views of people who watch crime drama shows expectations of evidence. Again, the mean of all items seems to be slightly above the midpoint with the lowest at 3.77 (SD = 1.50) and the highest 5.53 (SD = 1.20). A high percentage 80.9 % of participants agreed with the item that read, “jurors who watch crime drama shows will expect more forensic evidence.” Alternatively, participants were split on the item that read, “jurors who watch crime drama shows, will not convict without forensic evidence” with 23.9 % agreeing, 37.3 % disagreeing and 36.3 % remaining neutral.

Television Viewing Behaviors and Participant Knowledge of the CSI Effect

Having heard of the CSI Effect was positively correlated with how many hours a week a participant viewed crime dramas (r (193) = .17, p < .05) and how many hours daily they viewed crime shows (r (198) = .23, p < .05). Participants who reported hearing about the CSI Effect (yes or no) had different crime drama viewing habits than those who reported they had not heard of the CSI Effect (χ² (2, 193) = 7.27, p = .03). Thus, participants who reported having heard of the CSI effect were more likely to be heavy viewers of crime shows than those who reported not having heard of the CSI effect.

Participants’ Daily/Weekly Television Viewing and Perceptions of the CSI Effect

To examine whether participants’ general, overall television viewing habits were related to their perceptions of the CSI Effect, we tested the bi-variate correlations between measures of television viewing behavior (weekly and daily) and the measures of participants’ perceptions of the CSI Effect (education scale, trial outcome scale, and expectation of evidence scale). There was a positive significant relationship between participants’ weekly television viewing and the education scale (r (182) = .16, p < .05). Specifically, the more hours of television participants viewed weekly, the more likely they were to state that crime shows educated people on investigation and evidence issues. The other relationships between daily and weekly viewing measures and perceptions of the CSI Effect scales were not significant (all rs < .13, all ps > .07).

Participants’ Daily/Weekly Crime Show Viewing and Perceptions of the CSI Effect

To test whether viewing crime shows was related to participants’ perceptions of the CSI Effect, we calculated another series of bivariate correlations between measures of crime show viewing (daily and weekly) and perceptions of the CSI Effect (education scale, trial outcome scale, and expectation of evidence scale). Similar to the relationship with overall television viewing, there was a significant positive relationship between participants’ crime show viewing weekly behavior and the CSI Effect education scale (r (181) = .18, p < .01). Specifically, participants who were heavier viewers of crime shows weekly also stated that the shows were educating people on investigation and evidence issues compared to those who were lighter viewers of crime shows weekly. The other relationships between viewing measures and perceptions of the CSI Effect scales were not significant (all r’s < .13, all p’s > .07).

Participants’ Crime Genre Viewing and Perceptions of the CSI Effect

Again, we ran bivariate correlations to examine the relationship between genre viewing measure scales (weekly and daily) and perceptions of the CSI Effect scales. Results indicate that both forensic drama daily viewing and general crime/courtroom drama daily viewing were negatively correlated with the expectation of evidence scale (r (172) = −.23, p < .01; r (184) = −.23, p < .01, respectively). Watching more forensic dramas or general crime/courtroom dramas daily was associated with lower scores on the expectation of evidence scale. That is, those who watch shows like CSI daily were less likely than those who did not watch those shows daily to believe that watching said shows would influence people’s expectations of evidence. Results also indicated that general crime/courtroom drama weekly viewing was positively correlated to the education scale (r (177) = .15, p < .05); watching more crime/courtroom dramas weekly was associated with believing that they were more educational. All other relationships were not significantly correlated (all rs < .10, all ps > .05).

Exploring Socio-demographics and Perceptions of the CSI Effect

Last, whether socio-demographic variables (female, white, marriage, income level and whether they had served as a juror) were related to the scales measuring perceptions of the CSI Effect was tested using a series of ANOVAs. There were significant relationships between being white and the education scale F (1,169) = 5.20, p < .05 and the evidence scale F (1,171) = 4.12; p < .05. Participants who were white were more likely than those of other races to agree that crime shows educate people (M = 4.59 vs. M = 4.18, respectively). In addition, participants who were white were more likely than those of other races to agree that people who watch crime shows expect more forensic evidence (M = 4.61 versus M = 4.27, respectively). No other relationship between socio-demographic variables and the CSI Effects scales were significant (all Fs < 2.56, all ps > .08).

Bivariate correlations between the perceptions of the CSI Effect scales and participants’ and education levels & political beliefs indicated a significant relationship between level of education and the trial outcome scale, r (175) = −.153, p < .05. So, being more educated was associated with a greater belief that watching shows like CSI could affect jurors’ trial decisions. No other bivariate relationships between perceptions of CSI Effect scales and socio-demographic variables were significant (all rs < .11, all ps > .13).

Exploring Television Viewing & Socio-Demographic variables and Perceptions of the CSI Effect

Multivariate binary logistic regression regarding whether participants had knowledge of the CSI effect revealed that television viewing significantly influenced whether participants reported having knowledge of the CSI Effect in all models (all p’s < .05). In the model including general weekly television viewing both being non-white and of lower income also influenced knowledge of the CSI Effect. As participants’ weekly hours of viewing television increased, the odds of them selecting they had heard of the CSI Effect also increased, B = .02, S.E. = .01, Wald’s χ 2 (1 167) = 4.89, p < .05, exp(B) = 1.03. In addition, as participants’ income decreased, the odds of them selecting they had heard of the CSI effect also decreased, B = −.69, S.E. = .36, Wald’s χ 2 (1, 167) = 3.71, p < .05, exp(B) = .50. Last, those who were non-white were more likely to report that they had heard of the CSI Effect, B = −1.13, S.E. = .58, Wald’s χ 2 (1, 167) = 3.84, p < .05, exp(B) = .32.

For the following analyses, we included socio-demographic variables (race, gender, marital status, education, and income) and with television viewing behaviors as predictors on the three CSI Effect scales (education, trial outcome and evidence) as dependent measures in a series of multivariate analyses. The multivariate analyses for these relationships revealed a similar picture to the bivariate analyses regarding the influence of television viewing on participants’ views of the CSI effect as educational. That is, the more hours of television participants viewed weekly, the more likely they were to state that crime shows educated people on investigation (β = .24, t (154) = 3.09, p < .05). In addition, being white also significantly predicted whether participants viewed crime shows as educational (β = .19, t (154) = 2.42, p < .05). Furthermore, educational level significantly predicted whether participants rated crime shows as educational, β = .16, t (154) = 1.93, p < .05. In a comparable model, daily television viewing did not significantly predict whether participants reported higher on the education scale (p = .96), but being white significantly predicted higher ratings on the education scale (β = .43, t (155) = 2.18, p < .05).

The only other significant model was for the CSI Effect scale measuring perceptions of evidence. Participants who were higher in daily television viewing were less likely to agree that shows such as CSI have an effect on participants views of evidence (β = −.05, t (156) = −2.07, p < .05). No other predictors were significant. None of the television viewing or socio-demographic variables significantly predicted participants’ perceptions of trial outcomes (all p’s > .13).

Exploring Crime Show Viewing & Socio-Demographic Variables and Perceptions of the CSI Effect

Similarly to what was described above, binary logistic regression regarding whether participants had knowledge of the CSI effect revealed that crime show viewing also significantly influenced whether participants reported having knowledge of the CSI Effect in all models (all p’s < .05). The multivariate analyses that explored crime show viewing and perceptions of the CSI effect overall mirrored the findings of the television viewing and socio-demographic variables models.

For the following analyses, we included socio-demographic variables (race, gender, marital status, education, and income) and crime show viewing behaviors as predictors of the three CSI Effect scales (education, trial outcome and evidence) in a series of multivariate analyses. The model examining the effect of crime drama viewing and socio-demographic variables on the CSI Effect education scale revealed that both crime drama viewing and race were significant predictors (p < .05). Participants who reported more hours a week viewing crime dramas also reported higher on the perceptions of the CSI Effect as educational (β = .03, t (151) = 3.38, p < .01). In addition, white participants also were more likely than non-white participants to perceive that the CSI Effect scale was educational (β = .45, t (151) = 2.34, p < .05). The results including the crime show viewing weekly and socio-demographic variables mirrored the above findings. However, the model including crime show viewing daily and socio-demographic variables only revealed race as a significant predictor. That is, again being white significantly influenced participants reporting higher perceptions of the CSI Effect as educational (β = .44, t (147) = 2.19, p < .05).

For the analyses examining the other two CSI Effect scales, most relationships were not significant (all p’s > .05). However, in the model including daily crime show hours and socio-demographic variables; race was a significant predictor of participants’ perceptions of the CSI effect and evidence. That is, participants who were white were more likely than non-white participants to believe that the CSI Effect influenced views of evidence (β = .40, t (154) = 1.92, p = .057).

Discussion

The current study shed some light on how community members’ television and crime show viewing relates to perceptions of the so-called CSI Effect. Generally, people in this sample watch a significant amount of television and quite a few watch crime shows. Additionally, the majority of people reported that they had not heard of the CSI effect. Race and income appeared to influence this relationship. Those who had heard of the CSI Effect defined it as an unrealistic expectation of evidence. When participants described the CSI Effect in their open ended responses, it was only defined as a pro-defense effect (i.e., that jurors would not find the defendant guilty without forensic evidence), which could mean that the media and others are only portraying it as a hindrance on the prosecution. Peoples’ perceptions seem to be in line with the media’s message about the CSI Effect. That is, the news articles describe the CSI as a phenomenon in which defendants are found not guilty due to a lack of forensic evidence (see Willing, 2004). Furthermore, most participants replied that their information about the CSI Effect came from the media, specifically television, which is consistent with cultivation theory’s perspectives of television being the most important medium constructing peoples’ social reality. In reality, the CSI Effect could be either a pro-defense or a pro-prosecution effect; jurors could be inclined to not find a defendant guilty when not presented with forensic evidence (pro-defense), or jurors could weigh forensic evidence (even bad forensic evidence) heavier than other types of evidence (pro-prosecution; Tyler, 2006). Recent research found a marginal CSI Effect where participants who watched crime shows were more likely to favor the defense in certain trial conditions (Hayes-Smith & Levett, 2011).

Overall, we found that the majority of participants had not heard of the CSI Effect. It is possible that those who heard of the CSI Effect (and defined it correctly) would correct for it in a trial situation. That is, if jurors know that the CSI Effect may affect the way they make a decision in trial, and if jurors are sufficiently motivated and able, they may correct for it in their decision making (Fleming et al., 1999; Petty & Wegener, 1993). Future research could investigate this possibility by testing whether knowledge of the CSI Effect affects juror decision making and measuring the motivation and ability of jurors to correct for that bias. While several studies have investigated the effect of watching shows like CSI on jurors’ decisions (e.g., Hayes-Smith & Levett, 2011; Podlas, 2006; Reardon et al., 2007), to date, no studies have examined whether learning about the dangers of a CSI Effect causes jurors to render more legally accurate verdicts.

After giving participants the definition of the CSI Effect, the scaled statements attempting to measure their perceptions of the CSI Effect did not factor into one construct. That is, the scaled items factored into three separate constructs which may indicate there are multiple faucets to consider when measuring or defining the CSI effect. Specifically, the items that were scaled to create the trial outcome construct resemble both a possible pro-prosecution or pro-defense influence on jurors decision making (Tyler, 2006). Overall, the majority of participants believed that watching shows like CSI could affect trial outcomes without specifically describing the nature of that effect. Participant responses to the expectations of evidence construct closely resembles the pro-defense argument, suggesting that the influence of CSI watching on juror decision making lies in the expectation of more evidence (Tyler, 2006). This in turn could relate to the trial outcome construct, thus mediating the relationship. On most items, the majority of community members agreed that jurors expected more evidence as a result of watching shows like CSI. Similar to attorney beliefs about the CSI Effect, this could indicate that community members believe the CSI Effect has potential to be a pro-defense effect.

Our findings also indicate that perhaps another dimension of the CSI Effect warrants further exploration. Specifically, the third scale created from our items measuring perceptions of the CSI Effect measured community members’ perceptions of the CSI Effect as an educational mechanism. Media and criminal justice explanations of the CSI Effect explain it only as a juror influencing phenomenon. The education description highlights that community members may also believe that crime shows are educating criminals and/or the general population. Again, similar to the evidence construct, this could be related to the trial outcome construct in that those who watch crime shows (and are educated either correctly or incorrectly by them) could make different trial conclusions than non-viewers. This construct simply shows that the CSI Effect as defined by community members may go beyond what is defined by the media and criminal justice actors.

In addition to describing participants’ perceptions of the CSI Effect, we explored whether those perceptions varied as a function of whether they regularly watched crime shows. Participants who watched crime shows more often perceived the CSI Effect as more educational than participants who did not watch crime shows as often. This result stayed consistent even when other socio-demographic predictors were entered into the model. This finding is consistent with Watkins (2004) findings on attorneys’ perceptions of CSI; attorneys who reported that they viewed shows such as CSI were more likely than those who did not to think the shows were educational. Similarly, community members who were heavier viewers of crime shows were also less likely than those who were lighter viewers to think that watching crime shows influence trial outcomes negatively. Additionally, those who watched shows like CSI daily were less likely than those who did not to believe that watching CSI-type shows would affect a jury’s verdict or expectations of evidence.

For these effects, crime show viewing was associated with seeing CSI as educational and as influencing trial outcomes, however, there was not a significant relationship between general television viewing and perceptions of CSI as educational and influencing trial outcomes. This suggests that splitting television viewing into show or genre categories may show stronger associations between peoples’ television viewing behavior and their perceptions of social reality, supporting arguments made in previous research (Potter, 1991; Cohen & Weimann, 2000 and Hawkins and Pingree, 1980, 1981). That is, in predicting specific outcomes of television watching (like examining whether watching crime shows is related to peoples’ perceptions of whether watching those shows is educational or will change trial outcomes), it seems that a better predictor is the genre of interest rather than general television watching.

When running multivariate analyses we found that white participants had stronger opinions regarding the CSI Effect than the non-white participants. This could be due to the likelihood that white respondents’ overall exposure to the criminal justice system is different than non-white respondents. Non-whites are more likely than whites to have both formal and informal exposure to the criminal justice system through friends/family because of the disproportionate involvement of people of color. For example, Blacks are more likely to be arrested for homicides than Whites (UCR, 2010). Another example is that Black, non-Hispanics rates for violent victimization and aggravated assault were higher than for White, non-Hispanics (Truman, 2011). In addition, racial profiling is a topic that the public believes exists and is a substantial issue particularly among people of color (Carlson, 2004), therefore, it is no surprise that race is also a predictor of negative opinions towards criminal justice actors, such as law enforcement (Weitzer, 1999). Thus, whites and non-whites may also perceive crime shows differently or gather different amounts of information from those crime shows, which may result in different perceptions of the effect of viewing those shows on juror decision making. Future research on this topic should include questions regarding exposure to the criminal justice system through both informal and formal mechanisms.

In this study, we did not find a strong relationship between several of the measures examining participants’ perceptions of the CSI Effect and television viewing behavior. This could be for a few reasons. First, because the main proponent of the CSI Effect has been the media through stories relating to the CSI Effect as influencing trial decisions, it is possible that participants formed perceptions of the CSI Effect completely separate of having viewed the shows. That is, if the media is socially constructing the CSI Effect through the news, it is possible that participants (both those who watch crime shows and those who do not) had an opportunity to form opinions about the CSI Effect. Perhaps a better predictor of whether people know about the CSI Effect would have been a person’s news watching habits. Or, it is possible that these two viewing habits may interact to affect peoples’ perceptions of the CSI Effect. That is, it is possible that those who watch CSI and the news would be more likely than those who do not to pay attention to this type of news item, and therefore be more likely to know about the CSI Effect. It is also possible that those who watch CSI may be less likely to pay attention to these types of news items than those who do not watch CSI. That is, if an individual believes crime shows are generally accurate, then they may be less likely to pay attention to news that disputes these beliefs. Also, the type of medium communicating information about the CSI Effect (e.g., newspaper, television, and internet) may also matter. Future research could investigate these possibilities.

In our study, we found that participants’ CSI viewing habits were related to their perceptions of the CSI Effect on juries and their perceptions of whether the shows are educational. It is possible that participants were justifying their television viewing habits. That is those who watched shows like CSI may have felt the show was more realistic, and therefore downplayed the fictional nature of the show more so than non-viewers.

Limitations

The limitations of this web-based survey include the usual survey error concerns of sampling, coverage, measurement and non-response (Dillman & Bowker, 2001). However, this survey poses special concerns that should be noted. According to Dillman and Bowker (2001) the explosion of the use of web surveys has not often been conducted with attention to the reduction of survey error. However, the main limitation with this survey is regarding the use of convenience sample.

Due to the exploratory nature of this study, a convenience sample was well suited for this research but does have some strong limitations. The convenience sample rendered an overrepresentation of both female and white respondents along with a high population of heavy television viewers. The make-up of the sample influences the applicability of the questions, as it was likely due to the title of the study that drew in a higher amount of television viewers. While having heavy television viewers is helpful regarding the possibility of them having heard of the CSI Effect, the downside is that they are probably more likely to be agreeable to the attitude scales.

Along with sampling, coverage is a unique issue with web-surveys as older community members may not have access or experience using computers. The requirement of access to and proficiency of using a computer makes it likely to bias the sample towards younger participants. The sample in a previous web-based study was younger than those in our survey, with only 18.2 % being over 42 years old (O’Neal, Penrod & Bornstein, 2003). While the requirement of access to a computer cannot be changed, we took efforts in our study to recruit participants in a venue that had computers available for public access (and many participants completed the survey using those computers). We did not use the age variable in our study due to collection and non-response issues, but there were many responses that indicated the participant’s age was over 60.

To reduce dropout rate, Dillman (2007) suggests using a password, giving more specific computer instructions, and not forcing respondents to answer each question before continuing on to other questions. Based on these suggestions, this survey required a password for continuing, gave specific computer instructions per question, and did not force respondents to answer a question prior to continuing on to other questions. A final limitation is the incapability of measuring non-response rate. The extent of difference between non-respondents and respondents cannot be evaluated due to the methods of passing out flyers and posting the survey online. This limitation in concert with the main limitation of a convenience sample makes it impossible to generalize to a larger population. However, the purpose of this study was primarily exploratory; future studies could use improved methods.

Conclusions

Crime shows appear everywhere on television; one could likely find at least one crime show to watch most hours of the day. It is clear that this sample of community members were watching crime shows, but only a modest amount of people had heard of the CSI Effect. When given the definition, participants varied on whether they believed the CSI Effect exists. People who are watching crime shows are more likely to think crime shows are accurate and educational. Viewing forensic and crime dramas, specifically, appear to be influencing people’s views of social reality and race influenced this perception. However, knowing the CSI Effect exists might ultimately temper the potential influence that it has (if any) on jurors or as York et al. (2006) found knowing more about the CSI Effect could backfire and make the jurors more or less skeptical of forensic evidence. Future research should address these questions.