Investigating the 'CSI effect' effect: media and litigation crisis in criminal law.

INTRODUCTION

I.   TYPOLOGY OF CSI EFFECTS

II.  EVIDENCE OF THE CSI EFFECT
     A. Anecdotes
     B. Surveys of Legal Actors
     C. Juror Surveys
     D. Psychological Experiments
     E. Acquittal Rate Data

III. MEDIA COVERAGE OF THE CSI EFFECT

IV.  A SELF-FULFILLING OR SELF-DENYING PROPHECY?

CONCLUSION

INDEX OF FIGURES AND TABLES

Table 1. Annual Rank of CSI Franchise Programs Among U.S. Television
    Programs
Table 2. Media Mentions of CSI Effect
Table 3. The Many Effects of CSI: Typology of CSI Effects Found in Media
    Accounts
Table 4. Percentage of Respondents Responding "Very Great Prestige" to
    Questions About the Prestige of Selected Professions
Table 5. Acquittal Rates for Nine Jurisdictions in All Years Available,
    Starting 1986
Table 6. Linear Regression Summary of the Relationship Between
    Acquittal Ra Before and After the Airing of CSI in 2001, 2002,
    and 2003 (n = 132)
Table 7. Aggregate Number of Trials and Acquittals from 1997-2006
Table 8. Comparisons Between Aggregate Acquittal Rates Before and
    After the Airing of CSI in 2001, 2002, and 2003
Table 9. Frequency of Various Versions of CSI Effect and Frequency of
    Mention Doubt

INTRODUCTION

Since 2002, popular media has been disseminating serious concerns that the integrity of the criminal trial is being compromised by the effects of television drama. This concern has been dubbed the "CSI effect" after the popular franchise Crime Scene Investigation (CSI). Specifically, it was widely alleged that CSI, one of the most watched programs on television, was affecting jury deliberations and outcomes. It was claimed that jurors confused the idealized portrayal of the capabilities of forensic science on television with the actual capabilities of forensic science in the contemporary criminal justice system. Accordingly, jurors held inflated expectations concerning the occurrence and probative value of forensic evidence. When forensic evidence failed to reach these expectations, it was suggested, juries acquitted. In short, it was argued that, in cases lacking forensic evidence in which juries would have convicted before the advent of the CSI franchise, juries were now acquitting.

The jury is central to American law. The right to a jury trial is "no mere procedural formality, but a fundamental reservation of power in our constitutional structure." (1) Although the jury has been much maligned, the law continues to treat the jury as almost sacred, and many legal scholars and social scientists continue to defend the jury system. (2)

Among the longstanding criticisms of juries has been the claim that juries are subject to media bias. Psychologists have argued that juries can be influenced by pretrial publicity in specific cases, lending support for the need for changes of venue in high profile cases. (3) But, they have also argued that there are more general forms of pretrial publicity, in which media influence may shape jurors' general views about law and crime in ways that affect jury deliberations and verdicts. (4)

The CSI effect is supposedly just such a general pretrial publicity effect. It is alleged that media influence causes potential jurors to have distorted views of the capacity--in the broadest sense of that term--of forensic science to generate evidence in criminal prosecutions. These distorted views, it is alleged, actually affect jury verdicts: cases in which jurors would have convicted absent the media influence of CSI and similar television programming now result in acquittals or hung juries. As we have argued elsewhere, such charges, if true, would constitute a serious challenge to law's continued faith in the jury and thus raise serious questions about the integrity of the criminal justice system itself. (5)

CSI: Crime Scene Investigation is a primetime American television crime drama. It first appeared on the CBS television network on October 6, 2000. (6) CSI is in some sense a standard television crime drama; its innovation is that the protagonists are forensic scientists rather than police detectives, and the plot is driven by the accumulation of forensic evidence rather than the revelation of psychological motive. (7) CSI defied conventional wisdom by daring to try to make science "sexy." This turned out to be a stunningly successful innovation, and CSI surprised the network by becoming, for a time, the most popular television series in the world. (8) Although there is no data on the penetration of CSI or on public awareness of the CSI effect, Nielsen data shows that there is enormous public exposure to the CSI franchise (Table 1). In addition, the program generates even more exposure through reruns. (9) CSI soon became not merely a television series, but a television franchise, and the original program, set in Las Vegas, was "spun off" into CSI: Miami in 2002 and CSI: New York in 2004. (10) There are several shows on television that center on forensic science that we consider to be imitators of CSI, including Without a Trace, Numb3rs, Criminal Minds, N.C.I.S.: Naval Criminal Investigative Service, The Closer, Crossing Jordan, Bones, and The Evidence. (11) In addition to these dramas, numerous forensic-themed documentaries and "reality" television programs could be found on American television, both before and after the advent of CSI, including Anatomy of a Crime, Autopsy, Dr. G. Medical Examiner, Medical Detectives, and The New Detectives. When media discuss the CSI effect, they appear to be describing the cumulative effect of all of this television programming, although the CSI franchise, because of its omnipresence, remains the supposed primary mover of the CSI effect.

The term CSI effect appears to have entered the popular lexicon late in 2002 in an article in Time magazine. (13) That article described "a growing public expectation that police labs can do everything TV labs can." (14) Even in this early article, the notion of jury taint was present: "This [expectation, forensic scientists] worry, may poison jury pools...." (15) The term appeared a couple of times the following year and more in 2004. In 2005, media coverage of the CSI effect exploded. A LexisNexis search found fifty-six newspaper and magazine articles mentioning the CSI effect in that year and seventy-eight articles in 2006, the peak year (Table 2). This coverage included a cover story in U.S. News & World Report, (16) as well as coverage in leading science magazines like National Geographic and Scientific American. (17) Also in 2006, the first full-length book devoted to the CSI effect appeared. (18) Media discourse conceptualizes the CSI effect as what Professor Manning has called a "media loop," (19) a series of back-and-forth interactions between media and what is called, without irony, "reality." (20) The argument is this: Rapid developments and improvements in the power of forensic science inspired media coverage and even fictional television dramas. These media portrayals cause changes in jury behavior in real criminal trials. These changes themselves become the subject of media coverage: media stories about the impact of CSI and similar programs on juror behavior. We refer to this last category of media stories as CSI effect discourse.

Media coverage shows remarkably little equivocation about the existence of the CSI effect. Media reports declare that "[t]here is no debating" the reality of the CSI effect, (21) and that "[t]he story lines are fiction. Their effect is real." (22) It is said that "TV is driving, jury verdicts all across America," (23) that "TV's False Reality Fools Jurors," (24) and that "CSI Has 'Major Effect' On Real Life Juries." (25) An online journal claims that "In many cases across the nation real-life jurors who are fans of CSI has [sic] either caused hung juries or acquitted obviously guilty criminals, claiming the investigators failed to test evidence the way CSI does on television." (26) A jury consultant says that "[t]he CSI effect is real, and it's profound." (27) The accusations leveled at CSI border on charges of jury tampering: one forensic scientist says that CSI is "polluting jury pools." (28) The impact of CSI is portrayed as irresistible: a prosecutor adds, "Jurors are so influenced by television ... that it makes it nearly impossible for us...." (29)

Not only is the media treating the CSI effect as a serious problem, but justice system actors are as well. The FBI has produced a video about it. (30) The Maricopa County Attorney (the presiding prosecutor over much of the major metropolitan area surrounding the city of Phoenix, Arizona) has declared that CSI has a "real-life impact on justice" and has called on CBS to insert a disclaimer on the program stating that it is fiction. (31) In addition to concerns about the integrity of the jury system, some prosecutors have claimed that the CSI effect has altered another pillar of the criminal trial--the standard of proof. They have claimed that jurors are now holding them to a higher standard of proof than the traditional "beyond a reasonable doubt" standard. In closing arguments, prosecutors have called this higher standard the "TV expectation." (32) Several legal scholars have noted that, while the notion that forensically-oriented television programs might influence jurors is theoretically plausible, there is, as yet, no convincing evidence of such an effect. (33) Legal scholars have further noted that, from a theoretical point of view, any media influence on juries would be equally likely to have an effect opposite to that most commonly discussed by the media--that is, forensically-oriented television programming might just as easily make juries more conviction prone as more acquittal prone. (34) Legal scholars have also noted that even if media influences jurors, that by no means necessarily translates into changed verdicts. (35) They have also noted that no increase in reported jury acquittals has been detected. (36) We will report new data on acquittal rates in Part II.

How do we explain this apparent contradiction between media coverage and practice on the ground in U.S. criminal courts? There has been some excellent scholarly work debunking the CSI effect. (37) There has also been some excellent work interpreting the program CSI itself. (38) But there is little work that purports to explain the phenomenon of the CSI effect. In this article, we suggest that we may be able to gain additional insight into the CSI effect by drawing on legal literature emanating from an earlier episode of interaction between media and law. After all, the CSI effect is not the first time that American media has been accused of having perpetuated beliefs about the legal system that are not supported by empirical data. Since the 1970s, American media has reported on a phenomenon it termed the "litigation explosion" (39) or "litigation crisis." Legal scholars described this phenomenon as "hyperlexis," (40) "litigation panic," (41) and "litigation anxiety," (42) among other things. Media reports claimed that litigation was increasing dramatically, that American litigation rates were much higher than those of comparable nations, that punitive damage awards were increasing rapidly, and that the legal system was out of control. In short, the litigation explosion was portrayed as an acute social problem, a "crisis." However, these claims have been widely debunked by socio-legal scholars, who have generally agreed that there has been no dramatic increase in American litigiousness or punitive damages, and that American litigation rates are not wildly out of line with those of comparable nations. (43)

In part, the notion of a litigation crisis was perpetuated by the insurance industry itself through a deliberate and well-funded advertising campaign. (44) However the paid message also penetrated the purportedly objective and therefore more credible mainstream media. Socio-legal scholars have argued that the dissemination of the notion of the litigation explosion was successful not merely because the message was disseminated but also because it resonated with American values such as individualism, responsibility, and self-reliance. (45)

In this Article, we will articulate the noticeable parallels--parallels that have not hitherto been noted in the legal or social scientific literature--between the litigation explosion and the CSI effect. We use the more general term "litigation crisis" to encompass both the litigation explosion and the CSI effect. (46) Echoing the litigation explosion, CSI effect discourse is widely disseminating through the American public the belief that television drama is disadvantaging criminal prosecutions. And yet, the available evidence does not support this claim. Indeed, the available evidence suggests that the opposite may just as easily be the case: forensic-themed police procedural dramas may actually advantage the prosecution in criminal cases. Thus, jurors may come to trial with the counterfactual preconception that the prosecution is disadvantaged, and some of these jurors may unconsciously compensate for that perceived disadvantage.

This Article will be modeled on this analogy with the litigation explosion literature. In the next Part, we will lay out a typology of effects that are all discussed in the media under the rubric of the CSI effect. In Part II, we will discuss the existing evidence in support of the most prominent of these effects, the claim that CSI is changing jury decision making. We show that there is scant empirical evidence to warrant concluding that such changes in jury decision making are occurring. We also introduce acquittal rate data that show only equivocal evidence of an increase in acquittals following the debut of CSI and its spinoffs and imitators. In Part III, we report results of a content analysis that shows that, like it was for the litigation explosion, media coverage is inconsistent with the lack of empirical evidence discussed in Part I. In Part IV and our Conclusion, we attempt to explain the CSI effect as a cultural phenomenon. In Part IV, we suggest that the CSI effect may be a "self-denying prophecy" on behalf of prosecutors; in our Conclusion, we suggest that the CSI effect embodies anxiety about science's threat to the law's role in society as a truth-generating institution.

I. TYPOLOGY OF CSI EFFECTS

As we have discussed elsewhere, the media and its sources use the term CSI effect loosely to denote a remarkable variety of different purported effects. (47) In our earlier work, we proposed a typology of six different specific causal claims that we discerned in the media coverage of the CSI effect, each named for the type of social actor who tended to articulate the supposed effect. Table 3 summarizes each effect. The perhaps canonical effect, which we dubbed the "strong prosecutor's effect," is the claim that television programming is altering juror behavior. Specifically, it is frequently claimed that CSI has increased juror expectations for forensic evidence in criminal trials. Because of CSI, jurors supposedly expect to see forensic evidence more often and expect it to be more probative. This, in turn, could lead to acquittals in cases where forensic evidence is absent or insufficiently probative. In other words, it is suggested that jurors are acquitting in cases lacking forensic evidence in which they would have convicted but for the creation of CSI and similar television programs.

Many prosecutors also make a weaker claim, which we called the "weak prosecutor's effect." This claim posits that CSI has altered prosecutor, not juror, behavior. Claimed changes in prosecutorial behavior include questioning potential jurors about their television viewing habits in voir dire, presenting negative evidence testimony, discussing CSI in summations, and requesting legally unnecessary forensic tests. (48)

Some defense attorneys advance an opposite effect, which we called the "defendant's effect." The claim is that CSI and similar television programming, through their positive and heroic portrayals of state-employed forensic scientists, enhance the perceived credibility of the government's forensic witnesses, thus advantaging the prosecution.

The producers of CSI, in rebutting charges that their product is contaminating the criminal justice system, appropriated the term CSI effect and reinterpreted it as an educational effect on the general public. What we called the "producer's effect" holds that CSI teaches science to the American viewing public.

The "educator's effect," in contrast, claims that CSI is attracting young people into careers in forensic science, much as law programs, such as L.A. Law, have been thought to increase law school applications, (49) medical programs, such as E.R., have been thought to influence medical students' choice of specialty (50) (and perhaps medical school applications as well), and the book and film The Silence of the Lambs has generated countless, mostly unfulfillable, dreams of careers in forensic profiling. (51)

Finally, some media sources posit the "police chief's effect." This claim holds that CSI has educated criminals on how to avoid detection. Examples of the supposed police chief's effect include wearing gloves and dousing crime scenes with bleach. (52)

It is important to emphasize that, of these six effects, only three of them--the strong prosecutor's effect, the defendant's effect, and the police chief's effect--would constitute serious problems for society. If jurors are acquitting defendants that they would have convicted had the television program CSI never existed, this would constitute a serious challenge for the legal system. Such acquittals could, in some sense, be construed as wrongful acquittals. Likewise, if television programming is bolstering the credibility of government witnesses (the defendant's effect), wrongful convictions (in the broadest sense of the term) could result. Either of these effects, if true, would raise serious doubts about the integrity of the jury system that forms the foundation of American criminal justice. If juries are so sensitive to irrelevant influences that the current primetime television schedule has a significant impact on their verdicts, can we really sustain the dogged faith in the jury that remains such a cherished principle of American criminal justice? Finally, the police chiefs effect would generate perhaps the greatest material harm. If true, the police chief's effect could mean that CSI is both increasing crime and decreasing detection of those crimes. However, it should be noted that the police chiefs effect is, strictly speaking, a criminological matter not a legal one. If it were occurring, it would probably be detected and addressed by criminologists, not legal scholars.

The other three effects, on the other hand, would not seem to constitute genuine problems for society. For the weak prosecutor's effect, it would seem to comprise only a minor harm if prosecutors feel compelled to change their voir dire questioning to include asking jurors about their television viewing habits. Prosecutors in cases with little or no forensic evidence might use peremptory challenges to strike heavy CSI viewers from the jury based on the supposition that such jurors would require forensic evidence to convict. Likewise, defense attorneys might strike heavy CSI viewers in cases that rest heavily on forensic evidence based on the supposition that such jurors would be more likely to afford great credibility to the prosecution's forensic expert witnesses. As discussed in the next Part, neither of these suppositions is necessarily correct. Even so, litigants deploying their peremptory challenges in this manner would not seem to pose a significant legal problem. Similarly, prosecutors adopting the practice of explaining the absence of forensic evidence at trial would not seem to undermine the legal system's ability to deliver justice. Although it is true that the ordering of unnecessary forensic tests could constitute a drain on resources and add to backlogs at forensic labs, this, again, is not, strictly speaking, a problem to be solved by the legal system. The producer's effect is posited as a positive effect, provided that the educational aspects gleaned from the show are not wholly unrealistic or inaccurate. So, too, might the educator's effect be considered favorable if it increases the quantity--and thus perhaps indirectly the quality--of applicant pools to forensic science degree programs. (54) There would seem to be few negative repercussions from the educator's effect beyond the disappointment of some young people when they learn that forensic science is neither as exciting nor as glamorous nor as easy as its depiction on television.

In our earlier work, we cautioned that it was necessary to be vigilant against what we called "hypothesis swapping," in which evidence supporting one supposed effect was used to support claims about the existence of a different effect. (55) In particular, it is not uncommon to see evidence of the weak prosecutor's effect advanced in support of claims that the strong prosecutor's effect is occurring. (56) For example, Maricopa County Attorney Andrew Thomas released a study that claimed that jurors are reaching "conclusions contrary to the interests of justice" because of "a significant CSI influence." (57) But, in fact, the study concedes that "verdicts have not yet noticeably changed from guilty to not guilty." (58) Instead, the study has merely found the weak prosecutor's effect: "prosecutors have had to take more and more preemptive steps to divert juries from reliance on television-style expectations." (59) Thus, evidence that supports the weak prosecutor's effect is presented in support of the strong prosecutor's effect.

Since the publication of our earlier article, Donald Shelton has reinterpreted the CSI effect as the "tech effect." (60) He suggests that any apparent changes in juror behavior should not be attributed to television programming, but rather to the underlying real technological developments that these programs depict. (61) Forensic science and technology have advanced enormously over the past century. Shelton asserts that the cause of changes in juror behavior is not CSI but rather the real-life technological improvements in forensic science. (62)

Shelton's argument raises an important caveat about the CSI effect. If we are to take seriously the notion of a CSI effect, it must be carefully disentangled from what Judge Shelton et al. call the "tech effect," the effect of changes in the actual capabilities of forensic science. For example, if, as posited by proponents of a CSI effect, we do find that jurors' expectations for forensic evidence have increased, we would have to assume that this increase is caused by at least two factors. One factor would be jurors' generally accurate perceptions of actual increases in the capabilities of forensic science. The other factor would be jurors' inaccurate perceptions of the capabilities of forensic science as imparted by CSI and similar television programming. Surely, for example, we would not insist that jurors' expectations for forensic evidence should remain completely static. Their expectations should be different today than they were, say, a century ago. In the intervening years, numerous forensic technologies have been developed; we would not expect jurors' commonsensical expectations for forensic evidence to ignore these developments.

Our point is that the baseline against which the CSI effect should be measured is not a static baseline with no change in jurors' expectations for forensic evidence. Presumably, jurors' expectations should, appropriately, increase over time, in response to actual advances in forensic technology. The CSI effect, if there is one, would have to refer to a marginal increase in juror expectations that is excessive of whatever increase in expectations we should reasonably anticipate, given the technological developments that have actually occurred. What this means is that the tech effect, as Shelton et al. characterize it, is not a societal problem. It represents an appropriate increase in juror expectations in response to actual increases in forensic technological capacity. Only the supposed strong prosecutor's effect of wrongful acquittals would represent a genuine problem for the legal system.

For the legal system, clearly, the competing hypotheses of greatest interest are the strong prosecutor's and defendant's effects. Does CSI advantage prosecutors or defendants? Our intuition would be consistent with the defendant's effect: the popularity of television programs that portray forensic scientists as hardworking, virtuous, honest, truthful, heroic, skilled, and attractive should benefit those litigants who employ forensic scientists as expert witnesses. (63) Our intuition would be that positive portrayals of professions in the media would benefit those professions by making the public's perception of them more favorable. (64)

As indicated by public opinion polls, scientists have consistently been ranked among the most prestigious professions over a long period of time, far more prestigious than, say, lawyers or police (Table 4). Having a witness associated with one of the most prestigious professions would seem to advantage the prosecution. (65) Thus, it would seem that the defendant's effect is the more intuitive one. The strong prosecutor's effect is, in fact, a counterintuitive one.

II. EVIDENCE OF THE CSI EFFECT

Although the media coverage portrays the CSI effect as a well-documented phenomenon, actual evidence of the various effects is difficult to come by. (67) Media accounts rely largely on anecdotes and conclusory statements by various criminal justice system actors. Little reference is made to empirical data, and when such references are made, they usually note the absence of such data. (68) A review of the socio-legal literature reveals a rather different picture. Most legal scholars characterize claims of the most common CSI effect--the strong prosecutor's effect which would lead to wrongful acquittals--as speculative, and many suggest that the defendant's effect is equally plausible, even if mentioned less often by the media. (69) In this Part, we review the various forms of evidence that have been mustered in support of claims that there is a CSI effect, and we present new data on jury acquittal rates from state criminal trials that indicate no significant change in acquittal rates in response to CSI.

A. Anecdotes

Socio-legal scholars have noted that media discussions of the litigation explosion relied heavily on anecdotes--"horror stories" about frivolous civil suits that yielded large punitive awards. (70) Such "horror stories" included the notorious McDonald's coffee spill, the poodle in the microwave, and similar such cases. (71) And yet, further examination of these notorious anecdotes has shown that, in many cases, the supposedly outrageous outcomes were actually less egregious than media reports indicated. (72) Media coverage of the CSI effect similarly relies heavily on anecdotes. Perhaps the highest profile anecdote is the acquittal of Baretta television star Robert Blake from charges of murdering his wife, in which District Attorney Steve Cooley called the jury "incredibly stupid." (73) The prosecution provided evidence of motive and opportunity, but forensic evidence was lacking. In particular, Blake tested negative for gunshot residue, which was inconsistent with the theory that he fired the weapon that killed his wife. (74) We have suggested elsewhere that the Blake acquittal may be as indicative of a "celebrity defendant effect" as it is of a CSI effect. (75) However, numerous other lower-profile anecdotes abound--cases in which juries supposedly acquitted based on the lack of forensic evidence and despite the non-forensic evidence presented at trial. There is the sexual assault case in which, despite incriminating DNA evidence, the jury supposedly acquitted because of the failure to test a soil sample from the victim's cervix, (76) or the case in which a juror supposedly wanted a lawn tested for fingerprints. (77) Such stories are perhaps destined to become the equivalent of the McDonald's coffee case. But as with the McDonald's coffee case, there may be more to such stories than conveyed in most media coverage. (78) Such media portrayals present anecdotal evidence based on journalists' interviews with prosecutors and jurors who claimed that the acquittals were in fact due to jurors' increased expectation of forensic science evidence and techniques based on the television depictions in forensic programs such as CSI. (79) It is possible that the jury had good reasons for acquitting. For example, in reference to the Blake case jurors, Professor Laurie Levenson remarked, "[i]t was a reasonable-doubt case, and disagreeing with [Mr. Cooley, the District Attorney,] doesn't make them stupid." (80)

B. Surveys of Legal Actors

Another form of evidence cited in support of claims of a litigation explosion were surveys of corporate executives which reported that fear of litigation influenced business decisions, sapped competitiveness, and stifled innovation. (81) There are obvious methodological difficulties with treating such surveys as actual evidence of the supposed pernicious effects of the litigation explosion. First, even executives' genuine perceptions of the degree of litigiousness might not be accurate. Second, the opinions expressed in such surveys might have been self-serving given that part of an executive's job description includes communicating in a manner than benefits the corporation.

Similarly, the second major form of evidence cited in support of the CSI effect is opinion surveys of legal actors: prosecutors, defense attorneys, and trial judges. These surveys focused on the perceived impact of forensic programs on jury verdicts, pretrial preparation, and trial strategy. Survey results generally indicate that, according to legal actors, the CSI effect is real and has had considerable impact on the carrying out of criminal trials.

Watkins's study surveyed approximately fifty-three prosecutors, public defenders, and private defense attorneys. (82) The questions included the perceived impact of forensic television programs such as CSI on their pretrial preparations and on jury verdicts. He found that seventeen percent of prosecutors adjusted their pretrial behavior by requesting more forensic tests since the advent of CSI, fifty-five percent included questions on jurors' forensic television program viewing habits during voir dire, and forty-nine percent claimed to observe actual acquittals in cases where they felt that there was sufficient circumstantial evidence presented at trial to warrant a conviction.

Similarly, the Maricopa County Attorney's Office surveyed 102 prosecutors with jury trial experience as to whether they perceived a CSI effect among juries. (83) Although there were no reported changes in the verdicts or rates of acquittal in Maricopa County, prosecutors reported changes in their own behavior pretrial and at trial. They reported taking preemptive steps to ensure that juries did not rely on inflated expectations of forensic evidence when determining their verdict decisions. Thirty-eight percent of the prosecutors surveyed reported at least one wrongful acquittal or hung jury that was the result of a lack of forensic evidence at trial, where they felt that the non-forensic evidence was sufficient to sustain a conviction. Accordingly, Maricopa County Attorney Andrew Thomas concluded that a CSI effect is "no myth." (84)

Monica Robbers found, like Watkins and Thomas, that legal actors believed that jury decisions had been influenced by inflated expectations of forensic evidence based on forensic programs, such as CSI. (85) Robbers surveyed approximately 290 prosecutors, defense attorneys, and trial judges sampled from all state counties and cities in the United States. She asked respondents to discuss specific instances in which they believed that jury verdicts were influenced by CSI viewing and whether CSI and forensic television programs affected their jobs generally. The general findings from this survey were that the majority of legal actors (79%) reported some specific instance in which they felt juries' decisions were influenced by forensic programs. In their reports of these cases, they cited a perceived jury preference for forensic evidence (50%), the jury discounting eyewitnesses (53%), and an increased use of negative evidence witnesses (28%). (86) In another survey, conducted by Dennis Stevens, prosecutors also reported that they themselves and other legal actors were heavily influenced by forensic television programming. (87)

Surveys of prosecutors, defense attorneys, and judges offer rich data as to how threats of a CSI effect have changed their behaviors at trial. However, such studies measure what we have called the weak prosecutor's effect. They measure attorney, not juror, behavior, and they tell us little about the effect with which the media is most concerned: the strong prosecutor's effect--the wrongful acquittal of criminal defendants. The surveys show that legal actors believe juries are affected by forensic television programs. Surveys also show that prosecutors engage in countermeasures, including questioning prospective jurors about the program in voir dire, presenting negative evidence witnesses to explain why forensic evidence was not found, increasing the time spent establishing the credibility of eyewitnesses and other non-forensic witnesses, and requesting forensic testing and procedures more frequently.

However, these surveys provide very little supporting evidence for the strong prosecutor's effect, which people typically think of when they think of the CSI effect. Many of the questions on these surveys focus on effects on legal actors' behavior, not jurors' behavior. Moreover, even when these surveys do seek to measure juror behavior, they do so indirectly. Rather than examining jurors about whether they were influenced by CSI, these surveys ask legal actors whether they think jurors were influenced by CSI. Thus, these surveys measure not whether jurors were influenced by CSI, but whether legal actors perceive jurors to be influenced by CSI. In some cases, the legal actors claim to have spoken with the jurors; in others, they offer opinions about jury behavior without having spoken to the jurors. In either scenario, this makes for a poor measure of juror behavior, not merely because it is indirect, but also because legal actors are hardly unbiased perceivers of jurors' behavior and motivations. (88) Asking a prosecutor who has just completed a trial whether a jury acquittal was caused by the CSI effect, rather than by, say, insufficiency of the evidence, is clearly an unsatisfactory way of measuring whether the jurors were indeed influenced by CSI. Thus, in order to measure the strong prosecutor's effect, it would be far preferable to survey jurors directly.

C. Juror Surveys

Two forms of evidence are available to CSI effect claimants that were not available to litigation-explosion claimants. One is to survey jurors, or potential jurors, in an effort to determine whether individuals with high exposure to CSI hold different attitudes toward forensic evidence than individuals with low exposure. The second is to conduct psychological experiments that simulate juror decision making to, again, determine whether high-exposure individuals behave differently from low-exposure individuals in simulated jury decision making.

In order to determine and measure whether there is a CSI effect, several studies surveyed jurors or potential jurors. In studies of this type, jury decision making can be compared between groups who watch CSI (and other forensic television programs) with those who do not.

Kimberlianne Podlas attempted to detect the CSI effect by using a rape trial scenario with a consent defense where forensic evidence was neither provided nor necessary. (89) She surveyed 306 college students and asked them to reach a verdict of guilty or not guilty, where the expected or "legally correct" verdict for the case was not guilty. Podlas compared students who regularly watched forensic television programs with those who did not and found that there were no significant differences in their decision-making processes or the handing down of "not guilty" verdicts. The survey results did not indicate any increased expectation of forensic evidence by CSI viewers compared to non-CSI viewers.

To test the effects of CSI on jurors' expectations, Shelton administered a survey to 1027 individuals called to jury duty in a county in southeast Michigan. (90) Respondents were asked about their television viewing habits of crime dramas such as CSI and were presented with various scenarios of criminal cases and charges. Respondents were asked what types of evidence they expected to be presented at trial and what verdict they would hand down based on certain types of evidence presented by the prosecution and the defense. The results indicated high expectations for forensic evidence by all subjects, and CSI viewers had higher expectations of all types of evidence (forensic and non-forensic) than did non-CSI viewers. Any differences found between CSI and non-CSI viewers were marginally significant and were counter to the strong prosecutor effect. Respondents' increased expectations of scientific evidence did not translate into a requirement for handing down guilty verdicts. For example, CSI viewers were more likely than non-CSI viewers to find eyewitness testimony valuable when reaching a verdict without any scientific evidence. The authors suggest that the increased expectations of forensic evidence might have little to do with whether or not jurors watch forensic television programs, but instead they may reflect a broader notion in society of an increased awareness about technological advances. Instead of a CSI effect, they posited a general tech effect experienced by everyone.

Kiara Okita surveyed more than 1200 Canadians about their attitudes toward forensic science. (91) Like Shelton, Okita found that CSI viewers and nonviewers did not differ significantly in their perceptions of the accuracy and necessity of forensic science for investigating crimes. Indeed, in some cases nonviewers perceived forensic science to be more accurate than viewers did. However, Okita notes that even those differences between viewers and nonviewers that she did find were so small that they were unlikely to be operationalized, say, by producing different verdicts. As she summarizes:

   Regardless of CSI viewership, respondents' [sic] appear to consider
   forensic science, in general, to be somewhere between accurate or
   usually accurate, and between somewhat necessary and necessary in
   determining criminal guilt. Therefore again, contrary to the
   assertions of CSI effect claimants, respondents do not appear to
   perceive forensic science as completely accurate and always
   necessary in determining criminal guilt. (92)

Finally, in a series of studies, Steven Smith and colleagues found evidence of changes in legal professionals' behavior (the weak prosecutor's effect), and found evidence suggestive of the defendant's effect. (93) But they found little evidence supporting the strong prosecutor's effect.

D. Psychological Experiments

Another approach to measuring the CSI effect is to conduct simulations of jury deliberations using mock jurors, usually college students. Although college student populations are not representative of actual jury pools, jury simulations allow researchers to conduct controlled experiments. Kimberlianne Podlas's second study included 538 mock jurors who deliberated in small groups about two crime scenarios where forensic evidence was neither necessary, nor referenced. (94) Podlas tested for the strong prosecutor's effect as measured by not guilty verdicts or wrongful acquittals for each scenario. She found no indication that CSI viewing jurors acquitted in cases that warranted convictions, nor did she find that CSI viewers relied on forensic evidence to a greater degree than their non-CSI viewing counterparts.

While previous studies examined the strong prosecutor's effect of potential jurors acquitting defendants due to their increased expectation of forensic science, Schweitzer and Saks tested both the strong prosecutor's effect and the defendant's effect--that potential jurors who watch CSI have exaggerated faith in the capabilities of forensic science and give this evidence more weight than it may deserve. (95) Their sample included 48 college students who were asked to review a transcript of a mock criminal trial where the key piece of inculpatory evidence was a hair left at the crime scene. They included testimony by a forensic scientist who performed the microscopic hair analysis that identified the defendant, which overstated the probative value of the evidence--something that is apparently not uncommon for hair evidence. (96) Subjects were asked about their television viewing habits and their perceptions about the case and forensic evidence presented. CSI viewers perceived themselves as having a better understanding of forensic scientists and their techniques than non-CSI viewers, and they were more critical of the forensic evidence presented in the transcript. Schweitzer and Saks interpreted this as indirect evidence of an increased expectation of high-tech forensic science perhaps consistent with CSI's depictions of forensic techniques, and a tendency to find less high-tech or glamorous techniques less convincing. Similarly, in a study of 140 college students, Jenkins found CSI viewers more sensitive to possible flaws in a forensic assay that was discussed in a mock trial transcript. (97)

E. Acquittal Rate Data

Socio-legal scholars skeptical of claims about a litigation explosion examined time-series data on the number of civil trials held over various periods of time. Contrary to litigation explosion claims, they concluded that there was little empirical evidence of a litigation explosion. (98) These scholars argued that it was irresponsible to disseminate claims about a supposed litigation explosion if empirical data did not support such claims. (99)

We can also adopt such an approach to claims about the strong prosecutor's effect. Even if surveys and jury simulations did provide evidence for the strong prosecutor's effect, one would presumably want to look for changes in the rate of jury acquittals in American criminal trials before concluding that CSI is influencing jury verdicts. The strong prosecutor's effect holds that jurors are acquitting in cases in which they would have convicted had the television program CSI never existed. If this effect is occurring, therefore, it would be expected to manifest itself through an increase in jury acquittals following the advent of the program. In earlier work, we examined data on the jury acquittal rate in federal courts, and we found no discernable increase in acquittal rates following the advent of CSI. (100) In fact, if anything, there appeared to be a decrease in the acquittal rate after CSI. Were this decrease to be significant it would support the defendant's effect, the claim that CSI actually benefits prosecutors. Recall that in some legal scholars' view this effect is equally theoretically plausible. (101) Loeffier supported this finding, determining that there was no evidence of an increase in acquittals after examining the acquittal rates of four large states. (102) Looking at Canadian data, Benoit Dupont also found no discernable increase in acquittals that was attributable to CSI. (103) Although overall Canadian acquittals did rise after 2000, Dupont notes that acquittals had been steadily rising for a long time before 2000, and he concludes that CSI does not appear to have had an influence on this trend.

In new research we report here, we sought to carry this project forward by gathering acquittal rate data from all U.S. jurisdictions. Over the course of six months, we conducted internet searches of state administrative offices of courts' websites and follow up contacts via email and phone with state court administrators. We surveyed all fifty states to determine whether suitable acquittal rate data was available. We were able to obtain acquittal rate data on felony jury trials from eleven states. (104) However, the states varied in terms of how long they had been compiling dispositional data from criminal jury trials. They ranged from Florida, which has such data from as far back as 1986, to Kentucky, which began compiling data in 2006. There were only eight states for which we were able to ascertain acquittal rates both before and after the advent of CSI in 2000: California, Florida, Hawaii, Illinois, New York, North Carolina, Texas, and Vermont.

We compiled acquittal rates for all jury verdicts from the felony trial courts of these eight states and the federal district courts. (105) In most cases, these were restricted to felony trials, although the types of crimes were not designated in this data. In one case, the acquittal rates include a small number of misdemeanors because the felony trial courts (the California Superior Courts) adjudicate a small number of misdemeanors as well as felonies. (106) Although data on jury verdicts in misdemeanor trials were available for a few states (Texas, Florida, and Vermont), we opted not to include this data in our analysis because of the strong possibility that misdemeanor jury trials differ from felony jury trials.

We were able to obtain data reporting the outcomes (conviction or acquittal) for all felony jury trials (plus a small number of serious misdemeanor trials from California) that went to verdict for the states and years shown in Table 5. It is apparent from Table 5 that acquittal rates are fairly stable over time although, not surprisingly, they fluctuate far more in the two smallest jurisdictions. Indeed, the data show a strong main effect of jurisdiction on acquittal rates. In other words, each jurisdiction's acquittal rate appears to be quite stable over time, and jurisdiction appears to have a far greater influence on the probability of acquittal than does year. A defendant would be better off being tried in Florida than in California both before and after the advent of CSI, and, even if there is a CSI effect, it would appear to be a minor issue compared to the influence of jurisdiction.

Although there is no reason to expect that any CSI effect would be felt differentially in different states, the various states do not follow a wholly consistent pattern. Overall, there does appear to be a slight rise in acquittals in 2001 and 2002. Interpreting this small rise as the strong prosecutor's effect, however, raises several concerns. First, after 2002, the acquittal rate drops back to 1998-2000 levels, suggesting that even if there had been a strong prosecutor's effect in 2001-2002, it was short-lived (or prosecutors compensated for it by ceasing to bring susceptible cases to trial or through countermeasures). Second, the acquittal rate was already rising before the advent of CSI. The acquittal rates of 2001-2002 might simply be extensions of this trend, rather than reactions to CSI. Third, the aggregate acquittal rate in 1996 was as high as the post-CSI acquittal rates of 2001-2002. It may be that it is the 1997 drop in acquittals that requires explanation, rather than the increase that followed it. At the same time, the trend toward a drop in acquittals that we noted earlier in the federal courts appears even more pronounced now that we have data for two additional years. Whereas in our prior study we found a drop to an 11% acquittal rate in 2005, we now see that the acquittal rate has remained at this historically low rate for three consecutive years.

In short, the acquittal rate data would seem to offer only equivocal support for only a very small and short-lived strong prosecutor's effect. Can we conclude anything more definitive from this data? How to deal with time-series data of this sort is not obvious. For instance, it is not clear whether the pre-CSI acquittal rate should be treated as the acquittal rate for 2000, the year immediately preceding the advent of the program or the aggregate acquittal rate of a greater range of years, such as 1997-2000. Similarly, it is not clear how best to account for temporal trends in the acquittal rate that preceded the advent of CSI. We were not able to find any studies that attempt to model changes in jury verdicts in response to a historical event. Without any such study in hand upon which to model our analysis, we conducted two analyses.

1. Analysis 1

In order to test whether or not acquittal rates significantly changed due to the advent of CSI, we conducted pooled time-series cross-section analysis, which combines a time-series analysis within several cross sections. In our case, the acquittal data is characterized as repeated observations (acquittal rates) by year (1986 to 2008) within fixed states. We had a total of 132 observations/acquittal rates for all nine jurisdictions. One limitation of our analysis is that this constitutes a relatively small sample size. Another is that errors from regression equations from pooled data may not be independent over time, and, indeed, as discussed above, they appear not be. Third, we cannot control for the possibility of heteroskedasticity where there may be differing variances across the ranges of acquittal rates for any given state; meaning, different jurisdictions' acquittal rates vary differently, rather than consistently.

We conducted linear regression modeling comparing acquittal rates before and after the first airing of CSI in 2000. If CSI viewership had an effect on jury verdicts, we would expect a change in acquittal rates as early as fiscal year 2001. We also looked at differences in acquittal rates in the following three years after the first airing of CSI, to account for the possibility of a lag effect in which CSI did not have an immediate impact but did have an impact after some years of media saturation. Indeed, one might expect that the CSI effect would be felt after a large number of aggregate exposures to CSI and similar programs.

We conducted linear regressions of acquittal rates before and after 2000, after the first airing of CSI and found no statistically significant difference. That is, we treated year and state as the independent variable and acquittal rate as the dependent variable. When we tested the change in acquittal rates between these two groups (pre- and post-CSI), we found that the difference between them may have occurred due to chance or by coincidence, rather than inferring the events are somehow correlated. When we tested before and after the years 2001, 2002 and 2003, we found no statistically significant difference in acquittal rates before and after any of these years. Table 6 presents the results of this regression. These findings indicate that for any given comparison year, there was less than a one percent increase in acquittal rates after the introduction of the program CSI and that this difference is not significant at p<.05 level. These results suggest that the changes in acquittal rates following the introduction of CSI are very likely the result of chance, and, certainly, the possibility that they are due to chance cannot be ruled out.

2. Analysis 2

One disadvantage of Analysis 1 is that it treats each state's annual acquittal rate as a single observation. But such observations refer to a great many more trials in the case of California than in the case of Vermont. Analysis 2 overcomes that disadvantage by treating each trial that went to a jury verdict as an observation. In other words, we treated our data as if it was a random sample of American jury verdicts. (108) States were irrelevant in this analysis, except as a means of obtaining a sample of American jury verdicts. In order to keep the sampling method constant, we analyzed trial outcomes from the same set of states before and after the advent of CSI. Since Vermont only provided acquittal rate data starting in 1998, we chose to exclude Vermont, so as to include data from 1997. (109)

We calculated the proportions of acquittals for periods before and after the advent of CSI. We then calculated the difference between these proportions and confidence intervals surrounding this difference in order to test whether or not there was a statistically significant change in the proportion of acquittals. The total number of felony trials in any given year ranged from approximately 18,000 in 2005 to over 24,000 in 1997 (Table 7). We calculated the proportion of acquittals pre-CSI (from 1997-2000). We then calculated the proportion of acquittals in three sets of post-CSI years (2001-2006, 2002-2006, 2003-2006), again in order to account for the possibility of a lag effect. We then calculated the differences between these proportions and the associated confidence intervals. The results of this analysis are shown in Table 8. We see that there is a statistically significant increase in acquittal rates from the years pre-CSI to post-2001 and post-2002, but not post-2003.

As discussed above, there are a number of plausible explanations for this increase in addition to a two-year CSI effect. One is that this increase in acquittal rates post-CSI may be attributed to the general trend of rising acquittal rates beginning 1997 and may not be attributable to any CSI effect. This possibility may be rendered more tangible by row four of Table 8, which shows that there was also a statistically significant increase in acquittals between 1997-1999 and 2000, the year before CSI went on the air. To better control for a possible trend prior to CSI, in rows five through seven, we compared the proportion of acquittals from 2000 to post-CSI years (2001-2006, 2002-2006, 2003-2006). There was no longer a significant increase in acquittal rates. In fact, there appears to be a statistically significant decrease in the proportion of acquittals comparing 2000 to post-CSI 2003-2006. It should also be noted that the statistical power in this analysis is very high, making statistically significant findings in either direction more likely.

Another possibility is that two or more different CSI effects may be canceling each other out. For instance, the strong prosecutor's effect and the defendant's effect might both be occurring and canceling one another out, one driving acquittals up, the other driving them down. Or, the strong prosecutor's effect may be occurring, but prosecutors may be compensating by not bringing the affected cases to trial, by effectively screening out jurors affected by it in voir dire, or by effectively explaining the absence or weakness of forensic evidence. Under such a scenario, the CSI effect would be occurring, but it would not be detected in acquittal rates. In a sense, the strong prosecutor's effect would be cancelled by the weak prosecutor's effect. Or, unknown other historical changes for which we have not accounted may have affected the acquittal rate during the period we analyzed and may have counteracted the CSI effect. For example, the September 11, 2001 attack on the World Trade Center and Pentagon comes to mind as an event that falls within the period of analysis that might have conceivably had an impact on acquittal rates. Similarly, a major legal change, such as a landmark Supreme Court ruling on evidence law might shift acquittal rates, although we are not aware of such a decision during the relevant period. In sum, given the equivocal nature of the data and the relatively small changes in acquittal rates, existing acquittal rate data would not seem to warrant panic about the existence of a CSI effect.

III. MEDIA COVERAGE OF THE CSI EFFECT

The dispute over civil litigation reform has been characterized as a battle between public intellectuals arguing for tort reform and drawing attention to a supposed litigation explosion and social scientists defending the civil litigation system against these charges. (110) Tort reformers relied largely on culturally resonant anecdotes for evidence. Indeed, Haltom and McCann's analysis of the empirical evidence contained in one prominent book in the tort reform canon found 272 anecdotes, one case study, and six citations to statistical evidence. (11l) Social scientists, meanwhile, relied on statistical evidence about the volume of civil litigation to debunk claims of a litigation explosion or that Americans were exceptionally litigious. "[T]he resulting contest," Haltom and McCann conclude, "has been a mismatch heavily favoring the scholars on intellectual grounds, but a veritable triumph for the reform proponents in the broader cultural terrains of American mass society." (112) Tort reformers' narratives captured a much greater share of media attention than did those of scholars. In addition, a concerted insurance industry advertising campaign disseminated the tort reform message.

Two scholars even went so far as to call the contest for the hearts and minds of the American public a "slaughter." (113) Indeed, in the wake of media claims about the litigation explosion, socio-legal scholars documented that media coverage of civil law overwhelmingly emphasized plaintiff victories and high punitive damage awards. (114) Thus, the supposed existence of a litigation explosion became educated common sense among jurors (115) and even judges. (116)

The result of this apparent bias in media coverage was that consumers of popular media were likely to have distorted understandings of the American civil legal system. Such consumers were likely to believe that there was a litigation explosion when, in fact, there was none. These consumers might have believed that the plaintiff victory rate was higher than it actually was, that average punitive damages were higher than they were in reality, and that this phenomenon imposed severe social costs (such as increased insurance premiums or more expensive consumer goods). The potential impact of this media bias on civil trials was not lost on socio-legal scholars. Elizabeth Loftus performed an experiment that "showed that it is possible to affect a jury award by even a single exposure to an insurance advertisement." (117) Similarly, in interviews with individuals who had completed service on civil juries in tort cases, Valerie Hans and William Lofquist found that most jurors believed that the litigation explosion existed and that many jurors believed it had negative social impacts. (118) They also found that the jurors who believed in the litigation explosion gave lower awards. (119) Hans and Lofquist stress that this result must be interpreted with caution, and, indeed, some contradictory evidence exists. (120) For example, another jury simulation study, although finding that "jurors who agree that damage awards are excessive and that tort reform is necessary generally gave lower awards," also found that belief in the litigation explosion drove awards higher. (121) The authors hypothesized that mock jurors calibrated their awards to what they perceived as the "going rate." (122) Overall the effect of knowledge of the supposed litigation explosion on civil juries appears to be "complex." (123)

In this Part, we apply a similar approach to the CSI effect by examining the message conveyed by media about the CSI effect. Despite the lack of empirical support described in the preceding section, a review of media coverage shows that it tends to characterize the CSI effect as a well-established phenomenon. For example, based on a content analysis of seventy mass media articles about the CSI effect published between 2002 and 2005, Harvey and Derksen found that the majority of coverage claimed that jurors had formed unrealistic expectations of forensic evidence because of CSI. (124) In order to investigate this issue further, we conducted a content analysis of 258 magazine and newspaper articles discussing the CSI effect. We generated this data set by searching for the terms CSI effect or "C.S.I. effect" in the "Magazines, combined" and "Newspapers, combined" databases in LexisNexis for the years 2002 (after using LexisNexis to determine that was the earliest year the term appeared in the media) through 2008. These searches generated 504 documents. After eliminating duplicates, academic articles, and a few other sources that were not relevant to our research questions, or sources which did not seem to qualify as popular media, we were left with 258 documents published by media sources, two-thirds from domestic articles and one-third foreign. (125) Around fifty-five percent of the documents came from local U.S. newspapers, around four percent originated from national magazines like Macleans, U.S. News & Worm Report, and the New Yorker, and around nine percent derived from technical or legal magazines, such as Lawyer's Weekly, Business Monthly, New Scientist, and Chicago Lawyer. Although we confine our analysis to 258 unique articles in order to avoid double coding reprinted stories, the 504 LexisNexis "hits" might be a better indication of the penetration of the notion of the CSI effect in the popular consciousness. It should also be noted that LexisNexis misses many media sources. For example, LexisNexis is limited in its capture of broadcast television content, and it is completely ineffective in capturing new media sources like online journals and blogs. For all these reasons, we think that more than five hundred articles in six years still probably understates the penetration of the notion of the CSI effect.

Although some of the articles were published in major media outlets like U.S. News & Worm Report, one notable aspect of these articles is how local CSI effect stories are. Unlike many news stories, including the litigation explosion, (126) in which many local papers publish the same story provided to them by a wire service, CSI effect stories lend themselves to what communications scholars call "localization," in which the same story is rewritten using local characters. (127) As we have posited elsewhere, CSI effect stories seem to lend themselves to this sort of treatment because every locality has its own forensic technician, prosecutor, and police chief, who serve as the standard "cast of characters" in CSI effect stories. (128)

We coded the documents for any discussion of the six effects defined above. Not surprisingly, many documents mentioned more than one effect. During the coding process we also discovered some new "effects." Only one of these, which we call the "victim's effect," was prevalent enough to add to our coding scheme. The others were not prevalent enough to warrant further discussion. (129) The victim's effect concerns the supposed effect of CSI programs on crime victims' behavior. The claim is that victims have increased expectations that law enforcement will collect forensic evidence at a crime scene.

We also coded whether the documents treated the specified CSI effect as real or whether they articulated doubt about whether the specified CSI effect was actually occurring. Documents were coded for doubt if, for example, they made statements like "[t]he 'CSI effect' is largely the product of anecdotal evidence." (130) Many documents that were coded for doubt also included statements asserting the reality of the CSI effect.

Harvey and Derksen found that the strong prosecutor's effect was far more salient in the media than the defendant's effect. (131) Discussions of negative impacts on jurors far outweighed positive ones. In addition, Harvey and Derksen coded whether media reports discussed positive or negative impacts for prosecutors and defense attorneys. The most prevalent category was negative impacts for prosecutors. Moreover, there were more media reports of positive than negative impacts for defendants. (132) Our findings were similar, not surprisingly (especially because our data sets overlap). As shown in Table 9, the primary version of the CSI effect found in the media is the strong prosecutor's effect of increased juror expectations, which dwarfs all other effects. The rarity of the defendant's effect is also quite striking. In instances where defense attorneys were interviewed, they often mentioned altering their behavior to highlight the lack of forensic evidence at trial (in support of the strong prosecutor's effect), rather than raising the issue that CSI has led to jurors viewing government experts and forensic science techniques as having increased credibility. The strong prosecutor's effect appears in the media around seven times as often as the defendant's effect (197 appearances for the strong prosecutor's effect versus 21 for the defendant's effect). If the CSI effect is seen as strategic gamesmanship by prosecutors and defenders, our analysis of media content shows that the prosecutors are clearly doing a better job disseminating their message to the media. As with the litigation explosion, it is a "slaughter." (133) This is particularly striking because, as some commentators have noted, the defendant's effect is equally plausible (134) and, we suggest, more intuitive. One reason for this may be that in both cases, the media find stories of an effect and a social problem--the litigation explosion or the CSI effect--more appealing than stories extolling the unheralded virtues of the status quo.

In addition, we found that media discussions of the CSI effect gave voice to remarkably little skepticism about the claims that the CSI effect is actually occurring (Table 8). For example, of 197 documents mentioning the strong prosecutor's effect, only 34 gave voice to skepticism about the effect actually occurring. This is particularly striking, given that most legal scholars have expressed doubt that CSI actually has changed jury behavior. Indeed, several of the 34 articles voicing doubt are profiles of academics, like Donald Shelton or Kimberlianne Podlas, who have done empirical research that casts doubt on the claims advanced on behalf of the strong prosecutor's effect. (135)

In sum, our analysis showed that the media is quite broad in its use of the term "CSI effect," using it to convey a wide variety of quite different and sometimes even incompatible ideas, often in the same article. At the same time, it is clear that, by and large, in media discourse, CSI effect means the strong prosecutor's effect--an influence on jury behavior in the direction of acquittal. In addition, the media tends to portray the CSI effect as an established phenomenon. For example, one article states, "[i]n an alarming number of cases, jurors found people not guilty of serious violent crimes because they believed police should have presented more, or different, forensic evidence." (136) Media consumers, therefore, are likely to believe that CSI is impacting juror behavior, making convictions more difficult to achieve.

Aside from the question of whether the CSI effect is actually occurring is the question of whether it constitutes a serious social problem. The media discussions of the CSI effect are remarkable for their alarmist tone. Media discussions of the litigation explosion invoked "[i]mages of a destructive, elemental force" and described it with terms like "epidemic," "avalanche," "flood," "tidal wave," "deluge," "apocalypse," and "doomsday." (137) Similarly, media discussions of the CSI effect use terms like "alarming," (138) "dangerous," (139) and "a big problem," (140) and they suggest that the CSI effect "could have serious ramifications for our justice system." (141) This is striking because, as discussed above, empirical evidence supporting this claim is conspicuously lacking.

IV. A SELF-FULFILLING OR SELF-DENYING PROPHECY?

As with the litigation explosion, the current media view of the CSI effect is, in some respects, the outcome of a conscious effort on the part of litigants' lobbying organizations: corporate defendants and their public intellectual allies in the case of the litigation explosion; prosecutors in the case of the strong prosecutor's effect. As socio-legal scholars have pointed out, widespread popular belief in the litigation explosion would seem to have at least two effects that would benefit the interests of civil defendants. First, policymakers convinced that there is indeed a litigation explosion and that it constitutes a significant social problem might be amenable to regulatory changes that would render civil litigation less desirable for plaintiffs--tort reform. (142) Proposed reforms include creating disincentives for litigation, increased sanctioning of frivolous cases, promoting alternative dispute resolution, and creating alternative compensation systems. (143) A second effect of media coverage, as discussed above, might be on the jury pool itself. Jurors who believe that frivolous lawsuits are common and are causing economic and social damage to the United States, that punitive damage awards are out of control and are causing economic and social damage to the United States, and that plaintiffs usually win civil lawsuits may be more sympathetic jurors from the perspective of civil defendants. This constitutes a sort of second-order version of media effects on juries: the impact that awareness (or purported awareness) of previous jury decisions obtained through media may have on current jury decisions. As Hans and Lofquist noted, current jurors' "concerns about deep pockets, the litigation crisis, and the integrity of plaintiffs were implicitly and explicitly linked to the presumed excesses of antecedent juries." (144) Such juries, they further noted, might try to correct for the perceived excesses of antecedent juries. (145) And yet, it is important to note that current jurors' perceptions of the general trends among antecedent juries may be media-biased and incorrect. (146) A jury that attempts to correct for perceived pro-plaintiff bias when in fact there is not pro-plaintiff bias, is, in effect, enacting anti-plaintiff bias.

The same argument would seem to hold for media coverage of the CSI effect. Jurors who are consumers of the popular media might believe that prosecutors are typically disadvantaged in criminal trials; that high expectations for forensic evidence are "unreasonable"; and that criminal convictions are becoming increasingly rare and difficult to achieve. Jurors who believe these things might be more sympathetic to prosecutors out of sympathy for the perceived underdog or in attempt to correct for the perceived excesses of antecedent juries. (147) Claiming to be disadvantaged is a familiar trope in trial advocacy, especially in opening and closing arguments; prosecutors frequently point out that they bear the burden of proof, whereas defense attorneys often refer to their lack of resources or to the awesome power of the state.

As with the litigation explosion, there may be a second-order media effect on juries in criminal law. We might call this the "CSI effect effect": juries that have become convinced through media that there is a strong prosecutor's effect that disadvantages prosecutors and has led antecedent juries to acquit inappropriately might tend to sympathize with the prosecution and enact a seemingly "corrective" pro-prosecution bias. But, if there is, in fact, no strong prosecutor's effect, the CSI effect effect is essentially an inappropriate proprosecution bias. As Okita notes, "the 'CSI effect' may not be an effect caused by the media, but one which has instead been promulgated by the media." (148)

The Maricopa County Attorney proposed to actually enact the CSI effect effect, using CSI itself, rather than the news media. He proposed that CSI "show the CSI effect in action" and communicate the message that the strong prosecutor's effect is a damaging social phenomenon, and proposed potential storylines:

   Programs in which jurors use outside influences and prejudices to
   "supplement" the facts of the case presented in court, or in which
   a jury foreman with an addiction to television crime dramas uses
   his "expertise" to intimidate and cajole young, inexperienced
   jurors, might be instructive. It might also be possible to present
   a case in which both sides put on heroic court performances but the
   jury ends up freeing a criminal who committed a serious crime, like
   child molestation, because of these influences. (149)

Ultimately, the strong prosecutor's effect, as easily as it might be--as Harvey and Derksen describe it--"a self-fulfilling prophecy," (150) might just as easily be a self-denying prophecy. By disseminating through the media the notion that the CSI effect is occurring, prosecutors may be preventing the strong prosecutor's effect from occurring. And, if the strong prosecutor's effect is not occurring, this counteraction may in fact be creating a new effect of its own, advantaging, rather than disadvantaging prosecutors.

CONCLUSION

And yet, as with the litigation explosion, we cannot attribute media interest in the CSI effect merely to the efforts of those who have sought to perpetuate the story. The media does not accept every narrative proposed to it, and not every narrative shows the sort of penetration of public consciousness that the CSI effect has clearly achieved. As with the litigation explosion, the successful dissemination of the CSI effect must also be attributed to the resonance of CSI effect narratives with large themes and values in contemporary culture.

Kelner has usefully divided the litigation explosion literature into two broad categories. (151) Some literature seeks to debunk claims of a litigation explosion. Other literature seeks not only to debunk, but also to explain--to explore the symbolic meaning of litigation explosion claims. (152) In addition to the work of debunking, legal scholars searched for explanations as to "why so large a gap exists between the widespread perception that the American litigation system is wildly out of control and the picture that emerges from an examination of the available evidence." (153) This Article is modeled on the latter category. We seek not merely to debunk overblown media claims about television's influence on jury verdicts, but also to attempt an explanation of why these narratives hold such great appeal for the media and what that tells us about the public's perceptions of the law and even law's own self-perception.

In the case of the litigation explosion, some socio-legal scholars proposed that the litigation explosion narrative evinced a severe distrust of the civil jury system. (154) Others have noted that these arguments are situated within a broader context in which the litigation explosion was viewed as a sign of the erosion of American rugged self-reliance and individualism. (155) Thus, the litigation explosion is situated within a profound ambivalence about, or even hostility to, the use of law--a "turn against law." (156) This notion, in turn, implies an imagined end to law--an America in which disputes are resolved without law, or at least without "too much law." (157)

Media discussions of the CSI effect display similar distrust of juries, though in criminal, not civil, law. But the CSI effect cannot be explained by reference to the American values of individualism, responsibility and self-reliance. What cultural values, then, account for the remarkable resonance of the CSI effect? To us, the answer seems clear: the rising authority and prestige of science in modern society. Science is popularly associated with such positive values as truth, certainty, goodness, enlightenment, progress, and so on. Law's relationship to science has always been somewhat uneasy.(158) While law has often held high hopes that science would prove effective at resolving disputes without ambiguity, this very potential to be truth-producer is a cause for understandable anxiety on the part of the law. (159) As we have suggested elsewhere, (160) the CSI effect would seem to embody the law's anxiety about the threat to its legitimacy as a truth-producing institution posed by a rival truth-producing institute called "science." The discourse among legal actors about the supposed CSI effect is rife with lamentations of the law's purported inability to provide proof with the strength that jurors supposedly desire. Whereas the litigation explosion may have resonated with a societal anxiety about relying on law too heavily, the CSI effect would seem to resonate with anxieties about using law too little, increasingly abrogating its truth-producing function to science. Whereas the litigation explosion may have articulated fears of hyperlexis, the CSI effect would seem to give voice to fears of what we might call "hyperscientia"--too much science.

(1.) Blakely v. Washington, 542 U.S. 296, 306 (2004).

(2.) See, e.g., Shari Seidman Diamond & Mary R. Rose, Real Juries, 1 ANN. REV. L. & SOC. SCI. 255 (2005).

(3.) See, e.g., Geoffrey P. Kramer et al., Pretrial Publicity, Judicial Remedies, and Jury Bias, 14 LAW & HUM. BEHAV. 409, 435 (1990); Amy L. Otto et al., The Biasing Impact of Pretrial Publicity on Juror Judgments, 18 LAW & HUM. BEHAV. 453, 454 (1994).

(4.) See, e.g., Edith Greene, Media Effects on Jurors, 14 LAW & HUM. BEHAV. 439 (1990); Valerie P. Hans & Juliet L. Dee, Media Coverage of Law: Its Impact on Juries and the Public, 35 AM. BEHAV. SCIENTIST 136 (1991).

(5.) Simon A. Cole & Rachel Dioso-Villa, CSI and Its Effects: Media, Juries, and the Burden of Proof 41 NEW ENG. L. REV. 435, 435 (2007).

(6.) Michael Alien, Introduction: This Much I Know ... to READING CSI: CRIME TV UNDER THE MICROSCOPE 3, 4 (Michael Allen ed., 2007) [hereinafter READING CSI].

(7.) Silke Panse, 'The Bullets Confirm the Story Told by the Potato': Materials Without Motives in CSI: Crime Scene Investigation, in READING CSI, supra note 6, at 153, 153.

(8.) Allen, supra note 6, at 5; see also Deborah Jermyn, Body Matters: Realism, Spectacle and the Corpse in CSI, in READING CSI, supra note 6, at 79, 79 ("That a programme built around the gruesome clues, secrets and promises embedded within, and articulated across, the image of the corpse could become the most successful television series in the world would have been unimaginable until relatively recently.").

(9.) Paige Albiniak, Sizzling 'CSI' Reruns: Off-net Procedurals Stay Strong, BROADCASTING & CABLE, Nov. 3, 2008, at 9.

(10.) CSI (franchise), WIKIPEDIA, http://en.wikipedia.org/wiki/CSI_franchise (last visited Apr. 1, 2009).

(11.) We drew on research by Jennifer L. Christian in compiling this list.

(12.) CSI: Crime Scene Investigation, WIKIPEDIA, http://en.wikipedia.org/wiki/ CSI:_Crime_Scene_Investigation#American_Ratings (last visited Jan. 8, 2009); CSI: Miami, WIKIPEDIA, http://en.wikipedia.org/wiki/CSI:_Miami#U.S._television_ratings (last visited Jan. 8, 2009); CSI: New York, WIKIPEDIA, http://en.wikipedia.org/wiki/CSI:_NY# Nielsen_Ratings (last visited Jan. 8, 2009).

(13.) Jeffrey Kluger, How Science Solves Crimes, TIME, Oct. 21, 2002, at 36, 45.

(14.) Id.

(15.) Id.

(16.) Kit R. Roane, The CSI Effect, U.S. NEWS & WORLD REP., Apr. 25, 2005, at 48.

(17.) Max M. Houck, CSI: Reality, SOL AM., July 2006, at 84; Stefan Lovgren, "'CSI' Effect" is Mixed Blessing for Real Crime Labs, NAT'L GEOGRAPHIC NEWS, Sept. 23, 2004, http://news.nationalgeographic.com/news/pf/80520796.html.

(18.) KATHERINE RAMSLAND, THE C.S.I. EFFECT (2006).

(19.) For a discussion of media loops, see PETER K. MANNING, POLICING CONTINGENCIES 76-77 (2003); see also Susan Bandes, Fear Factor. The Role of Media in Covering and Shaping the Death Penalty, 1 OHIO ST. J. CRIM. L. 585, 585 (2004) (discussing a "complex feedback loop" between law and the media).

(20.) For an argument about the indistinguishability of "fiction" and "reality" in this discourse, see Michael Mopas, Examining the 'CSI Effect' Through an ANT Lens, 3 CRIME MEDIA CULTURE 110 (2007).

(21.) Linda Deutsch, 'CSI' and 'Law & Order' Lead Jurors to Great Expectations, Associated Press, Jan. 14, 2006.

(22.) Glenn E. Rice, TV is Making Jurors Suspicious About Evidence: CSI: Courtroom Stalemates Increase, KAN. CITY STAR, Aug. 10, 2005, at A1.

(23.) Roane, supra note 16, at cover.

(24.) Kate Coscarelli, The 'CSI' Effect: TV's False Reality Fools Jurors. Newhouse News Service, Apr. 21, 2005.

(25.) CBS4 Denver: CSI Has 'Major Effect' on Real Life Juries (CBS television broadcast May 5, 2005).

(26.) Clarence Walker, CSI (T. V. Crime Dramas) Affects the American Criminal Justice System, AMERICANMAFIA.COM, June 2005, http://www.americanmafia.com/Feature_ Articles_301.html.

(27.) Amy Lennard Goehner et al., Ripple Effect: Where CSI Meets Real Law and Order, TIME, Nov. 8, 2004, at 69 (quoting jury consultant Robert Hirschhorn) (internal quotation marks omitted).

(28.) Allison Klein, Art Trips Up Life: TV Crime Shows Influence Jurors, BALT. SUN, July 25, 2004, at 1A (quoting forensic scientist Thomas Mauriello) (internal quotation marks omitted).

(29.) Id. (quoting Baltimore Deputy State Attorney Haven H. Kodeck) (internal quotation marks omitted).

(30.) DVD: CSI Effect Fact or Fiction (Fed. Bureau of Investigation Training Network #112 2000) (on file with SUNY University at Buffalo library).

(31.) MARICOPA COUNTY ATT'Y'S OFFICE, CSI: MARICOPA COUNTY: THE CSI EFFECT AND ITS REAL-LIFE IMPACT ON JUSTICE 1, 10 (2005) [hereinafter CSI: MARICOPA COUNTY].

(32.) Boatswain v. State, No. 408,2004, 2005 WL 1000565, at *2 (Del. Apr. 27, 2005) (internal citation omitted); cf Robin Franzen, TV's 'CSI' Crime Drama Makes It Look Too Easy, OREGONIAN, Dec. 10, 2002, at A01 (quoting Clatsop County District Attorney Josh Marquis expressing prosecutors' concern about "being held to an artificial 'CSI' standard").

(33.) See Kimberlianne Podlas, "The CSI Effect": Exposing the Media Myth, 16 FORDHAM INTELL. PROP. MEDIA & ENT. L.J. 429 (2006); Donald E. Shelton et al., A Study of Juror Expectations and Demands Concerning Scientific Evidence: Does the "CSI Effect" Exist?, 9 VAND. J. ENT. & TECH. L. 331 (2006); Tom R. Tyler, Viewing CSI and the Threshold of Guilt: Managing Truth and Justice in Reality and Fiction, 115 YALE L.J. 1050 (2006).

(34.) See Podlas, supra note 33; Tyler, supra note 33.

(35.) See, e.g., Kiara Okita, The CSI Effect: Examining CSFs Effects upon Public Perceptions of Forensic Science (Fall 2007) (unpublished Master's thesis, University of Alberta) (on file with author).

(36.) See Cole & Dioso-Villa, supra note 5; see also Charles Loeffler, "CSI" and the Criminal Justice System: Jury Trials, NEW REPUBLIC ONLINE, June 7, 2006 (on file with authors).

(37.) See, e.g., Podlas, supra note 33; Shelton et al., supra note 33; Tyler, supra note 33; Okita, supra note 35.

(38.) See, e.g., READING CSI, supra note 6; Gray Cavender & Sarah K. Deutsch, CSI and Moral Authority: The Police and Science, 3 CRIME MEDIA CULTURE 67 (2007); Martha Gever, The Spectacle of Crime, Digitized: CSI: Crime Scene Investigation and Social Anatomy, 8 EUROPEAN J. CULTURAL STUD. 445 (2005); Mopas, supra note 20; Thomas W. Nolan, Depiction of the "CSI Effect" in Popular Culture: Portrait in Domination and Effective Affectation, 41 NEW ENG. L. REV. 575 (2007).

(39.) See Marc Galanter, Reading the Landscape of Disputes: What We Know and Don't Know (and Think We Know) About Our Allegedly Contentious and Litigious Society, 31 UCLA L. REV. 4, 5 (1983) [hereinafter Galanter, Reading the Landscape].

(40.) Id. at 6.

(41.) Marc Galanter, The Day After the Litigation Explosion, 46 MD. L. REV. 3, 37 (1986) [hereinafter Galanter, Day After].

(42.) See Kimberlianne Podlas, The Monster in the Television: The Media's Contribution to the Consumer Litigation Boogeyman, 34 GOLDEN GATE U. L. REV. 239, 241 (2004).

(43.) See Galanter, Day After, supra note 41; Marc Galanter, The Hundred-Year Decline of Trials and the Thirty Years War, 57 STAN. L. REV. 1255 (2005) [hereinafter Galanter, Decline of Trials]; Galanter, Reading the Landscape, supra note 39; Deborah L. Rhode, Frivolous Litigation and Civil Justice Reform: Miscasting the Problem, Recasting the Solution, 54 DUKE L.J. 447 (2004); Randy M. Mastro, The Myth of the Litigation Explosion, 60 FORDHAM L. REV. 199, 200 n.9 (1991) (reviewing WALTER K. OLSON, THE LITIGATION EXPLOSION (1991)). Since there is an extensive literature on the litigation explosion, we will not rehash those arguments here.

(44.) See The Insurance Industry: Do Its Ads Undermine Jury Impartiality?, 75 A.B.A. J. 46 (1989).

(45.) See, e.g., WILLIAM HALTOM & MICHAEL MCCANN, DISTORTING THE LAW: POLITICS, MEDIA, AND THE LITIGATION CRISIS 61 (2004);

(46.) While both the litigation explosion and the CSI effect are portrayed as litigation crises, the CSI effect is better characterized as a "litigation crisis" than as a "litigation explosion." The CSI effect refers to a supposed problem in litigation but not to any supposed increase (i.e., "explosion") in litigation.

(47.) Cole & Dioso-Villa, supra note 5. The discussion in this Part draws heavily from that work.

(48.) Negative evidence refers to the notion of presenting testimonial evidence explaining the absence of physical evidence when the jury might construe that absence as significant. For example, the prosecution might call a forensic technician to testify that the crime scene was dusted for fingerprints but none were found to preclude the defense from insinuating that the police were too lazy or too focused on the defendant to search for fingerprints.

(49.) See Cynthia A. Hoffner et al., TV Characters at Work: Television's Role in the Occupational Aspirations of Economically Disadvantaged Youths, 33 J. CAREER DEV. 3, 15 (2006).

(50.) See, e.g., Michael M. O'Connor, The Role of the Television Drama ER in Medical Student Life: Entertainment or Socialization?, 280 J. AM. MED. ASS'N 854, 854 (1998).

(51.) See, e.g., John Randolph Fuller, So You Want to Be a Serial-Murderer Profiler ..., CHRON. HIGHER EDUC., Dec. 7, 2001, at B5.

(52.) For an interesting discussion of convicts' perceptions of the strength of DNA evidence, see generally Barbara Prainsack & Martin Kitzberger, DNA Behind Bars: Other Ways of Knowing Forensic DNA Technologies, 39 Soc. STUD. Set. 51, 64 (2009).

(53.) See infra note 129 and accompanying text.

(54.) It should be noted that many forensic scientists feel that aspiring forensic scientists are not well served by forensic degree programs and would be better served by mainstream science programs. See, e.g., KEITH INMAN & NORAH RUDIN, PRINCIPLES AND PRACTICE OF CRIMINALISTICS: THE PROFESSION OF FORENSIC SCIENCE 302 (2001).

(55.) Cole & Dioso-Villa, supra note 5, at 453.

(56.) See. e.g., Enric Volante & Kim Smith, 'CSI Effect' Impacts Justice in Tucson, ARIZ. DAILY STAR, May 8, 2005, at A1.

(57.) CSI: MARICOPA COUNTY, supra note 31, at 2, 5.

(58.) Andrew P. Thomas, The CSI Effect: Fact or Fiction, 115 YALE L.J. POCKET PART 70, 71-72 (2006), http://www.thepocketpart.org/2006/02/thomas.html; see also CSI: MARICOPA COUNTY, supra note 31, at 5.

(59.) Thomas, supra note 58, at 72; see also CSI: MARICOPA COUNTY, supra note 31.

(60.) Shelton et al., supra note 33, at 362.

(61.) Id. at 362-65.

(62.) Id.

(63.) See Nolan, supra note 38, at 580.

(64.) See Cynthia Hoffner & Martha Buchanan, Young Adults' Wishful Identification with Television Characters: The Role of Perceived Similarity and Character Attributes, 7 MEDIA PSVCHOL. 325, 327-28 (2005) (arguing that "[t]elevision role models ... influence young people's occupational aspirations").

(65.) It might be argued that, if prosecution witnesses were already scientists to begin with, the advent of CSI should not alter the jury's perception of them as scientists. This issue is complicated by the fact that some forensic disciplines--such as forensic document analysis, latent print analysis, and firearms and toolmark analysis--are practiced by individuals who do not possess formal academic scientific credentials. (Whether they might yet be considered "scientists" is a fraught question.) On CSI, however, the same scientifically credentialed individuals who run DNA analyses and other high-tech assays also perform latent print analysis and tool mark analysis. It is, therefore, possible that CSI might increase jurors' association of practitioners of the humbler forensic disciplines with the prestigious profession of science.

(66.) The Roper Center for Public Opinion Research, University of Connecticut, http://www.ropercenter.uconn.edu/data_access/ipoll/ipoll.html (last visited Apr. 4, 2009) (also on file with authors). Responses are compiled from telephone surveys of national samples of American adults conducted by Harris Interactive. Sample size varied from year to year but was always greater than 1000.

(67.) See, e.g., Cole & Dioso-Villa, supra note 5; Tyler, supra note 33.

(68.) See, e.g., Roane, supra note 16; Lovgren, supra note 17.

(69.) See, e.g., Podlas, supra note 33; Podlas, supra note 42; Tyler, supra note 33.

(70.) See, e.g., HALTOM & MCCANN, supra note 45; Robert M. Hayden, The Cultural Logic of a Political Crisis: Common Sense, Hegemony and the Great American Liability Insurance Famine of 1986, 11 STUD, L. POL. & SOC'Y 95, 104-08 (1991).

(71.) See, e.g., Rhode, supra note 43.

(72.) See, e.g., HALTOM & MCCANN, supra note 45, at 183-226; Shari Seidman Diamond, Truth, Justice, and the Jury, 26 HARV. J.L. & PUB. POL'Y 143, 145-47 (2003); Hayden, supra note 70, at 106-08.

(73.) Julie Keller, D.A.: Blake Jury 'Incredibly Stupid', E!ONLINE, Mar. 24, 2005, http://www3.eonline.com/uberblog/b49519_da_blake_jury_incredibly stupid.html.

(74.) Houck, supra note 17, at 86.

(75.) Cole & Dioso-Villa, supra note 5, at 455.

(76.) See Andy Kravetz, 'CSI' Making Impact in Courts. Area Prosecutors Battling TV Show's Reality with Real Life PEORIA J. STAR, Sept. 4, 2005, at A1.

(77.) See Art Aisner, 'CSI Effect' Put Under the Scope, ANN ARBOR NEWS, May 18, 2007, at A1.

(78.) See e.g., HALTOM & MCCANN, supra note 45, at 183-226; Diamond, supra note 72, at 146.

(79.) See, e.g., Klein, supra note 28, at 1A.

(80.) See Keller, supra note 73.

(81.) See John Lande, Failing Faith in Litigation? A Survey of Business Lawyers' and Executives' Opinions, 3 HARV. NEGOT. L. REV. 1 (1998); see also Galanter, Decline of Trials, supra note 43, at 1267 n.36; Podlas, supra note 42, at 244.

(82.) See Michael J. Watkins, Forensics in the Media: Have Attorneys Reacted to the Growing Popularity of Forensic Crime Dramas? 59 (Aug. 3, 2004) (unpublished M.A. thesis, Florida State University) (on file with authors).

(83.) See CSI: MARICOPA COUNTY, supra note 31, at 5; see also Thomas, supra note 58, at 70.

(84.) Thomas, supra note 58, at 70.

(85.) See Monica L.P. Robbers, Blinded by Science: The Social Construction of Reality in Forensic Television Shows and its Effect on Criminal Jury Trials, 19 CRIM. JUST. POL'Y REV. 84, 91-98 (2008).

(86.) See id. at 91-92.

(87.) See Dennis J. Stevens, Forensic Science. Wrongful Convictions, and American Prosecutor Discretion, 47 How. J. 31, 46 (2008).

(88.) See Cole & Dioso-Villa, supra note 5, at 458; see also Tyler, supra note 33, at 1078 (speculating that the CSI effect describes prosecutors' attempts to understand jury behavior).

(89.) See Podlas, supra note 33, at 455-61.

(90.) See Shelton et al., supra note 33, at 332-39.

(91.) See Okita, supra note 35, at 47.

(92.) Id. at 103.

(93.) See Steven M. Smith et al., Exploring the CSI Effect: Is It Real? If So, What Is It?, Address at the North American Correctional and Criminal Justice Psychology Conference, in PROCEEDINGS OF THE NORTH AMERICAN CORRECTIONAL & CRIMINAL JUSTICE PSYCHOLOGY CONFERENCE, 2008, at 81, 83.

(94.) See Kimberlianne Podlas, The 'CSI Effect' and Other Forensic Fictions, 27 LOY. L.A. ENT. L. REV. 87, 112-21 (2007).

(95.) See N.J. Schweitzer & Michael J. Saks, The CSI Effect: Popular Fiction About Forensic Science Affects the Public's Expectations About Real Forensic Science, 47 JURIMETRICS J. 357, 360-63 (2007).

(96.) See Margaret A. Berger, The Impact of DNA Exonerations on the Criminal Justice System, 34 J.L. MED. & ETHICS 320, 322 (2006).

(97.) See Gwen Jenkins et al., The CSI Effect: Mock Jurors' Sensitivity to the Reliability of Forensic Evidence, Address at Annual Meeting of the American Psychology-Law Society, Jacksonville, Fla. (Mar. 5-8, 2008) (on file with authors).

(98.) HALTOM & MCCANN, supra note 45, at 73; Galanter, Day After, supra note 41; Galanter, Decline of Trials, supra note 43; Galanter, Reading the Landscape, supra note 39; Michael Saks, If There Be a Crisis, How Shall We Know It?, 46 MD. L. REV., 63 (1986).

(99.) Galanter, Reading the Landscape, supra note 39, at 71.

(100.) Cole & Dioso-Villa, supra note 5, at 462.

(101.) See, e.g., Podlas, supra note 33, at 461; see also Tyler, supra note 33, at 1084.

(102.) Loeffler, supra note 36. The states were New York, Texas, Illinois, and California.

(103.) Benoit Dupont, The CSI Effect: Myths and Reality, Address at 29th Annual Conference of the Canadian Identification Association (Nov. 24, 2008).

(104.) These eleven states included California, Florida, Hawaii, Illinois, Kentucky, Missouri, New York, North Carolina, South Dakota, Texas, and Vermont.

(105.) The sources of the data are: BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, SOURCEBOOK OF CRIMINAL JUSTICE STATISTICS ONLINE, available at http://www.albany.edu/sourcebook; CIRCUIT COURTS OF ILL., CASELOAD AND STATISTICAL RECORDS, available at http://www.state.il.us/COURT/SupremeCourt/AnnualReport/2006/ Stat/caseload.pdf; FLA. OFFICE OF THE STATE COURTS ADM'R, FLORIDA'S TRIAL COURTS STATISTICAL REFERENCE GUIDE, available at http://www.flcourts.org/gen_public/stats/ reference_guide.shtml; JUDICIAL COUNCIL OF CAL., ANNUAL DATA REFERENCE FOR 1994-1995, available at http://www.courtinfo.ea.gov/reference/3_stats.htm; JUDICIAL COUNCIL OF CAL., COURT STATISTICS REPORTS, STATEWIDE CASELOAD TRENDS, available at http://www.courtinfo.ea.gov/reference/3_stats.htm; N.C. JUDICIAL BRANCH, ADMIN. OFFICE OF THE COURTS, CASELOAD INVENTORY BY COUNTY, SUPERIOR COURT CRIMINAL, DISPOSITION ACTIVITY, STATEWIDE SUMMARY (courtesy of Mr. Patrick Tamer, Statistical Programmer Analyst, Research and Planning Div., N.C. Admin. Office of the Courts); OFFICE OF COURT ADM'R, ANNUAL STATISTICAL REPORT FOR THE TEXAS JUDICIARY, available at http://www.courts.state.tx.us/pubs/annual-reports.asp; STATE OF HAW., JUDICIARY, OFFICE OF THE ADMIN. DIR., STATEWIDE CIRCUIT COURT CRIMINAL (on file with the Stanford Law Review); STATE OF N.Y., REPORT OF THE CHIEF ADMINISTRATOR OF THE COURTS, available at http://www.courts.state.ny.us/reports/annual; STATE OF VT. SUPREME COURT, JUDICIAL STATISTICS (2000) (on file with the Stanford Law Review); STATE OF VERMONT SUPREME COURT, JUDICIAL STATISTICS (1999) (on file with the Stanford Law Review); STATE OF VERMONT SUPREME COURT, JUDICIAL STATISTICS (1998) (on file with the Stanford Law Review); VT. JUDICIARY, ANNUAL STATISTICS, DISTRICT COURT, available at http://www.vermontjudiciary.org/Statistics/default.aspx.

(106.) The California data does not allow the removal of misdemeanors because, although convictions are broken out into felonies and misdemeanors, acquittals are aggregated. We do not feel that the inclusion of a small number of California misdemeanors in the felony data is likely to substantially distort our findings. To give some idea of what we mean by a "small number" of misdemeanors: in 2007, misdemeanors accounted for 4.5% of all convictions recorded by the California Superior Courts. See JUDICIAL COUNCIL OF CALIFORNIA, COURT STATISTICS REPORTS, STATEWIDE CASELOAD TRENDS (2008), available at http://www.courtinfo.ca.gov/reference/documents/csr2008.pdf.

(107.) Note: all tests were not significant at p>.05.

(108.) Strictly speaking, our sample was not random. It was a convenience sample dictated by which states compile acquittal rate data. Nonetheless, we think it is still appropriate to treat the sample as random because we did not exercise any choice in selecting which states would supply the sample data.

(109.) Had we wanted to use data from earlier than 1997, we would have had to exclude the State of New York, which accounts for a large amount of data.

(110.) See HALTOM & MCCANN, supra note 45.

(111.) Id. at 67.

(112.) Id. at 148; see also id. at 100, 109.

(113.) Id. at 148.

(114.) See id. at 165-66; Daniel S. Bailis & Robert J. MacCoun, Estimating Liability Risks with the Media as Your Guide: A Content Analysis of Media Coverage of Tort Litigation, 20 LAW & HUM. BEHAV. 419, 425-26 (1996); Steven Garber & Anthony G. Bower, Newspaper Coverage of Automotive Product Liability Verdicts, 33 LAW & SOC'Y REV. 93 (1999).

(115.) See Stephen S. Meinhold & David W. Neubauer, Exploring Attitudes About the Litigation Explosion, 22 JUST. SYS. J. 105 (2001); cf. Edith Greene et al., Jurors' Attitudes About Civil Litigation and the Size of Damage Awards, 40 AM. U. L. REV. 805, 814, 817 (1991) (finding a high degree of belief in a "litigation explosion" but a lack of belief in an insurance "crisis").

(116.) See, e.g., Galanter, Decline of Trials, supra note 43, at 1267-68.

(117.) Elizabeth Loftus, Insurance Advertising and Jury Awards, 65 A.B.A. J. 68, 70 (1979).

(118.) Valerie P. Hans & William S. Lofquist, Jurors' Judgments of Business Liability in Tort Cases: Implications for the Litigation Explosion Debate, 26 LAW & SOC'Y REV. 85, 93-97 (1992).

(119.) Id. at97.

(120.) Id.

(121.) Edith Greene et al., supra note 115, at 816-17.

(122.) Id. at 808-09 (citing S. Daniels, Civil Juries, Civil Jury Verdict Reporters, and the Going Rate (paper presented at the annual meeting of the Law and Society Association in Chicago, 1986)); see also H. Ross, SETTLED OUT OF COURT: THE SOCIAL PROGRESS OF INSURANCE CLAIMS 112-16 (1970).

(123.) Jennifer K. Robbennolt & Christina A. Studebaker, News Media Reporting on Civil Litigation and Its Influence on Civil Justice Decision Making, 27 LAW & HUM. BEHAV. 5, 23 (2003).

(124.) Elizabeth Harvey & Linda Derksen, The CSI Effect: Science Fiction or Social Fact?, in DECONSTRUCTING THE CSI EFFECT: PRODUCING NARRATIVES OF JUSTICE AND SCIENCE, PRODUCING TELEVISION DRAMA 18, 25-26 (Valerie Marie Johnson & Michelle Byers eds., 2009).

(125.) We eliminated academic articles because we were interested in studying the characterization of the CSI effect in popular media, not scholarly discourse.

(126.) HALTOM & MCCANN, supra note 45, at 203.

(127.) See generally Lisbeth Clausen, Localizing the Global: "Domestication" Processes in International News Production, 26 MEDIA CULTURE & SOC'Y 25 (2004) (describing a similar process of "news domestication").

(128.) Cole & Dioso-Villa, supra note 5, at 444.

(129.) One article mentioned what we call the "Hollywood effect." This held that CSI had induced forensic scientists to forego their careers in law enforcement in order to work as consultants in the entertainment industry. Richard Willing, Medium Sends Message to U.S. Court System: 'CSI' Effect Has Juries Wanting More Evidence, CHI. SUN-TIMES, Aug. 15, 2004, at 1. Another article mentioned what we call the "Joseph Smith effect." The claim here was that CSI had promulgated the notion that "DNA proves pretty much everything" and thus undermined faith in the teachings of the Book of Mormon by contradicting claims that Native Americans were descendants of the lost tribes of Israel. Carrie A. Moore, DNA Claims Rebutted on Book of Mormon, DESERET MORNING NEWS, Oct. 23, 2007. For a fascinating account of this controversy, see generally Gary Edmond & Simon Southerton, Almost Apostasy: Human DNA Genealogy and the Latter-day Saints (unpublished manuscript on file with author).

(130.) See, e.g., The Jury Is In: TV Doesn't Sway Justice, GRAND RAPIDS PRESS, May 21, 2007, at A3 [hereinafter The Jury Is In].

(131.) See Harvey & Derksen, supra note 124, at 25-26.

(132.) See id. at 28.

(133.) HALTOM & MCCANN, supra note 45, at 148 (internal quotation marks omitted).

(134.) See Podlas, supra note 33, at 452; Tyler, supra note 33, at 1063.

(135.) See, e.g., Walt Belcher, College Teacher Finds Some Evidence Debunking Alleged 'CSI Effect' on Verdicts, TAMPA TRIB., Feb. 16, 2006, at 4; The Jury Is In, supra note 130.

(136.) Susan Clairmont, Public Is a Victim of 'The CSI Effect', HAMILTON SPECTATOR, Sept. 29, 2005, at A01.

(137.) Galanter, Reading the Landscapes, supra note 39, at 65.

(138.) Clairmont, supra note 136.

(139.) Carlene Hempel, TV's Whodunit Effect. Police Dramas Are Having an Unexpected Impact in the Real World: The Public Thinks Every Crime Can be Solved, and Solved Now--Just Like on Television, BOSTON GLOBE, Feb. 9, 2003, (Magazine), at 13.

(140.) Roane, supra note 16, at 48.

(141.) Sonya Neufeld, Sorting Fact from Fiction, HOBART MERCURY, June 25, 2005, at B07.

(142.) See Robbennolt & Studebaker, supra note 123, at 22.

(143.) See Rhode, supra note 43, at 472.

(144.) Hans & Lofquist, supra note 118, at 111-12.

(145.) See id. at 112.

(146.) See, e.g., Robbennolt & Studebaker, supra note 123, at 15.

(147.) It should also be noted that judges who believe these things might also be unconsciously more sympathetic to the prosecution. Although some judges have evinced skepticism, e.g., Shelton et al., supra note 33, our data set contains numerous statements by judges attesting to the existence of the strong prosecutor's CSI effect. See, e.g., Maggi Newhouse, Real-Life Investigators Indict 'CSI' for Perjury, PITTSBURGH TRIB.-REV., Apr. 13, 2005; Dana Sullivan, Get Set for Jurors with CSI IQs, N.J. LAWYER, June 13, 2005, at 1.

(148.) Okita, supra note 35, at 106.

(149.) See CSI: MARICOPA COUNTY, supra note 31, at 11.

(150.) See Harvey & Derksen, supra note 124, at 18.

(151.) See Joshua D. Kelner, The Anatomy of an Image: Unpacking the Case for Tort Reform, 31 U. DAYTON L. REV. 243, 246 (2006) (noting that "[f]irst, a litany of scholars has adeptly refuted the arguments propounded by the proponents of tort reform" and "[s]econd, though with less frequency, scholars have devoted attention to understanding how the critique of the legal system functions, on both logical and symbolic levels").

(152.) The work of Professor Galanter is a prime example. See, e.g., Galanter, Decline of Trials, supra note 43; Galanter, Reading the Landscape, supra note 39.

(153.) Saks, supra note 98, at 74.

(154.) See Diamond, supra note 72, at 144; Kelner, supra note 151, at 266.

(155.) See Galanter, Reading the Landscape, supra note 39, at 66; see also Hayden, supra note 70, at 110.

(156.) Galanter, Decline of Trials, supra note 43, at 1269.

(157.) Galanter, Reading the Landscape, supra note 39, at 6.

(158.) See TAL GOLAN, LAWS OF MEN AND LAWS OF NATURE: THE HISTORY OF SCIENTIFIC EXPERT TESTIMONY IN ENGLAND AND AMERICA 263 (2004) (recounting law's "two centuries of growing ... dependence on, and frustration with, science").

(159.) This notion is a theme in, for example, KEN ALDER, THE LIE DETECTORS: THE HISTORY OF AN AMERICAN OBSESSION (2007); JAY D. ARONSON, GENETIC WITNESS: SCIENCE, LAW, AND CONTROVERSY IN THE MAKING OF DNA PROFILING (2007); MICHAEL LYNCH ET AL., TRUTH MACHINE: THE CONTENTIOUS HISTORY OF DNA FINGERPRINTING (2008); Jennifer L. Mnookin, Idealizing Science and Demonizing Experts: An Intellectual History of Expert Evidence, 52 VILL. L. REV. 763 (2007).

(160.) See Cole & Dioso-Villa, supra note 5, at 468-69.

Simon A. Cole, Associate Professor of Criminology, Law & Society, University of California, Irvine; Ph.D. (science & technology studies), Cornell University; A.B., Princeton University.

Rachel Dioso-Villa, Doctoral candidate, Criminology, Law & Society, University of California, Irvine; M.A. (criminology), University of Toronto; B.Sc., University of Toronto. For research assistance, we are grateful to Kasey Perry and Nicole Bucur. For assistance obtaining acquittal rate data, we are grateful to Thomas Keller and Sharon Iwai of the State of Hawai'i Judiciary, Patrick Tamer of the North Carolina Administrative Office of the Courts, and Christine A. Loso of the Vermont Court Administrators Office. We are also grateful for the assistance of Charles Loeffler, Jennifer Christian, Richard McCleary, John R. Hipp, and Elizabeth Loftus and for the thorough editing and convening of a symposium by the Stanford Law Review staff. This project was funded in part by the National Science Foundation (Award #SES-0347305). Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

Table 1. Annual Rank of CSI Franchise Programs Among U.S. Television
Programs (12)

Year   CSI   CSI: Miami   CSI: New York

2000    10
2001     2
2002     1           10
2003     2            8
2004     2            5              21
2005     3            6              22
2006     5           10              28
2007     5            8              23
2008     4           14              16

Table 2. Media Mentions of CSI Effect

Year   Mentions

2002          2
2003          2
2004         12
2005         56
2006         78
2007         65
2008         43

Table 3. The Many Effects of CSI: Typology of CSI Effects Found in
Media Accounts

Effect name             Effect on       Description

Strong prosecutor's     Jurors          Acquit in cases in which
effect                                  they would have convicted
                                        had CSI never existed

Weak prosecutor's       Prosecutors     Compensate for absence/
effect                                  weakness of forensic
                                        evidence

Defendant's effect      Jurors          Afford greater credibility
                                        to forensic expert
                                        witnesses

Producer's effect       Jurors          Know more science

Educator's effect       Students        Attraction to careers in
                                        forensic science

Police chief's          Criminals       Adopt countermeasures to
effect                                  prevent detection through
                                        forensic evidence

Tech effect             Jurors          Hold higher expectation for
                                        forensic evidence because
                                        of actual developments in
                                        forensic technology

Victim's effect (53)    Crime victims   Expect forensic testing for
                                        all crimes

Table 4. Percentage of Respondents Responding "Very Great Prestige"
to Questions About the Prestige of Selected Professions (66)

       Scientist   Doctor   Teacher   Police    Lawyer   Journalist
                                      officer

1992          57       50        41        34       25           15
1997          51       52        49        36       19           15
1998          55       61        53        41       23           15
2000          56       61        53        38       21           16
2001          53       61        54        37       18           18
2002          51       50        47        40       15           19
2003          57       52        49        42       17           15
2004          52       52        48        40       17           14
2005          56       54        47        40       18           14
2006          54       58        52        43       21           16
2008          56       53        52        46       24           18

Table 5. Acquittal Rates for Nine Jurisdictions in All Years Available,
Starting in 1986

                             North                               New
          Mean    Vermont   Florida   Carolina   Illinois        York
         Number
Fiscal      of         60      4131       2025            1009   2478
Year     Trials

1986                            37%
1987                            35%
1988                            37%
1989                            34%
1990                            34%
1991                            37%
1992                            35%        33%
1993                            34%        30%
1994                            33%        31%
1995                            33%        31%
1996                            38%        31%             29%

1997                            31%        27%             27%    26%
1998                  25%       33%        31%             26%    28%
1999                  31%       34%        30%              26    29%
2000                  30%       37%        30%             28%    29%
2001                  26%       37%        31%             33%    29%
2002                  44%       36%        30%             30%    29%
2003                  24%       35%        30%             31%    30%
2004                  33%       34%        30%             32%    29%
2005                  38%       33%        29%             35%    30%
2006                  37%       35%        29%             36%    30%
2007                  32%       33%        36%             32%
2008                  46%                  30%

         Hawaii   Texas   California   Federal   Aggregate

Fiscal      260    3180         5594      4141       22878
Year

1986                                       19%         26%
1987                                       19%         26%
1988                                       20%         27%
1989                                       19%         25%
1990                                       16%         23%
1991        21%                            17%         24%
1992        27%                            17%         25%
1993        31%                 l6%        16%         26%
1994        28%                  15%       15%         24%
1995        34%                 l4%        16%         23%
1996        35%     19%                    13%         26%

1997        31%     19%          13%       13%         22%
1998        31%     18%          14%       14%         24%
1999        29%     20%          16%       15%         24%
2000        24%     20%          17%       15%         25%
2001        24%     20%          17%       16%         26%
2002        25%     21%          19%       15%         26%
2003         34     21%          14%       15%         24%
2004        26%     19%          16%       15%         25%
2005        33%     21%          16%       11%         23%
2006        24%     20%          16%       11%         24%
2007        34%     21%          16%       11%         23%
2008        32%                                        25%

The horizontal line marks the advent of CSI. The dashed
box indicates data used in Analysis 2. The mean number
of trials per year for each jurisdiction is given as a
rough measure of the sample size for each jurisdiction.

Table 6. Linear Regression Summary of the Relationship Between
Acquittal Rates Before and After the Airing of CSI in 2001,
2002, and 2003 (n=132) (107)

                                                       95% CI

Pre/Post Comparisons   Coefficient   Standard     Lower      Upper
                                      Error       Bound      Bound

Pre vs. Post 2001       0.001975     0.005727   -0.009361   0.013311
Pre vs. Post 2002       0.005035     0.005679   -0.006206   0.016276
Pre vs. Post 2003       0.002294     0.005802   -0.009192   0.013780

Table 7. Aggregate Number of Trials and Acquittals from 1997-2006

Year   Total Number of   Number of    Acquittal
           Trials        Acquittals     Rate

1997       24,343          5,405        21.9%
1998       22,553          5,316        23.5%
1999       22,133          5,311        24.1%
2000       21,291          5,399        25.0%
2001       19,768          5,027        25.5%
2002       19,179          4,957        25.9%
2003       20,219          4,887        24.2%
2004       19,235          4,747        24.7%
2005       18,807          4,345        23.2%
2006       19,746          4,728        24.0%

Table 8. Comparisons Between Aggregate Acquittal Rates Before and
After the Airing of CSI in 2001, 2002, and 2003

Row   Pre/Post            Mean          Standard
      Comparisons      Difference        Error
1     1997-2000 vs.
      2001-2006        0.008040107    0.001893866
2     1997-2000 vs.
      2002-2006        0.006213296    0.001974616
3     1997-2000 vs.
      2003-2006        0.002533247    0.002083443
4     1997-1999 vs.
      2000             0.021331112    0.003387205
5     2000 vs.
      2001-2006        -0.00826265    0.003236207
6     2000 vs.
      2002-2006        -0.01008946    0.003284115
7     2000 vs.
      2003-2006        -0.01376951    0.003350677

                   95% CI
           Lower          Upper
Row       Bound          Bound

1     0.0043281292   0.0117520852
2     0.0023430489   0.0100835429
3     -0.001550301   0.0066167941
4      0.01469219     0.027970034
5      -0.01460562    -0.00191969
6      -0.01652633    -0.00365260
7      -0.02033684    -0.00720219

Table 9. Frequency of Various Versions of CSI Effect and Frequency
of Mention of Doubt

Version of CSI effect        Mentions   Mentions   Percentage of
                                        doubting        mentions
                                          effect      expressing
                                                           doubt

Strong prosecutor's effect        197         34             17%
Weak prosecutor's effect           74          8             11%
Educator's effect                  39          3              8%
Police chiefs effect               27          4             15%
Defendant's effect                 21          4             19%
Tech effect                         9          0              0%
Victim's effect                     6          3             50%