Read 651_1.tif text version

PROSECUTING SEXUAL ASSAULT: A COMPARISON OF CHARGING DECISIONS IN SEXUAL ASSAULT CASES INVOLVING STRANGERS, ACQUAINTANCES, AND INTIMATE PARTNERS*

CASSIA SPOHN**

University of Nebraska at Omaha

DAVID HOLLERAN***

East Tennessee State University

In this study of prosecutors' charging decisions in sexual assault cases, we test the hypothesis that the effect of victim characteristics is conditioned by the relationship between the victim and the suspect. We categorize the victim/suspect relationship as one involving strangers, acquaintances/relatives, or intimate partners, and we examine the effect of victim, suspect, and case characteristics on charging decisions in each type of case. The results of our analysis reveal that the effect of victim characteristics, with one exception, is confined to cases involving acquaintances and intimate partners. In these types of cases, prosecutors were less likely to file charges if there were questions about the victim's character or behavior at the time of the incident. In contrast, the victim's reputation and behavior did not affect charging in cases involving strangers; in those types of cases, prosecutors were more likely to file charges if the suspect used a gun or knife or if the victim was white. We conclude that stereotypes of "real rapes" and "genuine victims" continue to influence the charging decision in at least some types of sexual assault cases. * This manuscript is based on work supported by the National Institute of Justice under Grant 98-WT-VX-0003. Points of view are those of the authors and do not necessarily reflect the position of the National Institute of Justice. We would like to thank the anonymous reviewers for their helpful comments on an earlier draft of the manuscript. ** Cassia Spohn is a professor of criminal justice at the University of Nebraska at Omaha. She is the co-author of two books: The Color of Justice: Race, Ethnicity, and Crime in America (with Sam Walker and Miriam DeLone) and Rape Law Reform: A Grassroots Movement and Its Impact (with Julie Horney). She has published a number of articles examining prosecutors' charging decisions in sexual assault cases and exploring the effect of race/ethnicity on charging and sentencing decisions. Professor Spohn's current research interests include the effect of race and gender on court processing decisions, victim characteristics and case outcomes in sexual assault cases, judicial decision making, sentencing of drug offenders, and the deterrent effect of imprisonment. In 1999 she received the University of Nebraska Outstanding Research and Creative Activity Award. *** David Holleran is an assistant professor with the Department of Criminal Justice and Criminology at East Tennessee State University. He is currently interested in judicial decision making. JUSTICE QUARTERLY, Vol. 18 No. 3, September 2001 © 2001 Academy of Criminal Justice Sciences

652

PROSECUTING SEXUALASSAULT Good victims have jobs (like a stockbroker or accountant) or impeccable status (like a policeman's wife); are well-educated and articulate, and are, above all, presentable to a jury: attractive--but not too attractive, demure---but not pushovers. They should be upset--but in good taste--not so upset that they become hysterical. --An experienced sex crimes prosecutor (quoted in Bryden and Lengnick 1997:1247)

Prosecutors acting in their professional capacity play a key role in processing defendants through the criminal justice system. They, along with the police, are what Kerstetter (1990:182) calls the "gatekeepers of the criminal justice system." The prosecutor decides who will be charged, what charge will be filed, who will be offered a plea bargain, and the type of bargain that will be offered. The prosecutor also may recommend the offender's sentence. As Supreme Court Justice Robert Jackson noted in 1940, "[T]he prosecutor has more control over life, liberty, and reputation than any other person in America" (Davis 1969:190). None of the discretionary decisions made by the prosecutor is more critical than the initial decision to prosecute or not. Prosecutors have wide discretion at this stage in the process; there are no legislative or judicial guidelines on charging, and a decision not to file charges ordinarily is immune from review. According to the Supreme Court, "So long as the prosecutor has probable cause to believe that the accused committed an offense defined by statute, the decision whether or not to prosecute, and what charge to file or bring before a grand jury generally rests entirely in his discretion" (Bordenkircher v. Hayes 1978). The fact that the "prosecutor controls the doors to the courthouse" (Neubauer 1988:200) may be particularly important in cases, such as sexual assault cases, in which the victim's credibility is a potentially important issue. Observers of the charging process conclude that prosecutors attempt to "avoid uncertainty" (Albonetti 1986, 1987) by filing charges in cases where the odds of conviction are high and by rejecting charges in cases where conviction is unlikely. These studies suggest that although prosecutors' assessments of convictability are based primarily on legal factors such as the seriousness of the offense (Atbonetti 1987; Jacoby et al. 1982; Mather 1979; Miller 1969; Myers 1982; Neubauer 1974; Rauma 1984; Schmidt and Steury 1989), the strength of evidence in the case (Albonetti 1987; Feeney, Dill, and Weir 1983; Jacoby et al. 1982; Miller 1969; Nagel and Hagan 1983), and the defendant's culpability (Albonetti 1987; Mather 1979; Miller 1969; Neubauer 1974; Schmidt and Steury 1989), legally irrelevant characteristics of the

SPOHN AND HOLLERAN

653

suspect I and the victim also come into play. Stanko (1988:170), in fact, concludes t h a t "the character and credibility of the victim is a key factor in determining prosecutorial strategies, one at least as important as 'objective' evidence about the crime or characteristics of the defendant." The victim's character, behavior, and credibility m a y play an especially important role in charging decisions in sexual assault cases. In these types of cases, little physical evidence m a y exist to connect the suspect to the crime, and typically there will not be eyewitnesses who can corroborate the victim's testimony. The likelihood of conviction thus depends primarily on the victim's ability to articulate w h a t happened and to convince a judge or j u r y t h a t a sexual assault occurred. Because of this, prosecutors' assessments of convictability, and thus their charging decisions, rest on predictions regarding the way in which the victim's background, character, and behavior will be interpreted and evaluated by other decision makers, and especially by potential jurors. As F r o h m a n n (1991:535) states, this "downstream orientation" leads prosecutors to rely on stereotypes about "genuine victims" and appropriate behavior. Victims whose backgrounds and behavior conform to these stereotypes will be taken more seriously, and their allegations treated more seriously, t h a n victims whose backgrounds and behavior are at odds with the stereotypes. Although some researchers contend t h a t victim characteristics come into play in all cases of sexual assault, others argue t h a t their effect is conditioned by the n a t u r e of the case. Estrich (1987:29), for example, maintains t h a t "all women and all rapes are not treated equally." Instead, criminal justice officials differentiate between the "aggravated, jump-from-the-bushes stranger rapes and the simple cases of u n a r m e d rape by friends, neighbors, and acquaintances" (Estrich 1987:28). This point is consistent with the assertions of Black (1976) and Gottfredson and Gottfredson (1988), who argue t h a t the offender-victim relationship is an important predictor of the outcome of legal proceedings and t h a t crimes between intimates are perceived as less serious t h a n crimes between strangers. 2 1 These studies suggest that the suspect's race (Spohn, Gruhl, and Welch 1987) or the racial compositionof the suspect-victimpair (Bienen et al. 1988; Keil and Vito 1989; LaFree 1980; Paternoster 1984; Radelet and Pierce 1985) affect the prosecutor's decisionto charge; chargingis more likely if the defendant is nonwhite, or if the defendant is black and the victim is white. Other research indicates that prosecutors are more likely to file charges against men (Spohn et al. 1987) and those who are unemployed(Blumsteinet al. 1983; Schmidt and Steury 1989). 2 For a review of studies comparingcase processingdecisionsfor cases involving strangers and nonstrangers, see Simon (1996).

654

PROSECUTING SEXUALASSAULT

If (as these scholars contend) crimes involving strangers are viewed as more serious than crimes involving nonstrangers, victim characteristics--particularly those relating to the victim's credibili t y - s h o u l d be more accurate predictors of case outcomes in sexual assault cases in which the victim and the defendant are intimates or acquaintances than in cases in which they are strangers. Previous research has demonstrated that in more serious cases the outcome is determined primarily by legally relevant factors such as the seriousness of the offense, the strength of evidence in the case, and the culpability of the defendant (Kalven and Zeisel 1966; Spohn and Cederblom 1991). In such cases, criminal justice officials have relatively little discretion and thus few opportunities to consider elements such as the victim's background or her behavior at the time of the incident. In less serious cases, on the other hand, the appropriate outcome is not indicated clearly by the nature of the crime nor by other legally relevant factors; thus criminal justice officials, including prosecutors, may be more strongly disposed to take extralegal factors into consideration. Our study addresses this issue. We examine the effects of victim, suspect, and case characteristics on prosecutors' charging decisions in sexual assault cases involving strangers, acquaintances, and intimates. Using data on cases that resulted in arrest in Kansas C i t y a n d in Philadelphia, we test the hypothesis that charging decisions in cases involving strangers will be determined by legally relevant factors and that the effect of victim characteristics will be confined to cases involving acquaintances and intimates. VICTIM CHARACTERISTICS A N D OUTCOMES OF SEXUAL A S S A U L T CASES A substantial body of research examines case processing decisions in sexual assault cases. This research reveals that outcomes of sexual assault cases, like outcomes in other types of cases, are influenced strongly by legally relevant factors such as the seriousness of the crime and the offender's criminal record. A number of studies also document the influence of victim characteristics: they reveal that sexual assault case processing decisions are affected by the victim's age, occupation, and education (McCahill, Meyer, and Fischman 1979), by "risk-taking" behavior such as hitchhiking, drinking, or using drugs (Bohmer 1974; Kalven and Zeisel 1966; LaFree 1981, McCahill et al. 1979; Nelson and Amir 1975), and by the victim's reputation (Fetid and Bienen 1980; Feldman-Summers and Lindner 1976; McCahill et al. 1979; Reskin and Visher 1986).

SPOHN AND HOLLERAN

655

Outcomes of sexual assault cases also are affected by the relationship between the victim and the offender; reports of rape by strangers are investigated more thoroughly than reports of rape by a person known to the victim (McCahill et al. 1979). Stranger rapes also are less likely to be unfounded by the police (Kerstetter 1990) or rejected by the prosecutor (Battelle Memorial Institute 1977; Loh 1980; Sebba and Cahan 1973; Weninger 1978; Williams 1978), but are more likely to result in a conviction (Battelle Memorial Institute 1977) or a prison sentence (McCahill et al. 1979). 3 Evidence such as this has led to the conclusion that the response of the criminal justice system to the crime of rape is predicated on stereotypes about rape and rape victims. LaFree (1989), for example, asserts that nontraditional women, or women who engage in some type of "risk-taking" behavior, are less likely to be viewed as genuine victims who deserve protection under the law. Frohmann (1991) similarly maintains that the victim's allegations will be discredited if they conflict with decision makers' "repertoire of knowledge" about the characteristics of sexual assault incidents and the behavior of sexual assault victims. A similar conclusion was reached by the authors of a recent comprehensive review of research on the treatment of acquaintance rape in the criminal justice system (Bryden and Lengnick 1997:1326): "[T]he prosecution's heavy burden of proof has played an important role in the justice system's treatment of acquaintance rape cases, but so have public biases against certain classes of alleged rape victims" (emphasis added).4

The Contextual Effect of Victim Factors

According to many feminists, research documenting the influence of victim characteristics on sexual assault case outcomes generally signals widespread distrust of rape victims and specifically indicates that women who are regarded as behaving in sex-inappropriate ways do not receive the full protection of the law. Estrich (1987), however, asserts that historically the processing of rape

3 Spohn a n d Spears (1996), on the other h a n d , found t h a t the effect of t h e victim/offender relationship was conditioned by the racial m a k e u p of.' the victim/offender dyad. This relationship did not affect any of t h e four outcomes examined when the offender was black and the victim was white or w h e n both the offender and victim were white. The relationship came into play only w h e n both the offender and victim were black: blacks charged with sexually assaulting black acquaintances were treated more leniently t h a n blacks charged with sexually assaulting black strangers. 4 The results of some studies are inconsistent with these conclusions. Spohn a n d H o m e y (1993), for example, found t h a t extralegal victim characteristics h a d little effect on case outcomes in Detroit e i t h e r before o r after t h e reform of rape taws. Kingsnorth et at. (1998) found t h a t t h e racial composition of the offender/victim dyad did not influence any of t h e decisions points they examined a n d t h a t legally relevant factors were the primary predictors of case outcomes.

656

PROSECUTING SEXUALASSAULT

cases has not been characterized by indiscriminate sexism. She contends (p. 29) that "all women and all rapes are not treated equally" and suggests that both case law and criminal justice decision makers differentiate between aggravated and simple rapes. In accordance with Kalven and Zeisel (1966), she defines simple rape as a rape by a lone acquaintance with no weapon and no collateral injury to the victim. An aggravated rape, in contrast, involves either an attack by a stranger, multiple assailants, the use of a weapon, or injury to the victim. Because of negative presumptions about victims of simple rape, says Estrich, these incidents are less likely to be reported to the police; if they are reported, they are less likely than cases of aggravated rape to result in prosecution and/or conviction. Estrich's assertions suggest not only that outcomes for simple and for aggravated rape cases will differ, b u t also that the influence of certain victim characteristics on case outcomes will depend on whether a rape is simple or aggravated. Estrich argues that because the essential features of aggravated rape cases meet the requirements of "real rape," there is no inherent reason to distrust the victim in such cases. Thus "blame and believability" characteristics of the victim--factors ~that might lead decision makers to blame her for being victimized or to question her credibility--are ordinarily irrelevant to the processing of aggravated rape cases. In other words, if a woman was raped by a complete stranger who held a gun to her head or a knife to her throat, criminal justice decision makers normally would not care whether she had been using drugs or was employed as a topless dancer. Because simple rapes are not considered "real rapes," such victim characteristics would play a more important role in determining the outcome of these cases, according to Estrich. If a w o m a n was raped by a man she was dating, for example, her behavior at the time of the incident, her background, and her "morals" might be taken into consideration in deciding how to handle the case. Bryden and Lengnick (1997) make a similar argument, but one that focuses more explicitly on the relationship between the victim and the defendant. They observe that the defendant in a sexual assault case involving acquaintances typically claims that the victim consented. Consequently "the woman's character is inevitably a critical issue" (p. 1204). In this type of case, in other words, the jurors must evaluate the victim's character and behavior at the time of the incident in order to determine whether she fabricated or exaggerated the sexual assault or had a motive for lying about the incident. In contrast, the defendant in a sexual assault involving strangers cannot plausibly claim that the victim consented; instead

SPOHN AND HOLLERAN

657

he acknowledges that she was sexually assaulted but claims that he was misidentified. Because the defendant denies that he had sexual intercourse with the victim, he does not need to persuade the jury that she is promiscuous, vindictive, or dishonest. Moreover, in a stranger rape, "the possibility that the parties misunderstood each other's signals does not arise. As a result, the woman's character and all the controversial issues of appropriate sex roles and behavior in dating situations ordinarily are not issues" (p. 1204). Several recent studies explore the effect of victim characteristics on prosecutors' charging decisions in different types of sexual assaults. These studies generally find little support for the propositions advanced by Estrich (1987) or Bryden and Lengnick (1997). Horney and Spohn (1996), for example, used data on a sample of sexual assault complaints received by the Detroit Police Department in 1989 to formalize and test Estrich's (1987) assertions. They hypothesized that the effect of victim characteristics would be greater in simple than in aggravated rapes. Their hypothesis was not confirmed: victim characteristics did not influence case outcomes more strongly in simple than in aggravated cases. The only exception was that a prompt report to the police increased the odds of prosecution in simple rape cases but had no effect on prosecution in aggravated cases. Similar results were reported by Spears and Spohn (1997), who explored the effect of victim characteristics on Detroit prosecutors' charging decisions in sexual assault cases with strong and weak evidence and in simple and aggravated sexual assaults. These authors found that victim characteristics had a significant effect on charging decisions in cases involving both adolescent and adult victims: prosecutors were more likely to file charges if there were no questions about the victim's moral character or behavior at the time of the incident and if the victim reported the crime within one hour. They also found, however, that the victim characteristics included in their model did not exert a differential effect on prosecutors' charging decision in cases with strong or weak evidence or in aggravated and simple rape cases. These results led Spears and Spohn (1997:520-21) to conclude that "Detroit prosecutors regard the victim's moral character and behavior at the time of the incident as relevant to convictability in all types of cases." Two studies (Kerstetter 1990; Kingsnorth, Macintosh, and Wentworth 1999) examined the factors that affect charging decisions in cases involving strangers and nonstrangers. Kerstetter (1990) used data on sexual assaults reported to the police in Chicago to compare the predictors of prosecutors' decisions to file felony

658

PROSECUTING SEXUALASSAULT

charges in two types of cases: those in which the assailant's identity was at issue (primarily cases involving strangers) and those in which the complainant's consent was the issue (primarily cases involving acquaintances). He found that the strongest predictors of charging in both types of cases were evidentiary and instrumental variables; victim characteristics did not affect decision making in either type of case. In cases involving strangers, prosecutors were more likely to file charges if the victim reported the crime promptly, could identify the suspect, and was willing to prosecute. In cases involving acquaintances, the only significant predictor was the use of a weapon by the suspect. Kingsnorth and his colleagues (1999) also compared the factors influencing processing decisions in sexual assaults involving strangers and nonstrangers. They found that stranger and nonstranger cases were equally likely to be prosecuted, but different fhctors determined the likelihood of prosecution for each type of case. When the victim and the suspect were strangers, prosecution was more likely if the suspect had a prior felony record, if the victim was willing to cooperate in the prosecution of the suspect, and if there were witnesses to support the victim's account. In contrast, when the victim and the suspect were acquaintances or intimates, prosecution was more likely if the victim was injured, if the suspect made incriminating remarks or was charged with more than one crime at arrest, if the victim reported promptly and was willing to cooperate with the prosecution, and if there were witnesses who could corroborate her testimony. Further analysis revealed that only one variab l e - w h e t h e r the victim reported the crime promptly--exerted a significantly different effect on the two types of cases. Because a prompt report was relevant only in cases involving nonstrangers, Kingsnorth et al. concluded that "report time may have less relation to loss of physical evidence than to the victim's perceived credibility" (1999:290).

Limitations of Previous Research

The research reviewed above provides somewhat contradictory evidence on the factors that affect prosecutors' charging decisions in sexual assault cases. Researchers generally agree that prosecutors' attempts to "avoid uncertainty" (Albonetti 1986, 1987) and their "downstream orientation" to judges and juries (Frohmann 1997) lead them to file charges only when the odds of conviction at trial are high, but they express less agreement on the factors that define or determine "convictability." Most empirical studies of charging decisions in sexual assault cases find that legally relevant factors--

SPOHN AND HOLLERAN

659

particularly the strength of evidence in the case--play an important role. The evidence concerning the role played by victim characteristics, on the other hand, is mixed. Some studies conclude that victim characteristics--particularly the relationship between the victim and the suspect and the victim's behavior at the time of the incident--play either a primary or a secondary role; others conclude that victim characteristics are largely irrelevant. Some commentators contend that these inconsistent results reflect a failure to differentiate between aggravated and simple rapes or between cases involving strangers and nonstrangers. They argue that victim characteristics will come into play primarily in the "less serious" simple rapes: that is, cases in which the victim and the suspect are acquainted, the suspect did not use a weapon, and the victim did not suffer collateral injury. Although this proposition is consistent with research exploring the effect of legally irrelevant factors on decision making by juries and judges and with anecdotal evidence regarding prosecutorial charging decisions, it is not supported by the findings of empirical research designed to test its applicability to charging decisions in sexual assault cases. As noted above, these studies generally conclude that the effect of victim characteristics is not conditioned by the nature of the case, nor by the relationship between the victim and the suspect. These negative findings, we suggest, may reflect the fact that researchers typically classify the relationship between the victim and the suspect into only two categories: strangers and nonstrangers. Doing so ignores the diversity of these relationships and, as Decker (1993:585) comments, "mask[s] important withingroup differences." The "nonstranger" category, which includes both close and distant relationships, is particularly problematic: it includes victims and suspects who are intimate partners, relatives, good friends, and casual acquaintances. Although consent certainly is far more likely to be the defense in each of these types of nonstranger cases than in cases involving strangers, it does not necessarily follow that victim characteristics will play an identical role in each type of case. The fact that the victim invited the suspect to her apartment, for example, might have little bearing on the decision to charge or not in cases in which the victim and the suspect had an ongoing sexual relationship, but might be an important consideration if they were casual acquaintances. Similarly, the victim's failure to report the crime promptly might influence charging decisions in cases in which the victim and the suspect were neighbors or coworkers, but not in cases in which they were closely related. A further limitation of existing research is that many of the studies were conducted in a single jurisdiction, and used different

660

PROSECUTING SEXUALASSAULT

models and different statistical techniques to analyze data from the 1970s (or earlier). This is problematic for two reasons. First, it obviously is difficult to generalize from single-jurisdiction studies that vary in methodological sophistication and that reach different conclusions. Second, conclusions drawn from data gathered more than 20 years ago may no longer be valid. Changes in attitudes toward rape and rape victims, coupled with changes in the laws and rules of evidence relevant to rape (Spohn and Horney 1992), may have rendered these conclusions invalid. These changes in attitude and statutes may have resulted in more sensitive treatment of rape victims and may have reduced the likelihood that the victim's character, reputation, and behavior would affect decision making. In this study we attempt to improve on past research by defining more precisely the nature of the victim/suspect relationship and by using current data on sexual assaults that resulted in arrest in two large urban jurisdictions. We categorize the relationship as one involving strangers, acquaintances, or intimate partners, and we examine the effect of victim, suspect, and case characteristics on prosecutors' charging decisions in each type of' case. We test the hypothesis that the effect of victim characteristics is conditioned by the relationship between the victim and the suspect. We predict that the victim's character, reputation, and behavior at the time of the incident will not affect charging decisions in cases involving strangers, but will affect such decisions in cases involving acquaintances and intimate partners. Below we discuss the context of case screening, focusing on the organization of the prosecutor's office and on the procedures used to assign cases to attorneys. We also describe the case screening procedures used in the two jurisdictions--Kansas City and Philadelphia-observed in this study. THE CONTEXT OF CASE SCREENING The organization of the prosecutor's office and the procedures used to assign cases to assistant prosecutors vary from one jurisdiction to another. Some jurisdictions assign assistant prosecutors to cases on the basis of the attorney's expertise and skill; others assign attorneys to courtrooms, either permanently or for a specified period. In many large urban jurisdictions, assistant prosecutors are assigned to courtrooms, and cases are prosecuted horizontally: different prosecutors handle the case at each stage in the process. In other jurisdictions, cases are prosecuted vertically: each case is assigned to an assistant prosecutor (typically after the felony review unit has made the decision to charge), who stays with the case until

SPOHN AND HOLLERAN

661

final disposition. A number of large jurisdictions combine horizontal and vertical prosecution (Abadinsky 1988). Routine cases are prosecuted horizontally, while targeted cases (e.g., homicides, sex offenses, white-collar crimes, cases involving career criminals) are prosecuted vertically. Typically the targeted cases are assigned to specialized units within the prosecutor's office. In some jurisdictions, the prosecutors assigned to the unit will handle the case from arrest through disposition; in others, the decision to charge is made by the felony review unit, and the case is assigned to the specialized unit after screening. Researchers assert that case assignment and case screening procedures "can have important consequences for the flow of cases through the system" (Nardulli, Eisenstein, and Flemming 1988: 180-81) and can influence case outcomes. Horizontal prosecution produces attorneys who are "experts" at handling cases at particular stages of the criminal justice process, but who m a y know little about individual cases. Also, victims or complainants must deal with different attorneys at each stage of the process. Vertical prosecution, particularly within a specialized unit, produces attorneys who are especially knowledgeable about certain types of cases and who are more familiar with the cases assigned to them. In addition, "the victim or complainant has the comfort of one assistant throughout the entire judicial process; he or she does not have to discuss the case anew with each new assistant" (Abadinsky 1988:125). These structural and organizational variations can influence the processing of sexual assault cases. Such cases raise unique evidentiary issues and m a y involve victims who are reluctant to prosecute or whose credibility is in question. Prosecutors who "specialize" in these types of cases m a y handle the cases more efficiently and more effectively, and m a y treat victims more sympathetically. 5 Because one of the goals of this research project was to compare case outcomes in jurisdictions with and without specialized units for the prosecution of sexual offenses, we selected two jurisdictions

5 Predictions about t h e effects of these u n i t s are largely untested. The one exception is a study by LaFree (1989), who compared police decision making in rape cases in Indianapolis before a n d after the creation of a sex offenses unit. Although LaFree found differences in attitudes toward rape a n d rape victims between officers assigned to this u n i t a n d officers assigned to the homicide a n d robbery unit, he did not find t h a t a r r e s t and felony filing rates increased after the unit was created. He also found little support for the hypothesis t h a t extralegal variables would have less effect on decision m a k i n g in the post-reform period. LaFree concluded (1989:88) t h a t "some factors are difficult to change no m a t t e r how enlightened or motivated the public official. Arrest and felony screening require elements, such as a reasonable suspect a n d a victim willing to testify, t h a t are often beyond the police officer's ability to influence."

662

PROSECUTING SEXUAL ASSAULT

that use different procedures for screening and prosecuting sex offenses. Kansas City has a specialized unit that makes the decision to charge and uses vertical prosecution from screening through disposition; Philadelphia has a specialized unit that receives cases after a decision to charge has been made. The procedures used in these two jurisdictions are described below in greater detail.

Case Screening in Kansas City

In Kansas City, sexual assault cases that result in an arrest are assigned to the Sex Crimes Unit (SCU) of the Office of the Prosecuting Attorney for the Sixteenth Judicial Circuit of Missouri. 6 In 1999, six prosecutors were assigned to the unit, which handles all cases (i.e., homicide, assault, abuse, sex crimes) involving children as victims and all cases of forcible rape and forcible sodomy. Unlike attorneys in other units in the Office of the Prosecuting Attorney, those assigned to the Sex Crimes Unit do not rotate out regularly; most remain in the unit at least two years, and some stay "more or less permanently." In the words of the unit's trial team leader, "If they're going to get burned out, it typically will happen within the first two years; if they last that long, they will probably stay, at least until a 'better' assignment comes along." Cases that lead to an arrest are referred to the SCU by the Kansas City Police Department, and then are assigned randomly to one of the attorneys in the unit. That attorney decides whether or not to file charges. Although his or her decision is final and generally cannot be overruled by other SCU prosecutors, the unit meets once a week to discuss and evaluate ambiguous cases. If charges are filed, the attorney to whom the case was originally assigned handles the case through final disposition. If charges are declined, the case can either be rejected outright, sent back to the police for additional investigation, or sent to the city attorney for prosecution as a misdemeanor. According to the SCU's trial team leader, the standard of proof at charging is proof beyond a reasonable doubt. "We file only if we believe that we could take the case to trial and get a conviction," he stated. In addition, the unit's policy is to file charges only if there is some type of corroboration; yet the existence of corroboration does not necessarily mean that the unit will file charges in the case. "Corroboration," added the team leader, "is a necessary but not a sufficient condition for charging." If the attorney handling the case has concerns about its convictability, he or she will schedule a pre6 We obtained the information on screening and prosecution of sexual assault cases in Kansas City during a personal interview with the trial team leader of the Sex Crimes Unit, in 1999.

SPOHN AND HOLLERAN

663

file interview with the victim. The interview is designed to assess the victim's credibility and genuineness, to evaluate the strength of evidence in the case, and to answer any questions about the incident t h a t the prosecutor might raise.

Case Screening in Philadelphia

In Philadelphia, sexual assault cases that lead to an arrest are reviewed by the charging unit of the Office of the District Attorney, which operates around the clock. 7 A prosecutor assigned to the unit reviews the information contained in the police report and evaluares the case for legal sufficiency. In other words, the screening that takes place at this stage in the process focuses on the presence or absence of the statutory elements of the offense, not on the likelihood of conviction at trial. According to the chief of the Family Violence and Sexual Assault Unit (FVSAU) of the Office of the District Attorney, the "real screening" takes place in the Sex Crimes Unit (SCU) of the Philadelphia Police Department. As a result, the charging unit rarely declines charges in sexual assault cases, s This point was confirmed by a lieutenant in the Sex Crimes Unit, who reported that the officers in the SCU work closely with prosecutors in the FVSAU, and seek their advice in ambiguous cases. This officer explained that the SCU applies a more stringent standard than probable cause in making an arrest, and added that the officer investigating the case will not issue an arrest w a r r a n t unless the case is solid and the evidence strong. "By the time we get to the point of making an arrest," he said, "questions regarding the credibility of the victim and the strength of evidence in the case already have been answered. "9 After the charging unit has reviewed the case and has made a decision to file charges, the case is referred to the Family Violence

7 We obtained the information on screening and prosecution of sexual assault cases in Philadelphia during personal interviews with the head of the Family Violence and Sexual Assault Unit and with a lieutenant in the Sex Crimes Unit of the Philadelphia Police Department, held in 1998. s The data obtained for this project reveal that 43 (16.1%) of the 267 cases that resulted in an arrest in 1997 were rejected at screening. An additional 92 (34.5%) cases were dismissed by the FVSAU after charges were filed. 9 Questions regarding the handling of rape complaints by the Sex Crimes Unit were raised in a two-part series, headlined '~Vomen Victimized Twice in Police Game of Numbers," which appeared in the Philadelphia Inquirer in October 1999. In January 2000, the Philadelphia Police Department reopened more than 2,000 rapes, indecent assaults, and other sex offenses reported from 1995 to 1999 which had been inappropriately closed and classified as "investigation of person." According to the Philadelphia Inquirer ("Police Reopen Thousands of Dumped Cases" 2000), the cases were flagged for reinvestigation by department auditors, who began reviewing sexcrimes files in October 1999. The reporter stated "The auditors identified as least 1,800 cases from 1995, 1996, and 1997 that were closed without adequate investigation and improperly given administrative designations, such as 'investigation of person,' that do not appear in the city's crime statistics."

664

PROSECUTING SEXUAL ASSAULT

and Sexual Assault Unit for prosecution. The FVSAU, which includes 15 prosecutors, is responsible for prosecuting all sexual offenses involving adults, physical and sexual abuse cases involving children, and all misdemeanor and felony domestic violence cases. Prosecutors assigned to the unit do not rotate out at regular intervals but remain with the unit as long as they wish; those who leave generally move to a supervisory position or to the homicide unit. As noted above, the screening standard applied by the charging unit is "legal sufficiency," not "convictability." Cases assigned to the FVSAU therefore undergo a second screening that focuses on the likelihood of conviction at trial. The attorney reviewing the case can dismiss the charges before the preliminary hearing or can file a motion to dismiss the case at the preliminary hearing; the attorney also can dispose of the case by agreeing to allow the defendant to plead guilty to a misdemeanor. Although cases assigned to the FVSAU are prosecuted vertically, the attorney who handles the case at the preliminary hearing is not necessarily the attorney who takes the case to trial. The chief of the FVSAU explained that cases bound over for trial after the preliminary hearing are assigned to attorneys in the FVSAU on the basis of the "difficulty" of the case: the more difficult cases are assigned to the more experienced attorneys.

RESEARCH DESIGN Sampling and Data Collection Procedures

In this paper we analyze data on sexual assaults that resulted in arrest in Kansas City and Philadelphia. We used different criteria for selecting cases and different procedures to collect data in the two jurisdictions. In Kansas City, we selected all cases that met the following criteria: the defendant was arrested in 1996, 1997, or 1998 for rape, forcible sodomy, sexual assault, deviate sexual assault, first-degree statutory rape, or first-degree statutory sodomy; 1° the case was referred to the Office of the Prosecuting

1o The Missouri statutes define forcible rape as sexual intercourse with another person by the use of forcible compulsion (RSMo §566.030). Forcible sodomy is defined as deviate sexual intercourse ("any act involving the genitals of one person and the mouth, tongue, or anus of another person or a sexual act involving the penetration, however slight, of the male or female sex organ or the anus by a finger, instrument, or object . . .") with another person by the use of forcible compulsion (RSMo §566.060). First degree statutory rape is defined as sexual intercourse with another person who is less t h a n 14 years old (RSMo §566.032); first degree statutory sodomy is defined as deviate sexual intercourse with a person who is less t h a n 14 years old (RSMo §566.062). All four of these offenses are ungraded felonies, for which the authorized term of imprisonment is life or a term of not less than five years. If, however, the actor inflicts serious physical injury or displays a deadly or dangerous i n s t r u m e n t in a threatening manner, or subjects the victim to intercourse with more t h a n one person, the authorized term of imprisonment is life or a term of

SPOHN AND HOLLERAN

665

Attorney for the Sixteenth Judicial Circuit of Missouri by the Kansas City Police Department; and the victim was age 12 or older. In Philadelphia, we selected all cases of rape, involuntary deviate sexual intercourse, and sexual assault 11 involving victims age 12 or older that resulted in an arrest during 1997. For this study, we eliminated cases involving male victims and female suspects. 12 This resulted in a data file that included 259 cases in Kansas City and 267 cases in Philadelphia. The procedures used to obtain the data also varied by jurisdiction. In Philadelphia, we obtained a list of all sexual assault cases that met our selection criteria from the Sex Crimes Unit of the Philadelphia Police Department. The unit gave us access to its case files and allowed us to use the department's computer, which was linked to the Philadelphia Court of Common Pleas, to determine the final disposition of the case. In Kansas City, we obtained the data from case files maintained by the Sex Crimes Unit of the Office of the Prosecuting Attorney. 13 That unit provided us with a list of the cases that met our selection criteria and allowed us to examine the files for these cases. In each jurisdiction, we read. through the documents and reports in the case file and recorded detailed information about the incident, the victim, the suspect, and the outcome of the case, using an optical-scan form designed for the project.

not less t h a n 10 years. Sexual assault, a class C felony, is defined as sexual intercourse with another person without t h a t person's consent (RSMo §566.040). Deviate sexual assault, a class C felony, is defined as deviate sexual intercourse with another person without t h a t person's consent (RSMo §566.070). 11 The Crimes Code of Pennsylvania defines rape, a first degree felony, as sexual intercourse u n d e r any of the following circumstances: forcible compulsion; t h r e a t of forcible compulsion; where complainant is unconscious a n d unaware; where complainant is substantially impaired by actor t h r o u g h use of intoxicants, drugs, or other m e a n s designed or intended to reduce resistance; where complainant is, by m e n t a l disability, incapable of consent; or where complainant is less t h a n 13 years of age (18 Pa.C.S.A. §3121). Involuntary deviate sexual intercourse, which also is a felony of the first degree, is defined as deviate sexual intercourse (oral or anal sexual penetration or pentetration with a n object) u n d e r any of the conditions outlined above, or where the complainant is less t h a n 16 years of age a n d the actor is four or more years older a n d the complainant a n d actor are not married to each other (18 Pa. C.S.A. §3123). Sexual assault, a felony of the second degree, is defined as sexual intercourse or deviate sexual intercourse without the complainant's consent (18 Pa. C.S.A. §3124.1). 12 There were 17 male victims and four female suspects in Kansas City, and 14 male victims a n d one female suspect in Philadelphia. 13 We first a t t e m p t e d to obtain permission to examine the files m a i n t a i n e d by t h e Kansas City Police Department. We were informed, however, t h a t the department's sexual assault case files were not stored in a separate location, but were intermixed with the case files for all other felonies. Because obtaining the files would have been very time-consuming a n d would have required the services of a departm e n t employee, our request was denied.

666

PROSECUTING SEXUAL ASSAULT

Dependent and Independent Variables

The dependent and independent variables used in this study are presented in Table 1. We present separate data for each of the three types of relationship. As explained above, the screening procedures vary in the two jurisdictions included in this study; therefore the operational definition of the dependent variable--the decision to file charges--also varies. In Kansas City, where the charging decision is made by the Sex Crimes Unit of the Office of the Prosecuting Attorney, all cases rejected at the initial screening were coded 0; all cases not rejected at the initial screening were coded 1. In Philadelphia, where the initial charging decision is pro forma and where cases undergo a more rigorous screening by the Family Violence and Sexual Assault Unit after formal charges are filed, we coded 0 all cases rejected at the initial screening or dismissed before the preliminary hearing; all cases not screened out at these two stages were coded 1. Table 1. D e s c r i p t i v e Statistics for Variables Included in Models o f O f f e n d e r N i c t i m Relationship: Means

Stranger (N = 113) Dependent Variable Prosecutor charged (charges filed=l) Independent Variables Victim characteristics White victim (white=l) Victim's age Risktaking by victim (yes=l) Questions about moral character (yes=l) Victim physically resisted (yes=l) Incident reported in one hour (yes=l) Suspect characteristics White suspect (white=l) Suspect's age Any prior felony convictions (yes=l) Case characteristics Gun/knife used (yes=l) Injury to victim (yes=l) Physical evidence available (yes=l) Witnesses to incident (yes=l) Philadelphia (Philadelphia=l) NOTE: Standard deviations in parentheses. .55 .41 28.50 (14.37) .48 .41 .77 .68 .21 30.41 (8.87) .53 .30 .39 .73 .52 .53 Acquaintance (N = 297) .53 .31 21.28 (10.30) .39 .29 .69 .21 .24 32.49 (12.92) .33 .10 .12 .51 .47 .52 Partner (N = 116) .52 .32 27.99 (9.61) .33 .52 .81 .36 .23 31.75 (8.41) .44 .20 .47 .62 .29 .46

The primary objective of this study is to compare the predictors of charging decisions in cases involving strangers, acquaintances, and intimates. Initially we classified the relationship between the

SPOHN AND HOLLERAN

667

victim and the suspect into 11 categories: stranger; not met, but have seen before; authority figure; boyfriend of mother, other relative; acquaintance or friend; prior dating or sexual relationship; current dating or sexual relationship; relative; living together as boyfriend and girlfriend; former spouse; and spouse. Because several of the categories contained only a few cases, we reclassified the 11 types of relationship into three categories. We classified cases in which the suspect and the victim were complete strangers (n = 82), or in which the victim had not met and could not identify the suspect (n = 31), as cases involving "strangers." We categorized cases in which the suspect and victim were relatives (n = 57) or friends or acquaintances (n = 202), or in which the suspect was either an authority figure (n = 8) or the boyfriend of the victim's mother or another relative (n = 30), as cases involving "acquaintances." The final category, "intimate partners," includes cases in which the victim and the suspect were (or had been) dating (n = 89), were currently living together (n = 16), or were (or had been) married to each other (n = 11). We labeled this category "intimate partners" r a t h e r than "partners" because most of the relationships involved prior consensual sexual intercourse; 58 of the 63 (92.1%) victims in Kansas City and 47 of the 53 (88.7%) victims in Philadelphia indicated that they had had a prior sexual relationship with the suspect. The other independent variables included in this study are also listed in Table 1. Previous research examining case processing decisions, including the decision to charge or not, typically differentiated between legal and extralegal predictors of decision making. We abandon this dichotomy and instead examine victim characteristics, suspect characteristics, and case characteristics. Our decision is motivated by two concerns. First, although the concepts of legal and extralegal factors have been used extensively by researchers analyzing and predicting the outcomes of criminal cases, they have been defined neither precisely nor consistently. There is disagreement as to the proper categorization of factors even among researchers examining the effect of these factors on the outcomes of sexual assault cases. Some researchers, for example, categorize a prompt report to the police as an extralegal variable; others contend that its relationship to the preservation of evidence makes it legally relevant (Bryden and Lengnick 1997; Kerstetter 1990). Second, m a n y of the extralegal factors traditionally examined by researchers are characteristics of the defendant (for example, race, gender, and social class). In this project, however, we are particularly interested in the effect of victim characteristics.

668

PROSECUTING SEXUAL ASSAULT

We further subdivide victim characteristics into background factors (the victim's gender, race, and age) and what we call "blame and believability" factors. As noted above, these are characteristics of the victim that might cause criminal justice officials to blame the victim and/or question her credibility. We control for whether the victim physically resisted her attacker (yes = 1; no = 0) or made a prompt report to the police (reported in one hour or less = 1; reported in more than one hour = 0); w h e t h e r there were questions about the victim's "moral character"; and whether the victim was engaged in any type of risk-taking activity at the time of the incident. The moral character variable is coded 1 ff the police file contained information about the victim's prior sexual activity with someone other than the suspect, out-of-wedlock pregnancy or birth, a pattern of alcohol and/or drug abuse, prior criminal record, work as a prostitute, or work as an exotic dancer or in a massage parlor. The risk-taking variable is coded 1 if the police file indicated that the victim, at the time of the assult, was walking alone late at night, was hitchhiking, was in a bar alone, was using alcohol or drugs, willingly accompanied the suspect to his residence, or invited the suspect to her residence. Three suspect characteristics--the suspect's age, race, and prior criminal record--are included in the analysis. We expect that suspects with prior criminal records will be more likely to be charged: prior criminal record is a dichotomous indicator of whether or not the suspect had a prior felony conviction. The suspect's age and race are included as control variables. The analysis includes several case characteristics that reflect either the seriousness of the offense or the strength of evidence in the case. Measures of offense seriousness include w h e t h e r the suspect used a gun or knife during the assault (yes = 1; no = 0) and whether the victim suffered collateral injuries such as bruises, cuts, burns, or internal injuries (yes = 1; no = 0 ) . 14 The strength of evidence in the case is measured by the existence of a witness to the assault (yes = 1; no = 0) and by the presence of physical evidence, such as semen, blood, clothing, bedding, or hair, that can corroborate the victim's testimony (yes = 1; no = 0).

14 We did not include a control for the most serious charge at arrest because more t h a n 80 percent of the defendants were arrested for forcible rape or sodomy. Inclusion of this variable in the model did not change any of the results reported here.

SPOHN AND HOLLERAN

669

RESULTS

M e a n s for all v a r i a b l e s a r e p r e s e n t e d in T a b l e i for e a c h t y p e of v i c t i m / s u s p e c t r e l a t i o n s h i p . 15 ( P e a r s o n p r o d u c t - m o m e n t c o r r e l a tions a r e p r e s e n t e d in A p p e n d i x T a b l e s A 1 , A2, a n d A3). 16 T h e s e d a t a s h o w t h a t t h e p r o s e c u t o r filed c h a r g e s in slightly m o r e t h a n h a l f of t h e cases in e a c h r e l a t i o n s h i p category. T h e y also show t h a t t h e t y p i c a l victim, r e g a r d l e s s of t h e r e l a t i o n s h i p b e t w e e n s u s p e c t a n d victim, w a s a b l a c k w o m a n in h e r m i d - to l a t e t w e n t i e s a n d t h a t t h e t y p i c a l s u s p e c t w a s a b l a c k m a l e in his e a r l y t h i r t i e s F We f o u n d s o m e differences in t h e c h a r a c t e r i s t i c s of c a s e s in e a c h category. A h i g h e r p r o p o r t i o n of v i c t i m s w e r e e n g a g e d in s o m e f o r m of r i s k - t a k i n g b e h a v i o r a t t h e t i m e of t h e i n c i d e n t in cases inv o l v i n g s t r a n g e r s t h a n in cases i n v o l v i n g a c q u a i n t a n c e s or i n t i m a t e p a r t n e r s . E v i d e n c e calling t h e v i c t i m ' s m o r a l c h a r a c t e r into q u e s tion w a s m o r e c o m m o n in c a s e s i n v o l v i n g s t r a n g e r s a n d i n t i m a t e p a r t n e r s t h a n in c a s e s i n v o l v i n g a c q u a i n t a n c e s . I n addition, we f o u n d a c o n s i d e r a b l y h i g h e r p r o p o r t i o n of p r o m p t r e p o r t s of s e x u a l a s s a u l t s i n v o l v i n g s t r a n g e r s t h a n of s e x u a l a s s a u l t s i n v o l v i n g acq u a i n t a n c e s or p a r t n e r s , is M o r e o v e r , s t r a n g e r cases m o r e often included p h y s i c a l e v i d e n c e or a w i t n e s s w h o could c o r r o b o r a t e t h e v i c t i m ' s a l l e g a t i o n s . S u s p e c t s in c a s e s i n v o l v i n g s t r a n g e r s also w e r e m o r e likely t h a n t h o s e in t h e t w o o t h e r c a t e g o r i e s to h a v e u s e d a g u n or k n i f e d u r i n g t h e a s s a u l t a n d to h a v e a t l e a s t one p r i o r felony conviction. S o m e w h a t s u r p r i s i n g l y , v i c t i m s in cases involving i n t i m a t e p a r t n e r s d i s p l a y e d a h i g h e r r a t e of c o l l a t e r a l i n j u r y t h a n t h o s e in c a s e s i n v o l v i n g s t r a n g e r s or a c q u a i n t a n c e s , m To m o d e l t h e r e l a t i o n s h i p b e t w e e n victim, s u s p e c t , a n d case c h a r a c t e r i s t i c s a n d t h e decision to file c h a r g e s in s e x u a l a s s a u l t 15 The arithmetic mean is not presented as the measure of central tendency for binary items; instead we present the proportion of cases in the category that are coded "1." 16 As observed by Belsley, Kuh, and Welsch (1980), Pearson product-moment correlations, in and of themselves, do not provide an adequate test for cotlinearity because ill-conditioning of regressors can exist elsewhere in the matrix of independent variables. Accordingly, we followed their recommendation: for each relationship type, we estimated an ordinary least squares regression through the origin and computed condition indices and variance decomposition proportions. The character~ istic root did not, exceed the suggested threshold value (~ < 30). 17 These are the characteristics of sexual assault cases that resulted in an arrest, not of all cases that occurred or that were reported to the police. is Although this may suggest that victims are reluctant to report assaults involving suspects with whom they are acquainted or intimately involved, it also may indicate that the police find it mere difficult to identify the suspect and make an arrest if the victim does not report the crime promptly. 19 This suggests that victims of sexual assault by an intimate partner did not report the crime to police unless they were injured, and/or that police did not arrest the suspect unless there was some type of collateral injury. Another plausible explanation is that sexual assault between intimate partners is the culmination of an incident of physical violence.

670

PROSECUTING SEXUALASSAULT

cases, we employed a general linear model. Because the outcome measure was binary (charges filed = 1, charges not filed = 0), we estimated binary logistic regressions. To test for the effect of type of relationship, we first estimated a model for the entire sample. At this stage in the analysis we included two d u m m y variables (ACQUAINTANCE and INTIMATE PARTNER) measuring the type of relationship between the victim and the offender; STRANGER is the omitted category. Then, to test our hypothesis concerning the contextual effects of victim characteristics, we estimated separate logistic regression models for each of the three types of relationships. Table 2 presents the results of our analysis of the decision to charge or not, using the entire sample of cases. In contrast to previous research (Albonetti 1987; Hepperle 1985), we found t h a t the type of relationship between the victim and t h e suspect did not affect the likelihood of charging: prosecutors were n o t more likely to file charges in cases involving strangers t h a n in cases involving acquaintances/relatives or intimate partners. 2° Instead, charging was more likely if there was physical evidence t h a t could corroborate the victim's testimony, if the suspect previously had been convicted of a felony, if the victim did not engage in any type of risktaking behavior at the time of the incident, and if there were no questions about the victim's moral character. In these two jurisdictions, then, charging decisions reflected a combination of legaUy relevant suspect and case characteristics and legally irrelevant victim characteristics. Our analysis of the data, partitioned by the victim/suspect relationship, is summarized in Table 3. The results show t h a t the presence of physical evidence connecting the suspect to the crime exerted a strong and statistically significant effect on charging in all three types of cases. To determine whether the magnitude of the effect of physical evidence varied depending on the type of relationship, we calculated the z test for the equality of regression coefficients (Paternoster et al. 1998). 21 We found t h a t physical evidence had a stronger effect on cases involving strangers t h a n on cases involving acquaintances (z = 2.01); however, the effect of physical evidence did not vary between cases involving strangers and those 20 We foundthis pattern ofresults in bothjurisdictions. The results ofthe sitespecific analysis were as follows:Kansas City: acquaintance/relative(b = .005; SE = .49) intimate partner (b = -.254; SE = .52); Philadelphia: acquaintance/relative(b = .54; SE = .38), intimate partner (b = .08; SE = .45). 21 The formula for the z4est is Zb 1 - 52

~]SEbl2+SEb22

SPOHN AND HOLLERAN

671

Table 2.

Logistic Regression Results for All Cases

B SEB .32 .01 .21 .23 .22 .23 .36 .01 .21 .30 .25 .22 .20 .21 -.29 .32 .49 .15 .20 79.93/16" 500 exp(B) 1.44 .99 .39 .74 .57 1.49 .84 1.01 1.52 1.72 1.43 2.36 1.18 .75 -1.29 .99

Victim Characteristics White victim (white=l) Victim's age Risktaking by victim (yes=l) Questions about moral character (yes=l) Victim physically resisted (yes=l) Incident reported in one hour (yes=l) Suspect Characteristics White suspect (white=l) Suspect's age Any prior felony convictions (yes=l) Case Characteristics Gun/knife used (yes=l) Injury to victim (yes=l) Physical evidence available (yes=l) Witnesses to incident (yes=l) Philadelphia (Philadelphia=l) Victim/Suspect Relationship Strangers (reference group) Acquaintances Intimate partners Constant Cox and Snell R 2 Nagelkerke R 2 Chi-Square/df Number of Cases * p < .05

.36 .00 -.95* -.56* -.30 .40 -.18 .01 .42* .54 .36 .86* .17 -.28 -.25 -.01 -.30

involving i n t i m a t e s (z = .89) or b e t w e e n cases involving a c q u a i n t ances a n d t h o s e involving i n t i m a t e s (z = 1.21). T h e r e s u l t s p r e s e n t e d in T a b l e 3 also s u p p o r t o u r h y p o t h e s i s t h a t the effect of victim c h a r a c t e r i s t i c s would be confined to cases involving a c q u a i n t a n c e s a n d i n t i m a t e s a n d t h a t c h a r g i n g decisions in cases involving s t r a n g e r s would be d e t e r m i n e d b y legally relev a n t factors. In k e e p i n g w i t h our hypothesis, w h e n t h e victim a n d t h e s u s p e c t w e r e s t r a n g e r s , p r o s e c u t o r s w e r e m o r e likely to file c h a r g e s i f t h e r e w a s physical evidence to c o n n e c t t h e s u s p e c t to t h e crime a n d i f t h e s u s p e c t u s e d a g u n or a knife d u r i n g t h e a s s a u l t . In fact, as s h o w n by the odds ratios, p r o s e c u t o r s w e r e n e a r l y eight t i m e s m o r e likely to file charges if t h e r e was physical evidence a n d m o r e t h a n 51/2 t i m e s m o r e likely to do so if t h e suspect u s e d a g u n or a knife. S o m e w h a t surprisingly, c h a r g i n g decisions in cases involving s t r a n g e r s also w e r e affected b y the victim's race: p r o s e c u t o r s w e r e 4% t i m e s m o r e likely to file c h a r g e s i f t h e v i c t i m w a s white. T h e r e s u l t s of o u r a n a l y s i s of the two n o n s t r a n g e r categories also are c o n s i s t e n t w i t h our hypothesis. As predicted, victim characteristics came into p l a y in each of t h e s e t y p e s of cases. In cases involving a c q u a i n t a n c e s / r e l a t i v e s , for example, p r o s e c u t o r s w e r e

Table 3.

Stranger SEB .75 .02 .58 .62 .62 .57 .82 .03 .56 .65 .59 .66 .54 .58 1.53 .34 .47 46.73/14" 109 5.62 1.67 7.90 1.24 .48 -.24 -.10 .62* .13 -.06 .02 .47 .41 .29 .26 .29 .51 .11 .14 31.77/14" 277 1.27 1.02 .84 .08 .01 .59* ,50 .01 .29 1.08 1.01 1.80 1.27 .90 1,85 1.14 .94 --1.41 .00 .30 -.29 1.01" 1.33" .30 -.80 -.64 4.47 1.02 .76 .62 1.07 2.41 -.11 .00 -.66* -.71" -.40 .11 .46 .01 .29 .30 .30 .33 .89 .99 .52 .49 .67 1.11 ,79 .02 -1.93" .72 -1.44" .97 exp(B) B SEB exp(B) B SEB 1.08 .04 .58 .52 .66 .52 1.12 .04 .49 .64 .53 .51 .51 .53 1.34 .32 .42 43.38/14" 114 Acquaintance Partner

Logistic Regression Results for Prosecutors' D e c i s i o n s to File Charges for Cases I n v o l v i n g Strangers, A c q u a i n t a n c e s , and Intimate P a r t n e r s

"41 b~

© o~ exp(B) 2.20 1.02 .14 1.91 .24 2,65 .24 1.00 1.34 .75 2.76 3.77 1.35 .45 --

B

1.49" .02 -.27 -.47 -.08 .88

.24 .02 -.17

Victim Characteristics White victim (white=l) Victim's age Risktaking by victim (yes=l) Questions about moral character (yes=l) Victim physically resisted (yes=l) Incident reported in one hour (yes=l) Suspect Characteristics White suspect (white=l) Suspect's age Any prior felony convictions (yes=l) Case Characteristics Gun/knife used (yes=l) Injury to victim (yes=l) Physical evidence available (yes=l) Witnesses to incident (yes=l) Philadelphia (Philadelphia=l) Constant Cox and Snell R 2 Nagelkerke R 2 Chi-Square/df Number of Cases

1,72" .51 2.07* .22 -.74 -3.57

* p < .05

SPOHN AND HOLLERAN

673

significantly less likely to file charges if the victim engaged in any type of risk-taking behavior at the time of the incident or if there were any questions about the victim's reputation or "moral character." In such cases, charging also was more likely if the suspect had a prior felony conviction. In cases involving intimate partners, two victim characteristics influenced the decision to file charges: prosecutors were less likely to file charges if the victim engaged in risktaking behavior 22 or if she resisted her attacker. Prosecutors also were more likely to file charges in these types of cases if the victim suffered some type of collateral injury during the assault. 23 The patterns revealed by our analyses are illustrated more clearly by the data presented in Table 4. Using the formula suggested by Liao (1994:12), we converted the logistic regression coefficients to predicted probabilities. We set all variables, except the variables of interest, at their mean. We used the following formula to calculate the predicted probabilities: P1 = exp(Z1)/1 + exp(Z1) where Z1 = E B~ik The first panel of Table 4 displays the predicted probabilities of charging in cases involving strangers. Because the victim's race and the use of a gun or knife during the assault were significant predictors of charging in these types of cases, we estimated probabilities for four victim's race/weapon combinations: black victim/no weapon, black victim/weapon, white victim/no weapon, and white victim/weapon. Prosecutors were least likely to file charges when the victim was black and the suspect did not use a weapon; they were most likely to file charges when the victim was white and the suspect used a weapon. The probability of charging in the other two types of cases also was high: 70 percent for white victim/no weapon cases and 75 percent for black victim/weapon cases. Use of a gun or knife, then, increases twofold the probability of charging in

22 The z-test revealed t h a t risk-taking behavior h a d a significantly greater effect on cases involving i n t i m a t e s t h a n on cases involving acquaintances (z = 1.96). 23 One reviewer suggested t h a t cases be disaggregated according to relationship type a n d t h a t we t h e n estimate separate logistic regression for each site. We could not do this because of the small n u m b e r of cases in each jurisdiction in which the victim a n d the suspect were s t r a n g e r s or i n t i m a t e partners. (We found 52 s t r a n ger cases in Kansas City a n d 57 in Philadelphia; t h e r e were 62 i n t i m a t e cases in K a n s a s City and 52 in Philadelphia.) Our logistic regressions produced coefficients via m a x i m u m likelihood. With a small n u m b e r of cases, however, the consistency of coefficients produced by this numerical method of estimation is questionable. Alt h o u g h a precise threshold n u m b e r of cases is not known, Long (1997:53-54) recomm e n d s not relying on estimates produced with samples of less t h a n 100. Moreover, the n u m b e r of regressors given t h e n u m b e r of cases m u s t be t a k e n into consideration (Eliason 1993:83; Long 1997:54). Given only 50 to 60 cases per model a n d 13 regressors, we were reluctant to estimate these equations.

674

PROSECUTINGSEXUALASSAULT

T a b l e 4. P r e d i c t e d P r o b a b i l i t i e s o f C h a r g i n g i n D i f f e r e n t T y p e s of S e x u a l Assault Cases Predicted Probability of Charging Sexual Assaults Involving Strangers: Victim's Race and Suspect's Use of a Weapon Black victim/no weapon Black victim/weapon White victim/no weapon White victim/weapon Sexual Assaults Involving Acquaintances/ Relatives: Risk-Taking Behavior by Victim and Questions About the Victim's Moral Character Risky behavior/questions about moral character Risky behavior/no questions about moral character No risky behavior/questions about moral character No risky behavior/no questions about moral character Sexual Assaults Involving Intimate Partners: Risk-Taking Behavior by Victim and Injury to the Victim Risky behavior/no injury Risky behavior/injury No risky behavior/no injury No risky behavior/injury

.34 .70 .75 .90

.34 .70 .75 .90

.33 .57 .77 .90

cases involving black women, but has a more modest effect in cases involving white women. The probabilities presented in the second panel of Table 4 are based on the two victim characteristics that were significant predictors of charging in sexual assaults involving acquaintances or relatives: risk-taking behavior by the victim at the time of the incident and questions about the victim's moral character. The predicted probabilities of charging were very similar if there was evidence that the victim had engaged in some type of risky behavior at the time of the incident (predicted probability = .55) or if there were questions about the victim's reputation or character (predicted probability = .57). The presence of both of these factors substantially reduced the likelihood of charging; the absence of both increased the odds significantly. For cases involving intimate partners, we calculated predicted probabilities based on two of the significant variables in the model: whether the victim engaged in risky behavior and whether the victim suffered some type of collateral injury during the assault. As shown in the third panel of Table 4, prosecutors were nearly three

SPOHN AND HOLLERAN

675

times more likely to file charges if the victim did not behave in a risky manner and if she was injured (predicted probability = .90) than if she engaged in risk-taking behavior and was not injured (predicted probability = .33). In these types of cases, the victim's behavior at the time of the incident has a more pronounced effect on the predicted probability of charging than does injury to the victim. This influence is reflected in the fact that the odds of charging are relatively high, regardless of injury to the victim, in cases in which the victim did not engage in any risky behavior; they are relatively low, again regardless of victim injury, when the victim behaved in a risky manner. 24

Effect of Moral Character and Risk Taking

The findings discussed thus far indicate that the effect of evidence about the victim's character and risk-taking behavior at the time of the incident was confined to cases involving nonstrangers. Our measures of moral character and risky behavior, however, are composite variables that reflect the presence or absence of eight character issues and nine types of behavior. It is possible that charging in cases involving strangers is affected by certain types of character or behavior evidence; similarly, the various types of character and behavior evidence may exert differential effects on charging decisions in cases involving nonstrangers. To test these possibilities, we created three more specific risktaking variables: the victim used drugs or alcohol at the time of the incident; the victim was in a bar alone, was walking alone late at night, or was hitchhiking; and the victim either invited the suspect to her home or went with him to his home. We also created three more specific moral character variables: the victim had engaged in prior consensual sex with someone other t h a n the suspect; the victim had a history of drug or alcohol use; and the victim was a prostitute or worked as a stripper or exotic dancer. We then reran the analysis, substituting these variables for the original measures of risk taking and moral character. 25

24 One of the anonymous reviewers suggested t h a t charging decisions in sexual assault cases between i n t i m a t e s m i g h t be affected by the n a t u r e of t h e relations h i p - - t h a t is, by w h e t h e r or not t h e relationship was ongoing. Accordingly we created a variable t h a t we coded 1 if, a t the time of t h e assault, the victim and suspect were (1) dating or involved in a sexual relationship, (2) living together, or (3) married. We coded this variable 0 if the victim a n d suspect h a d a prior dating or sexual relationship or were ex-spouses. W h e n we r e r a n the analysis, we found t h a t this variable did not significantly affect t h e likelihood of charging (b = -.78; SE = .56). 25 In only seven cases involving acquaintances and one case involving intimates was the victim walking alone, in a bar alone, or hitchhiking; in only six cases involving acquaintances a n d four cases involving p a r t n e r s was the victim a prostitute or a stripper/exotic dancer. Therefore we did not include these variables in t h e analysis of charging in these two types of cases. We also excluded, from the analysis

676

PROSECUTING SEXUAL ASSAULT

We found that none of the more precisely defined variables affected charging decisions in cases involving strangers. In cases involving acquaintances and relatives, however, charging was affected by one of the moral character variables: prosecutors were less likely to file charges if the victim had a history of drug or alcohol abuse. In contrast, charging decisions in cases involving intimate partners were affected by two of the risk-taking variables: the likelihood of charging was reduced significantly if the victim used drugs or alcohol at the time of the incident or if she invited the suspect to her residence or went with him to his home. 26

DISCUSSION

Our examination of prosecutors' charging decisions in sexual assault cases confirms that the "prosecutor controls the doors to the courthouse" (Neubauer 1988:200). Prosecutions occurred in only about half of the sexual assault cases that resulted in an arrest in the two large urban jurisdictions included in this study. Our results also confirm that the decision to charge or not is based on a combination of victim, suspect, and case characteristics. Prosecutors were more likely to file charges if there was physical evidence connecting the suspect to the crime, if the suspect had a prior criminal record, and if there were no questions about the victim's character or behavior at the time of the incident. This finding suggests that prosecutors' concerns about convictability lead them to file charges when the evidence is strong, the suspect is culpable, and the victim is blameless. 27 Our results also reveal that the relationship between the victim and the suspect had no effect on the decision to charge or not. Although this finding is consistent with the results of one recent study of sexual-assault case-processing decisions (Kingsnorth et at. 1999), it is inconsistent with the assertions of Black (1976) and

of cases involving strangers (N = 9), the variable m e a s u r i n g the victim's prior consensual sex with someone other t h a n the suspect. 26 Results of these analyses are available ~ o m the first author. 27 As suggested by one of the anonymous reviewers, the decision to prosecute or not also may be affected by the victim's willingness to cooperate; this, in turn, may be affected by the extralegal behavior and character factors. Although we could not obtain this information in either jurisdiction, in Kansas City we were able to determine the prosecutor's reasons for rejecting the change. (Because we were not allowed to examine the prosecutors' files in Philadelphia, we could not obtain this information there.) In 41 cases, the victim either could not be located or asked t h a t the charges be dropped; the prosecutor filed charges in only six of these cases. Moreover, risk-taking behavior by the victim (but not evidence regarding moral character) was associated with t h e victim's willingness to prosecute: t h a t is, victims whose behavior at the time of the incident was questionable were sigzlificantly more likely to disappear or to ask t h a t t h e case be dropped.

SPOHN

AND

HOLLERAN

677

others, who contend that the victim-suspect relationship is an important predictor of case outcomes and that crimes between intimates are perceived as less serious than crimes between strangers. Silberman (1978:265), for example, asserts that "prosecutors distinguish between 'real crimes'--crimes committed by strangers--and 'junk (or garbage) cases" in which the victim and the suspect are acquainted. The fact that the victim/suspect relationship did not affect charging is particularly surprising, given our operational definition of the relationship between the victim and the suspect. Unlike previous researchers, who dichotomized the relationship as involving strangers or nonstrangers, we differentiated between crimes involving strangers, acquaintances/relatives, and intimate partners. We reasoned that cases involving intimate partners were qualitatively different from those involving friends, relatives, and acquaintances. We anticipated that prosecutors would be most likely to file charges in cases involving strangers, and least likely to do so in cases involving intimate partners. Our finding that the likelihood of charging did not vary by relationship type suggests that prosecutors in these two jurisdictions do not automatically classify cases involving nonstrangers as "junk or garbage cases." Yet the fact that prosecutors were equally likely to file charges in all three types of cases does not mean that they used the same criteria to determine the likelihood of conviction for sexual assault cases in each category. Although the presence of physical evidence connecting the suspect to the crime exerted a strong and statistically significant effect on charging in all three types of cases, the effect was more pronounced in cases involving strangers than in cases involving acquaintances or relatives. Moreover, the other predictors of charging were not invariant. In line with our hypothesis, in cases involving strangers, the decision to charge or not was determined primarily by legally relevant factors; in such cases, the odds of charging were increased if there was physical evidence and if the suspect used a gun or a knife. Our findings regarding the influence of victim characteristics also are consistent with our hypothesis. With the exception of the victim's race, which influenced the decision to charge or not if the victim and the suspect were strangers, victim characteristics affected charging only in cases involving nonstrangers. In cases involving friends, acquaintances, and relatives, prosecutors were significantly less likely to file charges if the victim engaged in risktaking behavior at the time of the incident or if there were questions about her reputation or character. If the victim and the suspect were (or had been) intimate partners, prosecutors were less

678

PROSECUTINGSEXUALASSAULT

likely to file charges if the victim engaged in risky behavior or physically resisted the suspect; they were more likely to file charges if the victim was injured. These results add credence to the assertions of Estrich (1987) and Bryden and Lengnick (1997), who contend that victim characteristics are more likely to influence the outcome of sexual assault cases if the victim and the suspect are not strangers. As they note, the suspect in a case involving complete strangers cannot logically argue that the victim consented. Thus, in assessing the likelihood of conviction in this type of case, prosecutors need not worry that evidence concerning the victim's character or behavior at the time of the incident will dispose jurors to conclude that she fabricated the assault, was biased against the defendant, or had a motive for lying. In other words, because the defendant in a case involving strangers is most likely to argue that he was misidentified, convictability hinges largely on the prosecutor's ability to prove that he was not. In sexual cases involving nonstrangers, on the other hand, the defendant is likely to claim that the incident was fabricated or that the victim consented. Thus, in such cases, the prosecutor's evaluation of convictability may depend on the presence or absence of evidence that will cause potential jurors to blame the victim or question her credibility. Our conclusion that victim characteristics affect the decision to charge or not only in sexual assault cases involving acquaintances, relatives, and intimates is strengthened by the fact that evidence concerning the victim's character and risk-taking behavior at the time of the incident actually was more prevalent in cases involving strangers than in cases involving nonstrangers. As shown in Table 1, we found evidence of risky behavior by the victim in 48 percent of the stranger cases, but in only 39 percent of the acquaintance cases and 33 percent of the intimate partner cases. Similarly, questions about the victim's moral character surfaced more frequently in cases involving strangers (41%) than in cases involving acquaintances (29%). Thus our finding that these negative victim characteristics did not affect charging when the victim and the suspect were strangers cannot be attributed to their absence in such cases. Rather, a substantial number of these cases included damaging evidence regarding the victim's character and behavior at the time of the incident, b u t such evidence did not affect the prosecutor's decision to charge or not. Our conclusion is strengthened further by evidence on the types of risk-taking behavior and the types of moral character questions that we found in each of the three categories. Our measures of the victim's risky behavior and moral character reflected the presence

SPOHN AND HOLLERAN

679

or absence of nine types of victim behavior and eight character issues. With respect to the risk-taking behaviors, very similar proportions of the victims in each group either accompanied the suspect to his residence or invited him to their homes: the figures were 21.2 percent for cases involving strangers, 21.6 percent for cases involving acquaintances or relatives, and 20.7 percent for cases involving intimate partners. Also similar were the proportions of victims in each group who had been using alcohol or drugs before the assault: 18.9 percent in cases involving intimates, 20.4 percent in cases involving strangers, and 26.9 percent in cases involving acquaintances. Victims who were assaulted by strangers, on the other hand, were substantially more likely than those who were assaulted by nonstrangers to be walking alone late at night, hitchhiking, or in a bar alone; 34 (30%) of the women who were attacked by strangers engaged in these types of risky behavior, compared with only eight (2%) of the nonstrangers' victims. Women assaulted by strangers, in other words, were more likely than those assaulted by nonstrangers to engage in precisely the types of risky behaviors that might lead jurors to conclude that they precipitated or were to blame for the sexual assault. When we examined the frequency distributions tbr the eight items constituting the moral character variable, we found, for each type of relationship, very similar proportions of victims who had engaged in premarital or extramarital sexual intercourse with someone other than the suspect or who had a history of drug or alcohol use. The figures for prior sexual history were 8.0 percent (strangers), 10.3 percent (intimates), and 10.8 percent (acquaintances); for a history of drug or alcohol use, 13.3 percent (strangers), 14.7 percent (intimates), and 15.8 percent (acquaintances). In contrast, evidence suggesting that the victim was a prostitute or that she worked as a stripper, exotic dancer, or masseuse was found in significantly more of the cases involving strangers (20%) than of cases involving nonstrangers (less than 3%). Again, these m a y be precisely the characteristics that incline jurors to blame the victim or to question her credibility. Our conclusion is supported further by the results of our analysis using the more refined measures of the victim's behavior and moral character. In cases involving complete strangers, charging was affected by n o n e of these more refined measures, including the two types of behavior or character evidence that might be viewed as most damaging (the victim was walking alone late at night or was in a bar by herself; the victim was a prostitute, stripper, or exotic dancer).

680

PROSECUTING SEXUALASSAULT

Considered together, these results provide compelling evidence that prosecutors' charging decisions in sexual assault cases involving strangers are not affected by the victim's character or behavior at the time of the incident. Even though the behavior of women who were assaulted by strangers might be viewed as more risky than the behavior of women who were assaulted by nonstrangers, and even though the evidence concerning the moral character of women who were assaulted by strangers might be viewed as more damaging than the evidence concerning the moral character of women who were assaulted by nonstrangers, neither of these victim characteristics affected the likelihood of charging when the victim and the suspect were strangers. We cannot conclude, however, that victim characteristics played no role in charging decisions in cases involving strangers; in these cases, prosecutors were more likely to file charges if the victim was white. This finding is both contrary to our hypothesis and inconsistent with the results of previous research; the research demonstrates that the effect of legally irrelevant suspect and victim characteristics is confined primarily to less serious cases, in which decision makers have more discretion in determining the appropriate outcome (see Kalven and Zeisel 1966; Spohn and Cederblom 1991). We speculated that this race-of-victim effect might be attributed to differences in the racial makeup of the suspect/victim dyad. Although most of the assaults in each category were intraracial, a larger proportion of the stranger cases (21.2%) than of cases involving acquaintances (5.4%) or intimates (7.8%) were sexual assaults by a black m a n on a white woman. Because prior research has demonstrated that black-on-white sexual assaults are treated more harshly (LaFree 1989; Spohn and Spears 1996; Walsh 1987) and because these cases occurred more frequently in our group of stranger rapes, in other words, the effect we uncovered actually might be due to interaction between victim's race and suspect's race. To test this possibility, we reran the analysis of stranger rapes, substituting two d u m m y variables--black suspect/black victim and white suspect/white victim--for the original victim's race and suspect's race variables. Black suspect/white victim was the reference category (there were no cases involving white suspects and black victims). We found that neither blacks who assaulted other blacks (b = -1.103; SE = .76) nor whites who assaulted other whites (b = 1.12; SE = .92) were less likely to be charged than blacks who assaulted whites. In these two jurisdictions, then, prosecutors were more likely to file charges against men who assaulted white women who were strangers to them than men who assaulted black women

SPOHN AND HOLLERAN

681

who were strangers to them. The fact that this effect did not appear in the other two relationship categories suggests that prosecutors view aggravated sexual assaults on white women as particularly serious. We also wish to comment on our findings regarding the factors that affect charging decisions in cases involving intimate partners. Although we were not surprised that risk-taking behavior by the victim influenced charging decisions in these types of cases, we were somewhat surprised to find that injury to the victim had a positive effect on charging, whereas resistance by the victim had a negative effect. These results suggest that prosecutors believe that an injury may counteract jurors' skepticism about a woman's allegation of sexual assault by her boyfriend, lover, or spouse. In such cases, which are inherently ambiguous, the victim's credibility may be particularly important. The victim may be deemed more believable if she has injuries that can corroborate her assertion that the sexual intercourse was nonconsensual. In view of this possibility, it seems inconsistent that prosecutors would be less likely to file charges against women who physically resisted their attacker. Although we can only speculate, it may be that physical resistance by the victim may be regarded as evidence of "mutual combat" between the suspect and the victim; such resistance, coupled with the prior consensual sexual relationship between the victim and the suspect, may cause the prosecutor to doubt the validity of the complaint of sexual assault. A final comment concerns our finding that the relationship between the victim and the suspect did not affect the likelihood of charging: prosecutors were no less likely to file charges if the victim and suspect were acquaintances, relatives, or intimate partners than if they were complete strangers. This finding contradicts assertions that sexual assaults involving acquaintances are not regarded as "real rapes', (Estrich 1987) and that women victimized by these crimes are not regarded as "genuine victims" (LaFree 1989). We suggest that this result may be attributed at least in part to the rape law reforms enacted during the past three decades. Beginning in the mid-1970s, most states, including the two states represented here, adopted reforms designed to shift the focus in a rape case from the victim's character and behavior to the offender's behavior (see Estrich 1987; Spohn and H o m e y 1992). The most common reforms were changes in the definition of rape, elimination of the resistance and corroboration requirements, and enactment of rape shield laws designed to preclude the use of testimony on the victim's sexual history. As Spohn and H o m e y (1992) note, these reforms were designed primarily to increase the odds of successful

682

PROSECUTING SEXUALASSAULT

prosecution in cases in which the victim and the suspect were acquainted and the suspect claimed that the victim consented. Although research evaluating the impact of the rape law reforms generally concludes that the statutory changes did not produce the widespread instrumental changes anticipated by reformers, evidence suggests that the reforms encouraged arrest and prosecution in "borderline cases." Spohn and Horney (1996:874), for example, found that the proportion of simple rape cases bound over for trial in Detroit increased from 17.6 percent in the pre-reform period to 24.4 percent in the post-reform period. Consistent with this fact is our finding that prosecutors were no more likely to screen out cases involving acquaintances and intimate partners than cases involving strangers.

CONCLUSION

The results of our study of charging decisions in sexual assault cases are consistent with prexdous research demonstrating that prosecutors attempt to avoid uncertainty by filing charges in cases where the likelihood of conviction is strong and by rejecting charges when conviction seems unlikely. In all three types of cases, and particularly in cases in which the victim and the suspect are strangers, the prosecutor is more likely to file charges if there is physical evidence that can corroborate the victim's assertion that a sexual assault occurred or can prove that the accused in fact is the person who committed the crime. Our results, however, also reveal "extralegal sources of uncertainty" (Albonetti 1987:311). Although the relationship between the victim and the suspect did not affect the likelihood of charging, victim characteristics influenced the decision to charge or not in cases in which the victim and the suspect were acquaintances, relatives, or intimate partners. In such cases, prosecutors' anticipation of a consent defense and their downstream orientation toward judges and juries apparently lead them to scrutinize the victim's character and behavior more carefully. Evidence challenging the victim's credibility or fostering a belief that she was not entirely blameless increases uncertainty about the outcome of the case and thus reduces the odds of prosecution. Despite the rape law reforms promulgated during the past three decades, victim characteristics continue to influence charging decisions in at least some types of sexual assault cases.

SPOHN AND HOLLERAN

683

REFERENCES

Abadinsky, H. 1988. Law and Justice. Chicago, IL: Nelson-Hall. Albonetti, C. 1986. "Criminality, Prosecutorial Screening, and Uncertainty: Toward a Theory of Discretionary Decision Making in Felony Case Processings." Criminology 24:623-44. --. 1987. "Prosecutorial Discretion: The Effects of Uncertainty." Law and Society Review 21:291-313. Battelle Memorial Institute Law and Justice Study Center. 1977. Forcible Rape: A National Survey of the Response by Prosecutors. Washington, DC: U.S. Government Printing Office. Belsley, D.A., E. Kuh, and R.E. Welsch. 1980. Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. New York: Wiley. Bienin, L., N. Weiner, D. Denno, P. Allison, and D. Mills. 1988. '~TheReimposition of Capital Punishment in New Jersey: The Rote of Prosecutorial Discretion." Rutgers Law Review 48:386-98. Black, D. 1976. The Behavior of Law. New York: Academic Press. Blumstein, A., J. Cohen, J. Roth, and C. Visher. 1983. Research on Sentencing: The Search for Reform. Washington, DC: National Academy Press. Bohmer, C. 1974. "Judicial Attitudes Toward Rape Victims." Judicature 57:303-07. Bryden, D.P. and S. Lengnick. 1997. "Rape in the Criminal Justice System." Journal of Criminal Law and Criminology 87:1194-1384. Davis, K. 1969. Discretionary Justice. Baton Rouge, LA: Louisiana State University Press. Decker, S.H. 1993. "Exploring Victim-Offender Relationships in Homicide: The Role of Individual and Event Characteristics." Justice Quarterly 10:585-612. Eliason, S.R. 1993. Maxiumum Likelihood Estimation: Logic and Practice. Newbury Park, CA: Sage. Estrich, S. 1987. Real Rape. Cambridge, MA: Harvard University Press. Feeney, F., F. Dill, and A. Weir. 1983. Arrests Without Conviction: How Often They Occur and Why. Washington, DC: Department of Justice. Feild, H. and L. Bienen. 1980. Jurors and Rape: A Study in Psychology and Law. Lexington, MA: Lexington Books. Feldman-Summers, S. and K. Lindner. 1976. "Perceptions of Victims and Defendants in Criminal Assault Cases." Criminal Justice and Behavior 3:135-50. Frohmann, L. 1991. "Discrediting Victims' Allegations of Sexual Assault: Prosetutorial Accounts of Case Rejections." Social Problems 38:213-26. --. 1997. "Convietability and Discordant Locales: Reproducing Race, Class, and Gender Ideologies in Prosecutorial Decisionmaking." Law and Society Review 31: 531-55. Gottfl'edson, M.R. and D.M. Gottfredson. 1988. Decision.Making in Criminal Justice: Toward the Rational Exercise of Discretion. 2nd ed. New York: Plenum. Hepperle, W. 1985. 'WComenVictims in the Criminal Justice System." In The Changing Role of Women in the Criminal Justice System, edited by I. Moyer. Prospect Heights, IL: Waveland. Horney, J. and C. Spohn. 1996. "The Influence of Blame and Believability Factors on the Processing of Simple Versus Aggravated Rape Cases." Criminology 34:13562. Jacoby, J., L. Mellon, E. Rattedge, and S. Turner. 1982. Prosecutorial Decision Making: A National Study. Washington, DC: U.S. Department of Justice. Kalven, H. and H. Zeisel. 1966. The American Jury. Boston, MA: Little, Brown. Keil, T. and G. Vito. 1989. "Race, Homicide, Severity, and Application of the Death Penalty: A Consideration of the Barnett Scale." Criminology 27:511-33. Kerstetter, W. 1990. "Gateway to Justice: Police and Prosecutoriat Response to Sexual Assaults Against Women." Criminology 81:267-313. Kingsnorth, R., J. Lopez, J. Wentworth, and D. Cummings. 1998. "Adult Sexual Assault: The Role of Racial/Ethnic Composition in Prosecution and Sentencing." Journal of Criminal Justice 26:359-71. Kingsnorth, R.F, R.C. Macintosh, and J. Wentworth. 1999. "Sexual Assault: The Role of Prior Relationship and Victim Characteristics in Case Processing." Justice Quarterly 16:275-302.

684

PROSECUTING SEXUAL ASSAULT

LaFree, G. 1980. "The Effect of Sexual Stratification by Race on Official Reactions to Rape." American Sociological Review 45:842-54. --. 1981. "Official Reactions to Social Problems: Police Decisions in Sexual Assault Cases." Social Problems 28:582-94. · 1989. Rape and Criminal Justice: The Social Construction of SexualAssault. Belmont, CA: Wadsworth. Liao, T.F. 1994. Interpreting Probability Models: Logit, Probit, and Other Generalized Models. Thousand Oaks, CA: Sage. Loh, W. 1980. "The Impact of Common Law and Reform Rape Statutes on Prosecution: An Empirical Study." Washington Law Review 55:543-625. Long, J.S. 1997. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage. Mather, L. 1979. Plea Bargaining or Trial? Lexington, MA: Heath. McCahill, T., L. Meyer, and A. Fischman. 1979. The Aftermath of Rape. Lexington, MA: Lexington Books. Miller, F. 1969. Prosecution: The Decision to Charge a Suspect With a Crime. Boston, MA: Little, Brown. Myers, M. 1982. "Common Law in Action: The Prosecution of Felonies and Misdemeanors." Sociological Inquiry 52:1-15· Nagel, H. and J. Hagan. 1983. "Gender and Crime: Offense Patterns and Criminal Court Sanctions." Pp. 91-144 in Crime and Justice: An Annual Review of Research, vol. 4, edited by M. Tonry and N. Morris. Chicago, IL: University of Chicago Press. Nardulli, P.F., J. Eisenstein, and R.B. Flemming. 1988. The Tenor of Justice: Criminal Courts and the Guilty Plea Process. Urbana, IL: University of Illinois Press· Nelson, S. and M. Amir. 1975. "The Hitchhike Victim of Rape: A Research Report." Pp. 103-27 in Victimology: A New Focus, voL 3, edited by I. Drapkin and E. Viano. Lexington, MA: Lexington Books. Neubauer, D. 1974. "After the Arrest: The Charging Decision in Prairie City." Law and Society Review 8:475-517. · 1988. America's Courts and the Criminal Justice System. Pacific Grove, CA: Brooks/Cole. Paternoster, R. 1984. "Prosecutorial Discretion in Requesting the Death Penalty: A Case of Victim-Based Racial Discrimination." Law and Society Review 18:43778. Paternoster, R., R. Brame, P. Mazerolle, and A. Piquero. 1998. "Using the Correct Statistical Test tbr the Equality of Regression Coefficients." Criminology 36:85966. "Police Reopen Thousands of Dumped Cases." 2000. The Philadelphia Inquirer, January 16. Posted at http'J/www.pbrpC~hilly.infi.net. Radelet, M. and G. Pierce. 1985. "Race and Prosecutorial Discretion in Homicide Cases." Law and Society Review 19:587-621. Rauma, D. 1984. "Going for the Gold: Prosecutorial Decision Making in Cases of Wife Assault." Social Science Research 13:321-51. Reskin, B. and C. Visher. 1986. "The Impacts of Evidence and Extralegal Factors in Jurors' Decisions." Law and Society Review 20:423-38. Schmidt, J. and E. Steury. 1989. "Prosecutorial Discretion in Filing Charges in Domestic Violence Cases." Criminology 27:487-510. Sebba, L. and S. Cahan. 1973. "Sex Offenses: The Genuine and the Doubted Victim." Pp. 29-46 in Victimology: A New Focus, vol. 5, edited by I. Drapkin and E. Viano. Lexington, MA: Lexington Books. Silberman, C. 1978. Criminal Violence, Criminal Justice. New York: Vintage. Simon, L.M.J. 1996. "Legal Treatment of the Victim-Offender Relationship in Crimes of Violence·" Journal of Interpersonal Violence 11:94-106. Spears, J.W. and C.C. Spohn. 1997. "The Effect of Evidence Factors and Victim Characteristics on Prosecutors' Charging Decisions in Sexual Assault Cases." Justice Quarterly 14:501-24. Spohn, C. and J. Cederblom. 1991. "Race and Disparities in Sentencing: A Test of the Liberation Hypothesis." Justice Quarterly 8:305-27. Spohn, C., J. Gruhl, and S. Welch. 1987. "The Impact of the Ethnicity and Gender of Defendants on the Decision to Reject or Dismiss Felony Charges·" Criminology 25:175-91. Spohn, C. and J. Homey. 1992. Rape Law Reform: A Grassroots Revolution and Its Impact. New York: Plenum.

SPOHN AND HOLLERAN --.

685

1993. "Rape Law Reform and the Effect of Victim Characteristics on Case Processing." Journal of Quantitative Criminology 9:383-409. --. 1996. "The Impact of Rape Law Reform on the Processing of Simple and Aggravated Rape Cases." Journal of Criminal Law and Criminology 86:861-84. Spohn, C. and J. Spears. 1996. ~ h e Effect of Offender and Victim Characteristics on Sexual Assault Case Processing Decisions." Justice Quarterly 13:649-79. Stanko, E. 1988. "The Impact of Victim Assessment on Prosecutor's Screening Decisions: The Case of the New York County District Attorney's Office." In Criminal Justice: Law and Politics, edited by G. Cole. Pacific Grove, CA: Brooks/Cole. Walsh, A. 1987. "The Sexual Stratification Hypothesis and Sexual Assault in Light of the Changing Conceptions of Race." Criminology 15:153-73. Weninger, R. 1978. "Factors Affecting the Prosecution of Rape: A Case Study of Travis County, Texas." Virginia Law Review 64:357-97. Williams, K. 1978. The Rote of the Victim in the Prosecution of Violent Offenses. Washington, DC: U.S. Government Printing Office. '~WomenVictimized Twice in Police Game of Numbers." 1999. The Philadelphia Inquirer, October 17. Posted at http://~-w.pbrpCCphilly.infi.net.

CASE CITED

Bordenkircher v. Hayes 434 U.S. 357, 364 (1978)

OO O'~

A p p e n d i x Table A1.

x2 x3 x4 x5 x6 x7 x8 x9 xl0 xll

P e a r s o n C o r r e l a t i o n s o f All V a r i a b l e s I n c l u d e d i n A n a l y s e s for C a s e s I n v o l v i n g S t r a n g e r s

x12 x13

©

5O

( N = 113)

c~

Y

xl

Prosecutor Charged (Y) White Victim (xl) Victim's Age (x2) Risk Taking (x3) Moral Character (x4) Victim Resisted (x5) Report in Hour (x6) White Suspect (xT) Suspect's Age (x8) Prior Felony (x9) Gun/Knife (xl0) Victim Injured (xll) Physical Evidence (x12) Witnesses (x13) Philadelphia (x14) .43* -.07 -,18" .i0 .14 .01 -.12 ,14 .01 -.02 .19" .02 -.21" .03 .05 .02 .05 .00 .21" .08 .21" .03 .12 -.09 -.05 -.19" .05 -.19" .19" -.05 -.08 -.06 .04 .20* .04 .10 .18" -,03 .09 .11 -.19" .06 -.09 .18" -.13 .47* -.14 .04 .01 .14 .02 .04 ,10 .16" .14 -,07

.26* .13 -.15 -.10 -.06 ,23* .09 .04 .16" .35* .13 .39* .09 -.23*

.04 -.06 -.14 .03 -.05 .56" .14 .23* -.08 -,03 .01 .04 -.34*

.08 -.17" .03 .20* -.07 .27* .28* -.07 .23* .11 .07 .00

B

.00 .35* .05 .00

.15 .17" -.27*

-.06 .04

-.02

* p < .05

A p p e n d i x Table A2.

x2 x3 x4 x5 x6 x7 x8 x9 xl0 xll

P e a r s o n C o r r e l a t i o n s o f All V a r i a b l e s I n c l u d e d in A n a l y s e s for Cases I n v o l v i n g A c q u a i n t a n c e s (N = 297)

x12 x13

Y

xl

© .15" .05 .04 .00 .00 .18" .22* -.05 .12" .22* -.08 -.01 .11" .12" Z .09 ©

Prosecutor Charged (Y) White Victim (xl) Victim's Age (x2) Risk Taking (x3) Moral Character (x4) Victim Resisted (x5) Report in Hour (x6) White Suspect (x7) Suspect's Age (x8) Prior Felony (x9) Gun/Knife (xl0) Victim Injured ( x l l ) Physical Evidence (x12) Witnesses (x13) Philadelphia (x14) .24* .17" .04 .12" -.11 -.13" -.04 -.02 .12" .08 -.16" .10" .01 -.01 .01 -.04 .05 .08 .13" .13" -.04 .13" .02 -.13" -.01 -.06 .12" .09 .05 .11" -.18" -.05 .08 .19" .14" .18" .05 .08 .09 -.16" -.16" .00 -.17" -.08 -.40* .23* .05 .04 -.17" .01 -.02

-.09 -.03 -.17" -.15" -.14" .04 -.05 .06 .18" .07 .00 .07 .03 -.01

.04 .24* .06 .07 -.12 .77* .06 -.10 -.20* .00 -.10" -.02 -.37*

.03 .13" .21" .21" -.02 .30* .12" .27* .22* .11" -.06 .11"

*p < .05

Qo

A p p e n d i x T a b l e A3.

x2 x3 x4 x5 x6 x7 x8 x9 xl0 xll

©

C~

P e a r s o n C o r r e l a t i o n s o f All V a r i a b l e s I n c l u d e d i n A n a l y s e s f o r C a s e s I n v o l v i n g I n t i m a t e P a r t n e r s ( N = 116)

x12 x13

Y

xl

Prosecutor Charged (Y) White Victim (xl) Victim's Age (x2) Risk Taking (x3) Moral Character (x4) Victim Resisted (x5) Report in Hour (x6) White Suspect (x7) Suspect's Age (x8) Prior Felony (x9) Gun/I~life (xl0) Victim Injured (xll) Physical Evidence (x12) Witnesses (x13) Philadelphia (x14) -.05 .15 -.02 .14 .20* .05 -.12 .05 .05 .03 -.12 -.07 -.06 .14 -.28* .01 .09 .15 -.07 .16" -.06 .22* .01 .12 .03 -.03 .05 -.02 .07 .09 .10 .13 .07 .08 .21" .14 -.04 .09 .04 -.04 -.02 .22* .05 .18" -.18" .13 -.15 .19" -.04 -.12 .06 .13 .18" .05 .00 -.11

.03 -.01 -.32* .13 -.20 .18" -.04 -.05 .09 .13 .24* .28* .02 -.12

O~

.08 .15 .05 -.09 .11 .81" .04 -.05 -.06 .21" .12 .09 -.26*

.22* -.31" .17" .14 .05 .74* .01 -.08 .25* .11 -.10 .10

.23* .12 .06 -.20*

.16" .01 -.13

-.04 -.07

-.10

* p < .05

Information

651_1.tif

38 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

1125198

You might also be interested in

BETA
CZECHCD.doc
FactSheet
Public Access to Court Records