Abstract
Victims who express less emotion in response to a crime are perceived as less deserving, less sympathetic, and they have less punishment assigned to the offender who committed the crime. This study considers the extent to which emotion norms underlie perceptions of victims who testify. Two studies investigate the circumstances in which emotional reactions to a crime are seen as “unusual” and whether a more general emotion norm underlies responses to victim testimony. We test a “victim-role” norm against a “proportionality” norm by crossing the severity of victim's emotional response (severe or mild) with the seriousness of a crime (serious or less serious). Results across two studies lend greater support to the notion that people expect victims to match the intensity of their emotional response to the seriousness of the event (i.e., a proportionality rule), although we also find instances in which expectations of the victim are not strong. Gender of the victim exhibited small and contingent effects. We discuss the relevance of emotion norms to legal settings.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Since the 1980s, it has become routine for the criminal justice system to involve crime victims in the sentencing process. This involvement sometimes allows victims the chance to provide information about how the crime has affected them (so-called “victim-impact” evidence, or VIE). VIE statements take many forms (Greene, 1999). Apart from objective losses (e.g., financial damage), victims are also encouraged to provide a more subjective account of a crime's effect, in particular, what emotional consequences have followed the event. The emotional toll of crime can be high (Cohen, 1988), and victims report a therapeutic value to having voice in sentencing proceedings (Erez, 1999; Roberts & Erez, 2004). Nevertheless, the practice of allowing this type of voice in sentencing continues to be criticized on the grounds that it might produce overly harsh sentences (Arrigo & Williams, 2003; Bandes, 1996).
In this paper, our focus is not on whether emotion-based victim impact statements produce unjust sentences. Instead we examine whether observers make normative judgments about the emotional content of victim statements. In particular, we consider normative expectations surrounding statements in non-capital criminal cases where it is the victim (rather than a family member or other person) who discusses the emotional impact of the crime. We argue that observers’ reactions to victims’ emotional displays can illuminate our understanding of emotion norms. Norms and rules governing appropriate emotion expression derive from a multitude of sources, including socialization via national or ethnic culture, parents, peer groups, or workplaces (for reviews, see (Gordon, 1990; Thoits, 1990)). As we discuss later, current empirical understanding of the specific norms governing a victim's emotional reaction is quite limited.
In the next section, we review existing experimental, laboratory-based studies of victim impact statements in non-capital cases, and we describe why available evidence suggests the importance of examining observers’ normative expectations about victims. We then present two studies designed to isolate two different theories about emotion expectations. Finally, we discuss the implications of these results for understanding people's expectations surrounding emotional experience and further work on emotion norms in the law.
How do people respond to a victim's emotional expression?
In a typical experimental study of victim impact statements, researchers present participants with a crime and vary whether the victim describes the events (Tsoudis & Smith-Lovin, 1998) or its aftermath (Hills & Thomson, 1999; Nadler & Rose, 2003) in a strongly emotional manner or in a more mild manner (e.g., the victim has symptoms of depression and PTSD versus the victim has no long-term effects and continues to undertake prior activities; (Hills & Thomson, 1999)). These studies find that a victim's emotional expression influences perceptions of the victim. Observers view crime victims who respond with mild emotions less positively. For example, Nadler and Rose (Nadler & Rose, 2003) not only found that observers imposed less punishment when a robbery victim's emotional reaction was mild rather than severe, the victim who displayed a mild emotional reaction also generated less sympathy.
Tsoudis and Smith-Lovin (Tsoudis & Smith-Lovin, 1998) provide the strongest evidence that a victim's emotional displays can alter perceptions of that victim. These researchers draw upon a sociological account of emotion, entitled affect control theory, which argues that strong emotions provide information about a person's identity in a role because they signal the extent to which a given transient role (e.g., a victim giving testimony) is or is not consistent with the fundamental sentiment we might feel toward that person (Heise, 1989; Smith-Lovin & Heise, 1988). A crime victim who displays a great deal of negative emotion signals to the observers that being a victim of an offense is inconsistent with that person's more fundamental identity, where more emotion indicates greater inconsistency. That is, the person emoting is not usually a “victim-like” person. In contrast, a muted emotional response signals that the transient role of being a victim is consistent with the person's fundamental identity. This is not “victim-blaming” (Lerner, 1980), but rather reactions to how a person occupies the victim-role. Presumably, we place less value on people who are comfortable with their stigmatized state—in this case, the state of being a victim. A structural equation model largely supported Tsoudis and Smith-Lovin's hypotheses: Affect expressions altered ratings of how good a person the victim seemed to be; the latter, in turn, affected sentences assigned (higher punishment for the offense done to the more-upset victim).
The fact that emotion displays altered observers’ perceptions of victims suggests the existence of underlying normative expectations for how victims should behave and should display emotion, i.e., emotion norms. Although research on emotion norms—including “display rules” (Ekman & Friesen, 1969) and “feeling rules” (Hochschild, 1983)—has received some attention, many studies are qualitative and devoted largely to establishing norms about what particular type of emotion (e.g., anger versus sadness, negative emotion versus happiness) should be expressed in a given circumstance. Hochschild (Hochschild, 1983), for instance, observed and interviewed flight attendants and showed how demands of the job require strict regulation of emotion displays (e.g., being cheerful for customers), and that workers attempt to change the feelings they actually experience in order to conform to these norms. Other experimental work confirms that participants exhibit some consensus about what people are likely to feel in various situations (Heise & Calhan, 1995), and some studies find that emotion norms and expectations change as a function of status of the person emoting (e.g., Tiedens, 2001).
Existing experimental victim impact studies, however, do not represent instances of victims’ demonstrating the wrong emotion. Rather they suggest that victims are stigmatized for expressing an insufficiently strong emotion—an emotion of too little intensity or too short a duration. Normative expectations about the time and intensity of emotions is still not well-understood in the sociological or psychological literatures (see e.g., Thoits, 1989). Recent literature on affective forecasting (Wilson & Gilbert, 2003) has demonstrated that people overestimate how long their own emotional states are expected to last and how intense they will be; it is unclear, however, the extent to which beliefs about how long emotions should last (e.g., when it is appropriate to cease feeling sorrow over a relative's death) contribute to these estimates (although see Igou, 2004).
Further, the law and social science literature suggests that our expectations for what victims are “supposed” to do may vary with different types of victims. In the civil justice arena, for example, it has been suggested that jurors can be critical of accident victims who seem to be “over-reacting” or making too much of their “pain and suffering”; some jurors see the latter as a normal part of life (e.g., Hans, 2000; Vidmar, 1995). This complexity surrounding the precise nature of expectations for emotions in legal contexts indicates that further work should seek to specify reactions to crime victims’ emotions and how observers’ expectations can be violated.
Observers’ expectations about victim emotion: Two possible norms
We explore two distinguishable ways in which people might view a crime victim as emoting and coping inappropriately. As noted above, affect control theory suggests that the trouble lies in the fact that the mild-mannered victims appears to “comfortably occupy” the role of victim, which is a stigmatized state (Tsoudis & Smith-Lovin, 1998). The implication of this perspective is that whenever people are wronged—severely or even mildly—they should express strong emotion in order to distinguish themselves from the type of people who are, fundamentally, victims. Strong emotion would communicate: “This should not have happened to me!” We refer to this as the “victim-role” rule: that is, there is a consistent norm—regardless of severity or type of crime—for how people are to react to being crime victims.
The available data could also be explained in an entirely different way. Rather than a norm that suggests that all crime victims should powerfully emote, perhaps a proportionality rule underlies the norm: The intensity and duration of emotion in response to a crime should be proportional to the seriousness of the event. In this view, the crime victims in the existing studies—all victims of serious crimes (e.g., rapes, robberies, assaults)—have failed to muster a corresponding amount of strong emotion given the harm done to them. What is distinguishable about this perspective is that, in contrast to the victim-role rule, one can imagine a situation in which the violation is the expression of too much emotion—i.e., if the crime was not very harmful or serious. In the proportionality perspective, one can over-react as well as under-react; this seems less evident in the predictions of affect control theory.
Existing research outside of the legal context suggests people might use a proportionality rule to judge the appropriateness of emotions. Drawing upon interviews, vignette-based experiments, and content analysis of cultural expressions of sympathy (e.g., greeting cards), Clark (1997) argues that we strive for proportionality in managing our sympathetic feelings towards others. In our social relations, we maintain “accounts” with others in which both parties make “withdrawals” (asking for sympathy) and “deposits” (giving sympathy). Clark suggests that we seek to keep these accounts in balance. We are dissatisfied with those who draw upon the account too infrequently (i.e., when we are prevented from extending sympathy to others when we want to), as well as with those who overdraw (people who tax our sympathy by asking “too much” of us). In other words, we seek a good fit between the perceived severity of the circumstances and the emotions we both want to express and want others to express. Of course, the more general notion that people seek balance in social relations and social perceptions is a core social psychological concept (Heider, 1958).
The present studies were designed to distinguish the victim-role perspective from the proportionality perspective. Doing so requires the use of a less serious crime and corresponding severe or mild emotional responses to it. We presented participants with vignettes that represent, in Study 1, serious and less serious crimes, and, in Study 2, the same crime with serious or less serious harm. We then crossed these crimes with two types of emotional reactions, one very strong and one more mild. If normative expectations follow the victim-role rule, then we should observe a main effect for the perceived appropriateness of the victim's emotion: Regardless of the seriousness of the crime or injury, a strong emotional reaction should be viewed as more appropriate, and the victim seen more positively. In contrast, if the proportionality rule is correct, we should observe an interaction: Emotional reactions, and the victims themselves, should be judged more positively when the victims’ expressions match the crime's seriousness (a strong reaction for the serious crime, a more mild reaction for the less serious event), compared to when there is a mismatch between victim emotion and crime seriousness.
In both the studies, we varied gender of the victim. Our intuitive expectation is that mild responses from men are viewed as potentially less unusual than the same response from a woman; however, it bears noting that stereotypes about gender differences in emotional reactions are often overly general (Brody & Hall, 2000; Clark, 1997). One study, for instance, found that a confederate male was liked more when he cried openly during a movie, whereas a less-emotional female confederate was preferred (Labott, Martin, Eason, & Berkey, 1991). Thus, we do not make strong predictions about the effect of the victim's gender on perceptions.
Study 1: Emotional reactions to armed robbery and pick-pocketing
Method
Participants and procedure
We sought data from people who had more life experience with stress and coping than may be true of the typical undergraduate population. Participants were 118 University of Texas employees and were paid $10, as well as the opportunity to enter a lottery for a larger cash prize ($250).Footnote 1
Study 1 made use of a 2 (crime: serious or less serious)×2 (emotional reaction: severe or mild)×2 (victim gender) design. Participants read a two-page vignette about a crime and a sentencing hearing and then answered a series of questions regarding perceptions of the offense, views of the victim and his/her reaction, self-rated emotions, and appropriate punishment; we also included other filler questions on the offender (we do not report results for these items; a full copy of the inventory is available from the first author upon request). During the session, participants also filled out a questionnaire for a separate study on responsibility, which made the entire session last about 40 min. The current study randomly appeared either first or second in the questionnaire packet. We observed small and inconsistent effects of this ordering, but we control for ordering in all analyses.
The sample had a mean age of 39 (range=21–71; median=35), was 71% white, 66% female, and was well-educated (all had at least some college). Most (66%) reported having had something stolen from them at some point in time; 30% had been victims of an assault, and 33% had been victims of some other crime. Over half (53%) knew a victim of an assault. A check on random assignment indicated that groups of participants (e.g., men versus women, crime victims versus non-victims) were well-represented in all cells of the design (all chi-square tests of these factors crossed with manipulated factors were non-significant). Further, neither the demographic variables nor experience with crime predicted responses to the variables discussed later, or altered the main findings. Hence, we do not discuss them further.
Stimulus
Participants in the serious crime condition read about a robbery that involved a gun and an explicit threat of harm, and in which the victim was pushed down (but not otherwise physically injured). In the mild crime condition, participants read about a “pick-pocketing” theft in which the thief ran past (and lightly touched) the victim. The victim discussed either dramatic effects from the crime (severe emotional reaction) or no long-term ill effects (mild emotional reaction), as explained in detail later.
All vignettes described the victim as having just finished some shopping and as stopping briefly at a bus stop in a parking lot to tie her/his shoelace. In the serious crime condition, an offender ran up to the victim, put a gun to the victim's head, and threatened to kill the victim if she/he did not hand over money. The victim told the offender that a wallet was in her purse (or, in the male-victim condition, in his backpack); the offender then took the item, shoved the victim down, and ran away. In the less serious crime, an offender ran by the victim and lightly brushed her/him; following this, the victim belatedly realized that the offender had made off with her purse/his backpack. All vignettes reported that the offender was apprehended quickly (after trying to use the victim's credit card) and that he decided to plead guilty. Participants read that they would be reviewing a summary of the sentencing hearing, which due to various court delays, occurred 6 months after the crime.
Each of these crimes was paired with either a severe or mild emotional reaction. In the severe emotional reaction condition, the victim testified that she/he felt the crime was “devastating” and that although she/he used to be really active, the victim did not go out at night any more. The victim is described as sobbing during the testimony, as saying the event still affects her/him, and as expressing the belief that the crime will stay with the victim “forever.” The victim concludes by saying, “What happened is on my mind from the time I wake up in the morning until I go to bed at night. I get bent out of shape over any little thing, and it seems like I’m always feeling down.”
In the mild reaction condition, the victim describes the experience as “troubling” and goes on to say that despite an active schedule, she/he was for a while “a little more alert when I’m out.” The victim is described as speaking in a reflective manner and saying, “I get irritated every once in a while when something reminds me of that night. Other than that, I have mostly forgotten about it, and I go out and have a good time just as much as I always have.”
Measures
Except where noted, all items described were rated on a 7-point scale with endpoints variously labeled none, not at all (1), and extremely, a great deal, very (7).
Manipulation checks.
To ensure that participants perceived differences in the robbery versus pick-pocket crimes, they were asked, “How much physical (bodily) injury did the victim suffer as a result of this crime?” and “How serious was this crime?” The emotional reaction manipulation was checked by asking “How upset” the victim seemed and how much “emotional injury” the victim suffered.
Normality of the reaction.
Respondents rated “How unusual” the victim's reaction seemed. As an estimate of emotion duration, participants were also asked, “For how long do you expect the victim to suffer any emotional injury from this crime?” (1: “no time at all”; 7: “for many years”).
Ratings of the victim.
To assess perceptions of the victim (what affect control theorists term “victim identity”; (Tsoudis & Smith-Lovin, 1998)), participants rated the victim on the following attributes on a 7-point scale with endpoints labeled: (1) bad, awful/good, nice; (2) likeable/repulsive; (3) displeasing/appealing; (4) pleasant/annoying. After reverse-scoring where appropriate, we summed the four to create a composite measure (alpha=.88).
Participants’ emotional reaction.
Participants rated how they felt toward the victim for each of the following four emotions: Anger, disgust, sympathy, and contempt. (We asked parallel items about the defendant.)
Perceptions of the crime and punishment.
To assess effects, if any, of emotion-based harms on perceptions of the offense and appropriate punishment, we utilized a composite measure of the manipulation checks discussed above as a dependent variable (i.e., amount of physical injury and ratings of crime seriousness; alpha=.71). In addition, we asked participants how severe they thought the punishment for this crime should be and to indicate how much total time the defendant should spend in prison, by marking their assessment on a scaled line.Footnote 2
Results
Unless otherwise indicated, all analyses utilized a three-way analysis of variance (ANOVA), which included tests of main effects and all possible interaction terms between crime, reaction, and gender.
Manipulation checks
As intended, the armed robbery was rated as significantly more serious (M=4.65) than the pick-pocketing theft of the purse/backpack (M=3.30), F(1, 109)=36.89, p < .0001 (partial-η 2=.253). The armed robbery was also perceived as resulting in comparatively more physical harm (M=3.05) than the less serious theft (M=1.05), F(1, 109)=60.57, p < .0001 (partial-η 2=.357). In addition, we successfully manipulated the strength of the victim's reaction: Victims in the severe reaction condition were seen as significantly more upset (M=6.39 versus 2.71), F(1, 108)=394.46, p < .0001 (partial-η 2=.785), and as having sustained more emotional injury (M=5.84 versus 3.53). F(1, 109)=88.65, p < .0001 (partial-η 2=.449). Controlling for the manipulated reaction, the victim in the more serious crime also was seen as more emotionally injured than the victim in the less serious crime (armed robbery M=5.10; pick-pocket M=4.27), F(1, 109)=11.84, p < .001 (partial-η 2=.10). No other factors significantly predicted any of these manipulation checks.
Normality of the reaction
Consistent with the notion of a “victim-role rule,” there was a main effect for emotional reaction, F(1, 105)=10.90, p < .01 (partial-η 2=.094), indicating that the more mild reaction was more unusual (M=3.59) than the more severe reaction (M=2.63). However, this effect was qualified by a significant Crime×Reaction interaction, F(1, 105)=16.73, p < .0001 (partial-η 2=.137). Although the “proportionality-rule” predicted a significant interaction between crime type and reaction, the means in Table 1 suggest that the pattern did not wholly support the expectation that the reaction should “match” the crime. Instead, the mild reaction to the robbery was viewed as significantly more unusual than the reactions in the other conditions (all ps < .001, ds=.92–1.38). Contrary to the proportionality perspective, within the pick-pocketing crime, the mild versus severe reactions were seen as equally unusual (i.e., as not very unusual, p=.56). For the robbery vignette, where a mild reaction to a more serious crime was seen as unusual, both the proportionality and the victim-role rules were supported. By contrast, neither rule was supported by the results for the less severe pick-pocketing incident, in which we detected no differences in perceptions of unusualness between the mild and severe victim emotional reactions.
We tested for a Gender×Reaction interaction, which did not reach conventional levels of statistical significance F(1, 105)=3.46, p < .07 (partial-η 2=.032). The form of the interaction suggested that, controlling for crime seriousness, a female victim's mild reaction was perceived as more unusual (M=3.93) than her severe reaction (M=2.43, p < .001, d=.99). The male victim's mild versus severe reactions did not differ significantly (M=3.25 versus 2.83, respectively). The three-way Crime×Reaction×Gender interaction was non-significant, F < 1.0.
With respect to estimates of duration of victim's emotional reaction, there was a significant main effect for severity of reaction, F(1, 108)=47.21, p < .0001 (partial-η 2=.304). Not surprisingly, those describing themselves as severely affected by the crime were expected to be upset for a longer period (M=5.48) than those in the mild reaction condition (M=3.58). Neither the Gender×Reaction nor the Crime×Reaction interaction had an effect on these estimates (both Fs < 1.0), suggesting that people's tendency to make these judgments on the basis of what the victim said (i.e., in essence, “I’m over it” versus “I’ll always feel this way”) was not affected by the crime or gender of the victim. In addition to the reaction effect, there was a Crime×Gender interaction, which did not reach conventional levels of statistical significance, F(1, 108)=3.37, p < .07 (partial-η 2=.030). In particular, the serious crime was expected to have longer-lasting effects on the female victim (M=5.14 for the serious crime versus 3.81 for the less serious crime, p < .001, d=.90); in contrast, irrespective of crime-type, the male's reaction was expected to last about the same amount of time (M=4.75 versus 4.43, ns). This supports a proportionality rule for the female victims, which predicts that severity, and by implication, duration of emotional reaction should match the seriousness of the crime. There was no significant three-way interaction.
Victim perceptions
Perceptions of victim characteristics (e.g., likeability, etc.) were unaffected by any manipulated variable or interaction. There was little variability across conditions in these ratings: Omnibus F(8, 107)=1.39, ns.
Participant emotion
Self-reports of some emotion ratings are highly correlated with one another (Green, Goldman, & Salovey, 1993). To appropriately capture this individual-rater effect and to limit Type I error by conducting fewer individual tests, we utilized a repeated measures analysis (MANOVA) of the set of ratings regarding sympathy, anger, disgust, and contempt felt toward the victim. There was no significant variability across experimental conditions in the participants’ reported feelings (Wilks’ Lambda F-values all <1.91, ns).
Perceptions of punishment and of the crime
Our less serious crime was duly viewed as not warranting much punishment: 79% of the participants in the pick-pocket condition sentenced the offender to 2 years or less; 38% called for no prison time at all (comparable figures for the serious crime were 45 and 10%, respectively). Such “floor” effects likely contributed to our finding that the only factor affecting either punishment or perceptions of the crime was our manipulation of crime seriousness. Sentences and ratings of punishment severity were significantly higher for the armed robbery (M=3.69 years in prison) than for the pick-pocket theft (M=1.50 years), for sentence F(1, 107)=16.96, p < .0001 (partial-η 2=.137), and for punishment severity F(1, 107)=14.42, p < .001 (partial-η 2=.119). The two-item composite of crime seriousness was also affected significantly only by whether participants read about the serious versus the less serious crime, F(1, 109)=69.51, p < .0001 (i.e., consistent with the manipulation check of the individual items; partial-η 2=.389).
Discussion
Our first study provides further support for prior work (Nadler & Rose, 2003; Tsoudis & Smith-Lovin, 1998) which suggested, but did not directly test, the notion that people expect victims to exhibit a great deal of emotionality in response to a crime likewise regarded as serious. Compared to all other conditions, the “mild” reaction to a severe crime was rated as particularly unusual. There is some evidence that expectations about victims’ emotions are gendered: Compared to a severe reaction, a mild reaction was significantly more unusual for a female but not a male victim (bearing in mind that the overall F for the interaction was just shy of conventional significance). Interestingly, unlike the predictions of affect control theory, in which emotional reactions signal information about a victim's identity, our pairings of different reactions to different crimes did not alter perceptions of the victim. We would be interested in replicating this result, for it is possible that these particular vignettes contained too little information for people to make reliable judgments on issues of, for example, the victim's “likeability.” Participant emotion ratings indicated that all victims generated comparable levels of sympathy. This study did not find an association between VIE and punishment, likely due, in part, to the floor effects associated with the less serious crime condition. Even for the serious crime, the null effect is not particularly surprising given that, even when significant, the relationship between VIE and punishment tends to be small (see e.g., (Nadler & Rose, 2003) which produced effect sizes of between .04 and .06).
No matter the effects of emotional reactions on other variables, the pattern of results for “how unusual” the reaction seemed is intriguing indeed. First, a general principle governing the appropriate emotional reaction did not lend itself to easy characterization. We sought to contrast one emotion norm—a “victim-role rule,” in which it is always more appropriate to express strong emotion in response to a wrong—with another emotion norm—a “proportionality-rule,” in which victims should calibrate their reactions to the seriousness of the offense. However, results failed to provide definitive support for either perspective. Despite a significant interaction of crime and emotional reaction, as predicted by a proportionality perspective, across conditions the only cell that differed from all others involved the victim expressing “too little” emotion about a serious crime (i.e., someone who had largely moved past the event and did not get upset while testifying). In the pick-pocketing scenario, participants gave the victim quite a bit of latitude to emote either strongly or weakly, with no differences in perceived unusualness. In short, no victim was perceived as “over”-reacting to an event.
It is worth noting here that our emotional reaction manipulation included information about feelings (the victim felt either devastated or only irritated) as well as behavior (the victim either did not or did continue to go out at night). We chose this multifaceted manipulation to ensure, in the absence of facial and vocal cues, that participants would perceive a difference in severity between the two emotional reaction conditions, which in fact they did (as shown by the manipulation checks). The disadvantage of strengthening the manipulation in this way is that the differences we observed could be attributable the victim's reported feelings, his/her behavior, or both. Future research could easily disentangle these attributes. In any case, it is clear that the norms of emotion that emerged in this study point toward specific expectations for victims of the robbery (i.e., the reaction should be sufficiently severe) but no parallel expectation for victims of the pick-pocket (i.e., substantial leeway for a range of reactions from mild to severe). Thus, there were normative expectations of the robbery victims, regardless of whether they were triggered by the victim's emotional experience, subsequent behavior, or some combination of the two.
We were frankly surprised that our participants appeared so reluctant to label the severe reaction to the less serious crime as “unusual.” In essence, people reported nothing amiss about a victim who, 6 months after a pick-pocket event, is tearful, “devastated,” preoccupied with the theft, and in permanent despair. In considering this result, we wondered if the “less serious” scenario was perceived as more significant than we had expected, perhaps because there was a small amount of physical contact (the perpetrator lightly brushing the victim) and people may have easily imagined a counter-factual situation with a much worse outcome (Kahneman & Miller, 1986).
We felt it important to explore boundary conditions on the latitude afforded victims who are highly emotional following an offense. If people do not consider a severe reaction to be unusual because the less serious crime entailed some physical contact between victim and offender, then a natural follow-up study would explore an instance in which physical harm to the victim would be difficult to imagine. We therefore examined a property loss in which the victim was not present when the actual theft occurred. The “serious” versus “less serious” manipulation was accomplished by varying the value of the amount taken ($35,000 versus $100). If perceivers have a general expectation that victims should be affected severely by offenses, and, likewise, that strong emotions and difficult coping are routine even when the victim did not lose much, then we should see patterns similar to those observed in Study 1. On balance, because a victim is never accused of making too much of an event, then a replication of patterns from Study 1 would lend greater support to the “victim-role rule” (with the modification that victims of less severe crimes are afforded latitude to be either highly affected or not). If, however, perceivers view the victim of a less serious crime to be over-reacting, then this supports the proportionality perspective, with a modified understanding of what makes an event “entitled” to strong or mild emotion.
Study 2: Fencing of a small or large amount of property
Method
Participants and procedure
Study 2 was a 2 (financial loss: high or low)×2 (emotional reaction: severe or mild)×2 (victim gender) design. Data came from two sources: 50 University of Texas employees, who were part of the same recruitment effort as that described in Study 1 (these participants also participated in the unrelated other study, presented in random order. There were no order effects associated with filling out this study first or second). In addition, 30 undergraduates in an upper division sociology course participated, for a total of 80 participants. There were scattered significant differences between students and employees (e.g., students had an overall more positive impression of the victim and also regarded the financial harm as more serious than the employees); we therefore included student as a factor to control for any differences.
The sample had a mean age of 30 (range=18–60; median=26), was 68% white, 73% female, and was again well-educated (all had at least some college). Most (66%) reported having had something stolen at some point in time; 23% had been victims of an assault, and 27% had been victims of some other crime. Over one-third (37%) reported knowing a victim of an assault. Apart from the student effect discussed above, demographic factors did not correlate with the variables of interest or alter the results; in addition, people of various backgrounds were successfully randomized across the cells of the design. Victims of a theft tended to think the emotional reactions expressed were more unusual, but controlling for crime-victim status did not alter the findings and is not discussed further.
Stimulus
Subjects read a story of a person (the victim) who wanted to transport some personal property. Based upon a friend's recommendation, the victim contracted with someone whom the victim believed to be a legitimate mover. However, the mover (the defendant) in this case turned out to be a drug addict who set up the bogus company to occasionally fence goods to support his drug habit. Rather than taking the victim's items to their new destination, the defendant took the property to another city and sold it to “a shady group of men.” After being apprehended, the defendant said that he did not know what became of the stolen goods after he was paid for them. His bogus moving company had no insurance to cover the loss. As in Study 1, we described the defendant as opting to plead guilty; sentencing took place 6 months after the event.
Financial loss was manipulated by reporting that the victim either was moving a chest of drawers he/she had purchased for $100 from an estate sale (low financial loss) or was moving everything in his/her apartment (worth an estimated $35,000). Both the high and low financial loss was paired with either a strong or mild emotional reaction, which we modified only slightly from Study 1. In the severe reaction condition, the victim again described the experience as “devastating”; he/she sobbed when testifying; was preoccupied with the event and believed it would “stay with me forever.” However, because we did not expect the loss of property to affect a person's willingness to go out at night (as described in Study 1), the victim in this case talked about a dramatic loss of faith in people. In the mild condition, the experience is again described as “troubling” and that he/she gets “irritated every once in a while” when reminded of it. The victim also describes being a little “more guarded” with people but as having “mostly forgotten about it,” and remaining active.
Measures
Except where noted, all items described were rated on a 7-point scale with endpoints variously labeled none, not at all (1) and extremely, a great deal, very (7).
Manipulation checks.
In addition to crime seriousness, we asked about how much financial harm the victim suffered, as well as how long the financial injury lasted (1: “no time at all,” 7: “for many years”). The checks on the manipulation of reaction were similar to those used in Study 1 (how upset the victim seemed and how much emotional injury he/she suffered).
Normality of the reaction.
We enhanced our norms measure in this study by including an additional rating. Respondents rated “How unusual” the victim's reaction seemed, as well as how likely is it that another person would have the same reaction (reverse-scored). Together, these formed a reliable composite (alpha = .87). We also asked the participants to rate how long the emotional injury was expected to last.
Ratings of the victim.
The same four items used in Study 1 were combined to create a reliable composite measure (alpha=.88).
Participants’ emotional reaction.
Participants rated their feelings toward the victims for the same four emotional reaction items used in Study 1: Anger, disgust, sympathy, and contempt.
Punishment/perceptions of the offense.
We utilized the same two items assessing punishment (ratings of how severe the punishment should be and proposed prison time). To create a composite measure of crime seriousness, we considered combining ratings of financial injury, financial harm, and how serious the offense seemed, but the alpha for these three items was unacceptably low (alpha < .40). We therefore examined the two-items on financial harm and injury (alpha=.86) separately from the crime seriousness rating.
Results
Manipulation checks
Participants correctly distinguished between the high and low financial losses across the two crime conditions, with the former being seen as more serious (Ms=3.15 versus 4.07), F(1, 71)=12.35, p < .001 (partial-η 2=.148), as involving more financial harm (Ms=2.78 versus 5.62) F(1, 69)=119.59, p < .0001 (partial-η 2=.634), and as involving a longer-term financial injury, F(1, 68)=80.09, p < .0001 (partial-η 2=.541; Ms=2.10 versus 5.03). In addition, the victim was viewed as significantly more upset in the severe reaction condition (M=5.68) than in the mild reaction condition (M=2.93), F(1, 71)=99.31, p < .0001 (partial-η 2=.583). The manipulation also affected perceptions of how much the victim was emotionally injured (M=3.89 versus 4.68 in the mild and severe conditions, respectively), F(1, 70)=4.57, p < .05 (partial-η 2=.061). Finally, people in the serious crime condition perceived the victim as more emotionally injured (M=4.95) than did those in the less serious condition (M=3.63), F(1, 70)=11.61, p < .01 (partial-η 2=.142).Footnote 3
Normality of the reaction
Perceptions of how unusual/common the victims’ reaction seemed were predicted by a significant crime×emotional reaction interaction, F(1, 71)=10.56, p < .01 (partial-η 2=.129). The form of this interaction is depicted in Table 2. The pattern of means provides support for the proportionality-rule—with the highest mean (i.e., the reaction seen as most unusual) occurring in the less serious crime/severe emotional reaction condition, and the second highest mean occurring in the other mismatched condition (serious crime/mild reaction). However, tests of the differences indicated that the severe reaction was significantly more unusual than the mild reaction only in the less serious crime condition (p < .001, d=1.10). Within the serious crime condition, the unusualness of the reaction did not significantly differ (p < .29). In other words, in contrast to Study 1, participants appeared to give latitude to the victim to react either severely or mildly in the serious (rather than less serious) crime condition; here what seemed most “unusual” was a severe reaction to the low financial loss.
Emotion duration estimates for this property crime produced a significant crime×gender interaction, F(1, 69)=7.92, p < .01 (partial-η 2=.103), indicating that the magnitude of financial loss significantly affected perceptions of how long the victim would suffer emotionally when the victim was a man (M=4.36 versus 2.71 for the serious versus less serious loss, respectively, p < .01, d=.95) but not a woman (M=3.35 versus 3.84, ns). Additionally, there was a trend for the significant three-way interaction between crime, emotional reaction, and gender, F(1, 69)=3.48, p < .07 (partial-η 2=.048). Tests of differences across cells produced only weak and sporadic effects. These were never inconsistent with the lower-order crime×gender interaction, which suggested that for this crime, men would be affected for a longer time by the larger financial loss; for example, within the serious crime, the mild reaction produced somewhat shorter estimates of emotion duration than the severe reaction but only when the victim was female.
Victim ratings
Unlike Study 1, the manipulated variables did have an effect on impressions of the victim. A highly significant main effect for reaction type, F(1, 71)=24.59, p < .0001 (partial-ç 2=.263), indicated that victims were more favorably perceived when they expressed the mild reaction (M=5.87) rather than the severe reaction (M=4.65), regardless of the seriousness of financial loss.
Participant emotion
Participants’ ratings of their feelings towards the victim did not vary significantly across experimental conditions (F-values for Wilks’ Lambda tests all ns).
Punishment and crime seriousness
Again, the less serious crime exhibited evidence of “floor” effects: Nearly 70% of participants in the $100 loss condition indicated a sentence of 2 years or less; half the sample recommended this in the more serious loss. As such, sentences did not vary significantly across conditions (overall F for the model p < .10), although participants reported wanting more severe punishment for the $35,000 loss (M=4.42) than for the $100 loss (M=3.53), F(1, 71)=8.58, p < .01 (partial-η 2=.108). Further, in addition to the manipulation check already reported (larger loss more serious than the smaller loss), there was a trend toward seeing the crime as more serious if the victim expressed stronger (M=3.83) rather than more mild (M=3.39) emotion, F(1, 71)=3.01, p < .09 (partial-η 2=.040). Emotional reaction did not predict views of the two-item severity of the financial loss, F(1, 69)=0.01.
Discussion
Using a crime involving little or no physical risk to a victim, we observed in Study 2 results that nicely complement those of Study 1 for ratings of the unusualness of emotional reactions. In Study 2, perceivers viewed a victim's severe emotional reaction to a less serious crime (a theft of an item worth only about $100) as quite unusual. Regardless of victim gender, a severe emotional reaction to the $100 loss was seen as significantly more atypical/unusual than a mild reaction. In contrast, and unlike Study 1, in the serious crime condition, there was no significant difference in perceptions of unusualness across the severe or mild emotional reaction conditions, although the pattern or means tended to support a proportionality-rule of expecting more upset victim when the loss is serious.
Even the few gender effects we observed were different for this crime than for the theft described in Study 1. In Study 1, emotion duration estimates largely tracked the severity of the emotional reaction (worse coping=longer expected duration), and expected emotional reactions were longer for serious crimes than mild ones, but only for female victims. By contrast, in Study 2, the male (but not the female) victim suffering a large financial loss was expected to be affected for a longer amount of time—and this was true no matter how severe the emotional reaction. At this point, one may be tempted to conclude that there is some fundamental difference between the victims’ roles in the two studies—for example, perhaps the victim is blamed more in this study (for hiring an errant mover) than in Study 1. However, both Studies 1 and 2 included victim-blaming measures (e.g., “To what extent could the victim have prevented this crime from happening?”). These never differed significantly across conditions in either study, making it unlikely that victim blaming explains the results we report here.
Considering the results across the two studies in light of our original expectations, we conclude that the proportionality perspective for perceptions of emotion expression received stronger support. Victims were seen as having “unusual” reactions when they responded calmly to a serious crime that involved dangerous interpersonal contact between victim and offender (a gun pointed to the victim's head, with the victim pushed to the ground in the course of a robbery); victims were also seen as atypical if they reacted very strongly to an event that would be hard to imagine as being difficult to overcome (the theft of a $100 chest of drawers). Thus, across the two studies we saw a pattern of matching an output (emotional reaction) to its input (crime severity); however, to observe this pattern, we had to offer quite extreme events to the participants, in which the crimes differed in terms of the personal threat to the victim (a gunpoint robbery at one extreme, and theft of a $100 piece of used furniture at the other). In between these two extremes, participants seemed reluctant to characterize responses as unusual. In short, the results of these two studies compel us to expand our a priori ideas about the precise form of a proportionality norm for victims; future work would be useful to further elucidate the norm.
It is fair to ask whether emotional displays or the nature of emotion norms ultimately matter given that, in these studies, what was seen as unusual was not generally related to punishment. For several reasons, we see emotion norms as a critical area to develop. First, our goal for these studies was to develop situations in which a given emotional response could conceivably be labeled as “unusual” in order to identify a normative understanding. This guided our choice of crimes, the sentences for which exhibited low variability, especially in the less serious crime conditions. Even so, as a post-hoc check, we examined the raw correlations between punishment (prison time and severity) and our norms measure. As our theorizing would suggest, perceptions of the unusualness of the reaction correlated negatively with prison time, r=−.21, although this was just shy of conventional significance levels (p < .07).Footnote 4 Additionally, Study 2 found a relationship between emotional reaction and liking for the victim, and this effect was opposite of what some theories (e.g., affect control theory) would predict—namely, it was the “stoic,” positive victim who was preferred. Thus, no matter the sentencing results, we would maintain that perceptions of emotional coping have the potential to contribute to perceptions of victims—something we believe deserves further study and understanding.
General discussion
Attention to emotion norms is valuable because several domains in law make norms part and parcel of decision-making. Lay perceptions within certain legal domains suggest the presence of a “victim-role” expectation for behavior, i.e., people expect victims of certain offenses to respond with a lot of emotion, despite the fact that many victims do not. Those who have advocated for the admissibility of rape trauma syndrome evidence, for example, argue that victims are penalized (in terms of credibility assessment) if their response to a rape is initially more muted and less emotional, and that expert testimony is necessary to counter the expectations we have of victims (see e.g., (Fischer, 1989)). Although evidence for the existence of a “syndrome” specific to rape may be questionable (Boeschen, Sales, & Koss, 1998), the larger concern likely has merit—including expectations of how victims are supposed to behave on the witness stand itself. For example, recent research depicting videotaped variations in victim testimony in rape cases finds that victims who fail to muster enough emotion to an event are seen as less credible, even if the content of their testimony is exactly the same as a victim who displays more emotion (Kaufmann, Drevland, Wessel, Overskeid, & Magnussen, 2002). The offense in question need not be criminal to generate expectations of strong emotion. In an intriguing experimental simulation, Woodzieka and LaFrance (Woodzicka & LaFrance, 2001) document a disconnect between the more controlled way women actually respond to sexually harassing situations (e.g., feeling confused over inappropriate questions or ignoring the violations) versus the outrage and anger that women who merely imagine the situation say they would express.
On the other hand, this study documents potential limits to the “victim-role” perspective, and there are other instances worthy of investigation. Jury researchers have found that plaintiffs who ask to be compensated for the “pain and suffering” related to an injury are often highly scrutinized, accused of exaggeration, or thought to be asking for compensation for something that all people experience (Hans, 2000; Vidmar, 1995). A better understanding of how people expect victims to react to offenses and harms would shed a great deal of light on when and under what circumstances credibility and justice decisions will be most influenced by feeling rules.
The nuance and varieties of potential investigations strikes us as rich. In non-legal contexts, social status variables (e.g., authority position) affect expectations about what people are supposed to feel. Tiedens (Tiedens, 2001), for instance, reports that observers react more positively to a high-status person who expresses anger, rather than sadness/remorse, over accusations against him (e.g., Bill Clinton testifying before a grand jury; see also (Kemper, 2000; Tiedens, Ellsworth, & Mesquita, 2000)). Subcultural norms, such as those associated with the Culture of Honor (Nisbett & Cohen, 1996) also appear to influence people's expectations regarding emotion, for example, by socializing people about the need for strong anger in response to an offense. Finally, the present study suggests that social categories are relevant for the larger literature on affective forecasting. The time estimate for male victims’ reactions to a property loss, for instance, was related to the value of the loss; but severity of the crime did not affect duration estimates for men in Study 1.
Our final note concerns methodology. Our work, like many experimental studies of victims, would benefit from more sophisticated stimulus materials that could effectively and convincingly depict the emotion being experienced. This would allow for more control in communicating a “severe” and “mild” emotional response, and would increase the chances that study participants would have emotional reactions themselves. Such reactions may influence not only judgment and decision-making of the case, but also of what is “unusual” or “typical.” We thus join with others who have called for more creative research methods to understand how we react to emotional statements from victims (Myers & Greene, 2004). As we have explained, we see the pay-off to illuminating both practical and theoretical questions as worth the efforts.
Conclusion
Emotions are a fact of life in legal settings—either because they are an inherent part of a decision-making task (e.g., the right amount of compensation for “suffering”) or, as in the case of victim impact statements, because the law has expanded evidentiary rules to make emotional reactions relevant to decisions. We encourage social science and law researchers to consider in more depth what makes a given emotional response “normal,” “understandable,” or “appropriate” because our work suggests that sometimes victims of unfortunate events can get it wrong.
Notes
Due to various factors (e.g., no-shows), our design is not fully balanced. To account for this, we report only the unique effect for each variable, or the “Type III” sums of squares.
The reliability for this composite is good, but not excellent (alpha=.66). We therefore analyzed the items individually, although results are largely consistent with either the composite or the individual items.
Unlike Study 1, some trend effects for interactions emerged in the manipulation checks. In particular, crime and gender interacted in predicting the amount of emotional harm (p < .09), such that men seemed significantly less emotionally harmed in the less serious, compared to the more serious, crime condition (Ms=3.14 versus 5.13, p < .001), whereas crime type did not differentiate perceptions of the emotional harm to the female victim (4.11 versus 4.77, ns). To be confident that this effect did not alter our reported findings, we ran all models with emotional harm as a covariate. Results were not substantially altered. There were also trends for the three-way interaction term on how upset the victim seemed (p < .08), and the amount of financial harm suffered (p < .07), but the pattern of means indicated the forms of these interactions were largely uninterpretable and not theoretically significant.
The relationship with prison sentence was non-significant for Study 1 data, although unusualness and punishment severity did correlate negatively, r=−.17, p < .09.
References
Arrigo, B., & Williams, C. R. (2003). Victim vices, victim voices, and impact statements: On the place of emotion and the role of restorative justice in capital sentencing. Crime and Delinquency, 49, 603–626.
Bandes, S. (1996). Empathy, narrative, and victim impact statements. University of Chicago Law Review, 63, 361–412.
Boeschen, L. E., Sales, B. D., & Koss, M. P. (1998). Rape trauma experts in the courtroom. Psychology, Public Policy, and Law, 4, 414–432.
Brody, L. R., & Hall, J. A. (2000). Gender, emotion and expression. In M. Lewis & J. M. Haviland-Jones (Eds.), Handbook of emotions (2nd ed., pp. 338–349). New York: Guilford.
Clark, C. (1997). Misery and company. Chicago, IL: University of Chicago Press.
Cohen, M. A. (1988). Pain, suffering, and jury awards: A study of the cost of crime to victims. Law and Society Review, 22, 537–555.
Ekman, P., & Friesen, W. V. (1969). The repertoire of nonverbal behavior: Categories, origins, usage, and coding. Semiotica, 1, 49–98.
Erez, E. (1999). Victim impact statements as victim empowerment and enhancement of justice. Criminal Law Review, July, 545–556.
Fischer, K. (1989). Defining the boundaries of admissible expert psychological testimony on rape trauma syndrome. University of Illinois Law Review, 1989, 691–734.
Gordon, S. L. (1990). Social structural effects on emotions. In T. D. Kemper (Ed.), Research agendas in the sociology of emotions (pp. 145–179). Albany: State University of New York Press.
Green, D. P., Goldman, S. L., & Salovey, P. (1993). Measurement error masks bipolarity in affect ratings. Journal of Personality and Social Psychology, 64, 1029–1041.
Greene, E. (1999). The many guises of victim impact evidence and effects on jurors’ judgments. Psychology, Crime and Law, 5, 331–348.
Hans, V. (2000). Business on trial. New Haven, CT: Yale University Press.
Heider, F. (1958). The psychology of interpersonal relations. Hillsdale, NJ: Lawrence Erlbaum.
Heise, D. R. (1989). Effects of emotion displays on social identification. Social Psychology Quarterly, 52, 10–21.
Heise, D. R., & Calhan, C. (1995). Emotion norms in interpersonal events. Social Psychology Quarterly, 58, 233–240.
Hills, A. M., & Thomson, D. M. (1999). Should victim impact influence sentences? Understanding the community's justice reasoning. Behavioral Sciences and the Law, 17, 661–671.
Hochschild, A. (1983). The managed heart. Berkeley, CA: University of California Press.
Igou, E. R. (2004). Lay theories in affective forecasting: The progression of affect. Journal of Experimental Social Psychology, 40, 528–534.
Kahneman, D., & Miller, D. T. (1986). Norm theory: Comparing reality to its alternatives. Psychological Review, 93, 136–153.
Kaufmann, G., Drevland, G. C. B., Wessel, E., Overskeid, G., & Magnussen, S. (2002). The importance of being earnest: Displayed emotions and witness credibility. Applied Cognitive Psychology, 17, 21–34.
Kemper, T. D. (2000). Social models in the explanation of emotion. In M. Lewis & J. M. Haviland-Jones (Eds.), Handbook of emotions (2nd ed., pp. 45–58). New York: Guilford.
Labott, S. M., Martin, R. B., Eason, P. S., & Berkey, E. (1991). Social reactions to the expression of emotion. Cognition and Emotion, 5, 397–417.
Lerner, M. (1980). The belief in a just world. New York: Plenum.
Myers, B., & Greene, E. (2004). The prejudicial nature of victim impact statements: Implications for capital sentencing policy. Psychology, Public Policy, and Law, 10, 492–515.
Nadler, J., & Rose, M. R. (2003). Victim impact testimony and the psychology of punishment. Cornell Law Review, 88, 419–456.
Nisbett, R., & Cohen, D. (1996). Culture of honor: The psychology of violence in the South. Boulder, CO: Westview.
Roberts, J., & Erez, E. (2004). Communication in sentencing: Exploring the expressive function of victim impact statements. International Review of Victimology, 10, 223–244.
Smith-Lovin, L., & Heise, D. (1988). Analyzing social interaction: Advances in affect control theory. New York: Gordon and Breach Science Publishers.
Thoits, P. (1989). The sociology of emotion. Annual Review of Sociology, 15, 317–342.
Thoits, P. A. (1990). Emotional deviance: Research agendas. In T. D. Kemper (Ed.), Research agendas in the sociology of emotions (pp. 180–203). Albany: State University of New York Press.
Tiedens, L. (2001). Anger and advancement versus sadness and subjugation: The effect of negative emotion expressions on social status conferral. Journal of Personality and Social Psychology, 80, 86–94.
Tiedens, L. Z., Ellsworth, P. C., & Mesquita, B. (2000). Stereotypes about sentiments and status: Emotional expectations for high- and low-status group members. Personality and Social Psychology Bulletin, 26, 560–574.
Tsoudis, O., & Smith-Lovin, L. (1998). How bad was it? The effects of victim and perpetrator emotion on responses to criminal court vignettes. Social Forces, 77, 695–722.
Vidmar, N. (1995). Medical malpractice and the American jury: Confronting the myths about jury incompetence, deep pockets, and outrageous damage awards. Ann Arbor, MI: University of Michigan Press.
Wilson, T. D., & Gilbert, D. T. (2003). Affective forecasting. Advances in Experimental Social Psychology, 35, 345–411.
Woodzicka, J. A., & LaFrance, M. (2001). Real versus imagined gender harassment. Journal of Social Issues, 57, 15–30.
Acknowledgements
We were greatly assisted by financial support from the American Bar Foundation. A version of this paper was presented to the American Sociological Association's Annual Meeting in August 2004.