The Effect of Hypervigilance on the Relationship Between Sexual Victimization and Gait

Abstract

Two studies investigated the relationship between hypervigilance, vulnerable gait cues, and a history of sexual victimization. In Study 1 (N = 130), gait was coded for traits relating to vulnerability where half of the sample was unaware of being videotaped (Unaware condition) and the other half was aware (Aware condition) to induce hypervigilance (between-subjects design). Gait was associated with a history of victimization, but only in the Unaware condition. A mediation analysis found that perceived impact of victimization mediated the association between victimization and vulnerable gait. In Study 2, female university students (N = 62) were measured on their victimization history and hypervigilance. Walking styles of participants were coded for the presence of vulnerability cues in both an Unaware and Aware condition (within-subjects design). A regression analysis revealed an association between hypervigilance and a reduced change in walking style between the two conditions. More notably, hypervigilance was found to moderate the relationship between sexual victimization and vulnerable gait but not violent victimization and vulnerable gait. These results suggest that hypervigilance may be an adaptive response that reduces perceived vulnerability in sexually victimized women.

Keywords

walking style gait vulnerability sexual victimization hypervigilance

Research has consistently demonstrated that an individual’s vulnerability to victimization can be reliably detected based on nonverbal cues such as gait (i.e., the way a person walks; Book, Costello, & Camilleri, 2013; Grayson & Stein, 1981; Gunns, Johnston, & Hudson, 2002; Wheeler, Book, & Costello, 2009). More specifically, a history of victimization has been associated with a distinct walking style that is perceived as more vulnerable by independent observers (Book et al., 2013; Wheeler et al., 2009). Despite the consistent association between gait and vulnerability, research has also suggested that the relationship between vulnerability and gait may be influenced by hypervigilance (Johnston, Hudson, Richardson, Gunns, & Garner, 2004). Hypervigilance is defined as an enhanced sensitivity to threat cues in the environment that is found to be more severe in cases of sexual victimization than in other forms of victimization (Cortina & Kubiak, 2006). The current set of studies sought to examine the effect of hypervigilance on the relationship between victimization and vulnerable gait cues and whether this effect depends on type of victimization (sexual or violent). The potential effects of other cognitive variables, namely, self-perceived vulnerability and perceived impact of victimization were also explored.

Gait as an Indicator of Vulnerability

Previous research has demonstrated that gait cues can be used as an indication of an individual’s vulnerability to victimization (e.g., Grayson & Stein, 1981; Murzynski & Degelman, 1996). For example, after instructing offenders to watch a series of videos depicting women walking down a sidewalk alone and rate each woman’s vulnerability to victimization, Grayson and Stein (1981) reported agreement among the offenders in identifying the women who were more vulnerable to victimization. In addition, Grayson and Stein (1981) identified similar movement cues among the vulnerable walkers, such as long or short strides (as opposed to normal strides), fast or slow speeds, nonlateral weight shifts, and lifting feet higher. Based on these results, Grayson and Stein (1981) concluded that individuals perceived as more vulnerable to victimization displayed a lack of fluidity and synchronicity in their walk. Such findings were among the first to demonstrate walking style as a potential factor in victim selection. Extending this research, Murzynski and Degelman (1996) recorded female walkers walking in a manner that portrayed the vulnerability cues identified by Grayson and Stein (1981) or walking without exhibiting these cues (i.e., a nonvulnerable walk). Undergraduate students and law enforcement officials were then asked to rate the vulnerability and confidence of each female walker. Both the students and law enforcement officials rated the experimental walkers (those walking with vulnerability cues) as significantly more vulnerable and less confident than the control walkers.

The above findings are consistent with the kinematic specification of dynamics (KSD) principle (Runeson & Frykholm, 1983), which states that features of movement reflect the intrinsic causes of an event. Runeson and Frykholm (1981) first demonstrated this principle by instructing participants to judge the weight of a box based on movement information using a point-light technique, which involves the placement of light-reflective markers on an individual’s joints to analyze bodily movements. The results of this study demonstrated that participants’ judgments based on movement cues were almost as accurate as their judgments based on lifting the box themselves (Runeson & Frykholm, 1981). Based on these results, Runeson and Frykholm (1981) concluded that the dynamics of an event (e.g., weight of a box) produce certain movement cues that individuals are then able to detect. Therefore, it may be possible to determine an individual’s intrinsic motivations, expectations, and feelings through movement information.

Drawing from the KSD principle, it is important to determine whether judgments of vulnerability based on gait reflect the walker’s actual vulnerability to victimization. To explore this question, Wheeler et al. (2009) recorded female participants walking down a hallway without their knowledge (to capture their natural gait) and assessed their history of victimization (i.e., victimization experiences equal to or greater than bullying). The walking style of each walker was then coded according to the vulnerability cues identified by Grayson and Stein (1981). An independent group of male participants were then recruited to rate each walker on their vulnerability to being mugged. Results indicated that the walkers rated as more vulnerable to victimization displayed a greater number of vulnerable gait cues than those rated as less vulnerable to victimization. The walkers rated as more vulnerable to victimization were also more likely to have experienced victimization in the past than those rated as less vulnerable to victimization (Wheeler et al., 2009).

The Effects of Hypervigilance on Gait

The above research demonstrates that individuals who are vulnerable to victimization can be reliably detected based on specific gait cues. To reduce the risk of future victimization, it is therefore important to identify any factors that may protect individuals from manifesting these cues. One possible factor is hypervigilance, which is typically described as an attentional bias toward threatening stimuli and is a common symptom among anxiety disorders, such as posttraumatic stress disorder (PTSD; Bogels & Mansell, 2004). According to the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association, 2013), hypervigilance is a part of the cluster of arousal symptoms associated with PTSD and involves a heightened sensitivity to stimuli that the individual perceives as threatening. Due to the increased anxiety and exhaustion associated with a global awareness of threat, hypervigilance is considered a maladaptive psychological symptom of PTSD (Kimble et al., 2014). However, hypervigilance is also found to result in protective behaviors across species such as scanning the environment for threat cues and avoidance of the source of threat once it is identified (Garner, Mogg, & Bradley, 2003). Hence, while hypervigilance is considered a psychological vulnerability, it may also lead to adaptive behaviors that would increase one’s chances of survival.

Based on the assumptions of the KSD principle noted earlier, it is plausible that hypervigilance would be reflected in one’s walking style. However, given its adaptive and maladaptive features, it is important to determine whether hypervigilance would increase or decrease cues of vulnerability. In a series of three studies, Johnston et al. (2004) attempted to demonstrate the role that hypervigilance may have in reducing nonverbal vulnerability cues displayed by women. Female walkers were recorded using a point-light technique and were told to either walk naturally (normal condition) or as though they were alone in a city park at night (threat condition). A separate group of participants then rated each walker for their vulnerability to being attacked. This study also examined whether training could result in less vulnerable walking styles. In this training, walkers observed videos of gait patterns that were rated as difficult to attack and were given guided instruction as to how to inhibit specific vulnerability cues. Results indicated that walkers in the threat condition exhibited less vulnerability cues and were rated as more difficult to attack than the walkers in the normal condition (Johnston et al., 2004). In addition, walkers who received training had significantly lower vulnerability ratings than before the training (Johnston et al., 2004). Such findings are in line with previous studies (e.g., Runeson & Frykholm, 1983) demonstrating the impact of cognitive variables (e.g., awareness) on nonverbal behavior. Drawing from this research, hypervigilance may serve as an adaptive function in suppressing vulnerability cues in walking style.

Although evidence suggests that hypervigilance may reduce vulnerable gait cues, it is possible that the strength of this effect may depend on the type of victimization experienced by the walker. In particular, previous research has indicated that sexual victimization is associated with an increased risk of developing PTSD (Bolstad & Zinbarg, 1997; Norris, 1992) and a greater severity of PTSD symptoms, such as hypervigilance (Cortina & Kubiak, 2006; Van Ameringen, Mancini, Patterson, & Boyle, 2008). For example, Cortina and Kubiak (2006) conducted a telephone survey that assessed the extent of PTSD and the prevalence of sexual assault, physical assault, and stalking in men and women. Their results demonstrated that, compared with men, women had an increased risk and severity of PTSD-related symptoms, including hypervigilance. In addition, the effect of gender on symptom severity was mediated by experiences of sexual victimization but not by experiences of physical assault or stalking (Cortina & Kubiak, 2006). This suggests that sexual victimization may be associated with a more severe symptomatology compared with other types of victimization, particularly for women. Despite experiencing more maladaptive symptoms of PTSD, sexually victimized women may also exhibit more pronounced effects of hypervigilance in their walking style.

The Current Set of Studies

Taken together, it is evident that inducing hypervigilance decreases walking style cues indicative of vulnerability (Johnston et al., 2004). Given the established relationship between cognition and nonverbal behavior, it is also possible that cognitive factors play a role in the relationship between walking style and previous victimization. The purpose of the first study was to investigate the effect of hypervigilance on the relationship between victimization and walking style and whether self-perceived vulnerability and perceived impact of victimization mediated these relationships. Study 1 was a between-subjects design, where half of the participants were aware of being videotaped, and the other half were unaware. Based on research demonstrating the effects of situational threat on walking style (e.g., Johnston et al., 2004), we expected a relationship between vulnerable gait cues and victimization history in the Unaware condition but not the Aware condition, given that the knowledge of being watched induces hypervigilance. We further hypothesized that self-perceived vulnerability and perceived impact of victimization would mediate the relationship between walking style and victimization history in the Unaware condition.

Drawing from previous research indicating the greater severity of hypervigilance in sexual assault victims (Cortina & Kubiak, 2006), the second study examined whether the influence of hypervigilance on walking style would be more pronounced in those who were sexually victimized compared with those who were violently victimized. This study employed a repeated measures design, in which participants were taped without their awareness, and then taped with their awareness, followed by their completion of a hypervigilance measure. In line with Study 1, we expected that vulnerable walking style cues would be reduced in the Aware condition relative to the Unaware condition due to the induction of hypervigilance. In addition, we predicted that this reduction in vulnerability cues would be less pronounced for hypervigilant individuals than non-hypervigilant individuals because they presumably exhibit less vulnerability cues in general. Finally, we expected that hypervigilance would moderate the relationship between sexual victimization and walking style but not the relationship between violent victimization and walking style.

Study 1

Method

Participants

One hundred thirty female undergraduate students from a Canadian University (M_age = 20.22, SD = 2.31, age range: 18-29 years) participated in Study 1. In exchange for participation, participants were given $10 and, if they chose, could also use the experience as the basis for a half credit in a psychology course. In the Unaware condition, each participant was independently videotaped walking down a hallway without their awareness (see procedure) and were later asked for consent to use the videos for vulnerability coding. In the Aware condition, participants were told that they would be taped as they walked down the hallway in order to examine their gait and its relationship to victimization history. Upon reviewing the videos, five of the 65 videos from each condition were excluded from the analysis, as the video clips were deemed inappropriate (e.g., the participants natural movement was disrupted because they were carrying something, the quality of the video did not permit accurate interpretation of gait cues, participants were impeded by others in the frame). A final sample of 120 female participants (60 in each condition) was used for the analyses.

Measures

Victimization history

A modified version of the Life Events Checklist (LEC; Brand & Johnson, 1982) was used to assess each participant’s previous victimization. The measure included 49 items and assessed positive (e.g., got married, had children, graduated) and negative (e.g., infidelity, life-threatening illness or accident, unemployment) life events. Participants completed the entire measure of positive and negative events, but we only examined the violent (which includes sexual assault) events for the purpose of Study 1. Violent victimization included physical bullying, assault, as well as any hands-on sexual victimization. For each event, participants were asked to document the frequency of occurrence during adulthood. The total number of violent victimization events that was endorsed was summed to create a violent victimization history score for each target where higher scores indicated higher violent victimization. Of the 120 participants, 95 (79.2%) reported at least 1 incident of violent (including sexual) victimization.

Perceived impact of victimization

The subjective impact of victimization was examined by asking participants to rate the impact that each life event (from the LEC) had on their everyday functioning. Participants rated perceived impact of victimization on a 10-point Likert-type scale ranging from 1 (no impact) to 10 (large impact). Targets reporting 0 incidents for a particular event were assigned impact ratings of 0 based on the rationale that they could not have been impacted by something that has not happened to them.

Self-perceived vulnerability

Each participant was asked to report her vulnerability to victimization (e.g., “overall, how vulnerable to victimization do you feel?”). The definition of victimization was left open to interpretation as previous research has shown that self-identification as a victim is more important than severity (Theriot, Dulmus, Sowers, & Johnson, 2005). Participants scored their vulnerability on a 7-point scale ranging from 1 (not vulnerable) to 7 (very vulnerable).

Walker videos

Gait cues were coded based on Grayson and Stein’s (1981) vulnerable body language cues, as employed by Wheeler et al. (2009). The cues are stride length, synchronicity, energy, knees, and head. Scores of 0, 1, or 2 were given on each of the relevant gait cues (0 = not present, 1 = some indication of the feature in question, 2 = feature is quite distinct/pronounced). Scores on these cues were summed for an overall gait-vulnerability score. Two independent raters coded the gait cues achieving adequate to high interrater agreement for all cues (Cohen’s kappa ranged from .63 to 1.00).

Procedure

Unaware condition

Participants were greeted in Study Room A, where they were asked to privately read the information and consent form while the researcher stepped out of the room. After a brief period, the researcher returned to the room and verbally described the study, associated tasks, compensation, participant rights, and obtained consent. The researcher then informed participants that the questionnaires would be completed in an alternate location with a collaborating researcher to ensure complete anonymity. While transitioning from Study Room A to B, participants were videotaped surreptitiously from behind by a video camera (Sony Handycam 8GB DCR-SX41/L) hidden in the hallway. Because participants were videotaped from behind, faces were not visible, allowing for relative anonymity. Videotaping participants without their awareness permitted researchers to capture the natural gait movements within a controlled environment. The hallway between Study Room A and B is in a relatively familiar location of the university, and is approximately 40 feet long, and 10 feet wide. Upon arriving at Study Room B, participants completed the demographic and victimization questionnaires and were subsequently sent back to Study Room A for a thorough debriefing.

Aware condition

Upon arriving, participants completed the demographic and victimization questionnaires. At this point, they were informed that the researcher would like to videotape them walking down the hallway. While transitioning from Study Room A to B, participants were videotaped from behind by a video camera (Sony Handycam 8GB DCR-SX41/L). Because participants were videotaped from behind, faces were not visible, allowing for relative anonymity.

Results

Descriptive statistics for variables in Study 1 are given in Table 1. An examination of the data revealed severe positive skews for the frequency of violent victimization reported on the modified LEC in both the Aware and Unaware conditions. Moreover, examination of the scatterplots revealed a violation of homoscedasticity. Due to the extreme, though understandable skewness for victimization frequencies, each victimization event on the LEC was dichotomized (yes/no). The frequency of victimization events was summed to create a total violent victimization score where higher scores indicated higher violent victimization, which met the assumption of normality.

Prediction 1: Gait vulnerability will correlate with victimization history but only in the Unaware condition.

Table 1.

Descriptive Statistics by Condition for Study 1.

Variable	Actual Range	M	SD
Unaware
Vulnerable gait	1-8	4.13	2.24
Violent victimization	0-7	1.88	1.58
Impact of victimization	0-7	3.51	2.44
Self-perceived vulnerability	1-6	2.86	1.30
Aware
Vulnerable gait	1-6	2.95	2.01
Violent victimization	0-7	1.93	1.42

Note. Total sample size is 120 (60 per condition).

To test whether gait vulnerability was positively correlated with victimization history, we conducted correlational analyses. As expected for the Unaware condition, vulnerable gait was significantly associated with overall victimization history, r = .44, p < .01. In other words, participants with more extensive victimization histories were more likely to display a vulnerable gait. Also as expected, the relationship between gait and previous victimization was nonsignificant in the Aware condition, r = .11, p = .23.

Prediction 2: In the Unaware condition, the relationship between gait vulnerability and victimization history will be mediated by cognition (i.e., self-perceived vulnerability and/or perceived impact of the victimization).

We have demonstrated that gait vulnerability is associated with violent victimization history; however, we also wanted to test whether cognitive variables such as self-perceived vulnerability or perceived impact of victimization could explain this relationship. We therefore conducted mediation analysis; if, after accounting for these cognitive variables, the relationship between vulnerable gait and victimization history is no longer significant, it can be concluded that these factors mediate (or help explain) the original relationship. Prior to mediation analysis, we examined correlations between gait and potential mediators. Self-perceived vulnerability was unrelated to participant gait, r(58) = .07 p = .60. That is, participants who displayed a greater number of vulnerability markers in their gait did not necessarily rate themselves as more vulnerable to victimization. The second potential mediator was the subjective impact of previous victimization. As expected, gait was significantly correlated with the perceived impact of victimization, r(58) = .32, p < .05, implying that there is some part of the vulnerable walk that relates to how victims perceive or react to their previous victimization.

To test whether the relationship between gait and victimization history is dependent upon the two potential mediators, we conducted a mediation analysis in PROCESS (Hayes, 2013). The final model included victimization history, self-perceived vulnerability, and impact in predicting gait cues. The overall regression equation was significant, F(3,55) = 3.47, p = .02; R² = .16, indicating that together, victimization history and the two mediators significantly predicted vulnerable gait, and the combination of variables accounted for 16% of the variance in gait cues. As seen in Table 2, violent victimization history is no longer related to gait vulnerability once the mediators are also included in the model. More importantly, when examining the mediators, it is impact of victimization that was a significant predictor of vulnerable gait; self-perceived vulnerability did not significantly predict vulnerable gait. Therefore, perceived impact of victimization helps to explain the original relationship between violent victimization history and gait vulnerability. Confidence intervals confirm this result, with impact of victimization not including zero in its confidence interval.

Table 2.

Results for Self-Perceived Vulnerability and Perceived Impact of Victimization as Mediators for the Relationship Between Victimization History and Gait Vulnerability.

Predictor	b	SE	t	p	95% CI
Violent victimization history	.01	.01	1.18	.24	[−.01, .03]
Self-perceived vulnerability	−.09	.21	−0.44	.66	[−.52, .33]
Impact	.31	.12	2.65	.01	[.08, .55]

Note. Sample size is 55 for all models. CI = confidence interval.

Study 1: Summary

In Study 1, we predicted that (a) gait would be related to a history of victimization only in people who are unaware of being observed; (b) cognitive variables, such as the perceived impact of the victimization, would predict a vulnerable gait in the Unaware condition; and (c) cognitive variables would mediate the relationship between gait and victimization history. All three hypotheses were supported. Our finding that gait was only related to victimization history in the Unaware condition suggests that attention to environment is an important factor in ratings of vulnerability. These results are also consistent with Johnston et al.’s (2004) finding that people exhibit a vulnerable walk when they feel safe, but do not exhibit a vulnerable walk in (a) unsafe contexts, and (b) with defense training, which presumably increases attention to environment. We also confirmed that the perceived impact of victimization (i.e., cognitive factor) plays an important role in the relationship between victimization history and vulnerable gait cues.

While the difference between the Aware and Unaware conditions supports the hypervigilance hypothesis, we did not directly measure hypervigilance. It would also be beneficial to replicate the current findings within individuals as we would expect that vulnerable gait cues would be related to victimization history when the person was unaware of being taped, but would not be related when the same person was aware of being taped. Thus, we conducted a second study to determine whether (a) results would be replicable in a repeated measures design, (b) hypervigilance moderates the relationship between gait and victimization history in the Unaware condition, and (c) the moderating effect of hypervigilance differs for sexual and violent victimization.

Study 2

Method

Participants

In contrast to Study 1, Study 2 was a within-subjects design where all participants served in both the Unaware and Aware conditions. Participants could not sign up to complete Study 2 if they had already completed Study 1. The sample for Study 2 originally consisted of 73 undergraduate female students from a Canadian University recruited through an online research portal. The mean age was 19.1 (SD = 1.6, range: 17 to 26) and the majority of participants were Caucasian. Two participants were removed from the study due to poor quality of the video recording. Nine more sets of videos were excluded from analyses for one of two reasons: (a) other individuals were in the video frame, and (b) the target individual was carrying something that would affect natural arm movement. The analyses are therefore based on a final sample size of 62 participants recorded twice (124 videos).

Materials

Demographics

Participants were instructed to complete basic demographic information such as their age and nationality. Gender (male or female) was also recorded, although only females were included in this study.

Victimization history

The modified LEC from Study 1 was also administered in Study 2. For each experience, participants are asked to indicate how many times it occurred, the impact of the event, and its probability of occurrence in the general population (only the occurrence of the event was used in this study). In Study 2, violent and sexual victimization were mutually exclusive. Examples of the experiences of physical victimization include physical assault, bullying, and robbery. Examples of the experiences of sexual victimization include sexual assault, attempted sexual assault, and sexual abuse. The number of times participants experienced each event was summed creating continuous variables reflective of violent and sexual victimization history.

Hypervigilance Questionnaire (HVQ)

The HVQ (Kimble, Frueh, & Marks, 2009) is an 11-item self-report scale that measures the respondent’s degree of hypervigilance. This scale includes items such as “I often feel as if I’m on guard” and “When I am out on the street, I am often looking over my shoulder.” Each item is scored from 1 to 5 where 1 represents “not at all true” and 5 represents “extremely true,” with total scores ranging from 11 to 55 (higher scores indicative of more hypervigilance). Previous research has demonstrated high reliability for the HVQ, with a Cronbach’s alpha of .92 and a split-half reliability of .89 (Kimble et al., 2009). In addition, the HVQ was found to be highly correlated with the PTSD Symptoms Scale Self-Report (Foa, Riggs, Dancu, & Rothbaum, 1993) demonstrating good convergent validity (Kimble et al., 2009).

Walker videos

Videos from two participants were removed due to a lack of quality in the recording, and nine others were removed due to (a) carrying items that would affect arm movement and (b) having other individuals in the video clip (final sample = 62). All participants were recorded twice (Unaware vs. Aware condition), resulting in a total of 124 videos. Like Study 1, gait cues were coded according to the vulnerability cues outlined by Grayson and Stein (1981) and were then summed to yield a total vulnerability score where higher scores represented higher vulnerability (Wheeler et al., 2009). Two coders blind to the victimization history and vigilance scores of the targets coded 10 of the videos from the Unaware condition and achieved acceptable interrater reliability (range: .72 to .85; Mdn = .81). The remaining videos were divided and coded by these two coders.

Procedure

Recruitment of participants and the procedure for filming students without their knowledge proceeded as in Study 1 (participant provides consent in Room A, is told to move to Room B, and is filmed as they transition between rooms). Upon arriving in Study Room B, participants were administered the questionnaire package that included the demographics sheet, LEC, and HVQ. After completing the questionnaire package, participants were informed by a researcher that they were recorded without their knowledge for the purpose of the study. The researchers subsequently asked the participant for their consent to film them going back into Room A to capture their gait. For all participants who consented, the researcher recorded the participant walking back to Room A using a different camera (i.e., Aware condition). Once the participant arrived at Room A, they were asked for their consent to use the videos for various research purposes. The participant was then debriefed and thanked for their participation. Each participant received additional course credit and 10 dollars for their participation.

Results

Prediction 1: Participants will display less vulnerable gait cues in the Aware condition compared with the Unaware condition. This change in gait across conditions will be smaller for participants scoring higher on hypervigilance.

Descriptive statistics for Study 2 can be found in Table 3. A paired samples t test confirmed that there was a significant reduction in the average number of vulnerable gait cues displayed from the Unaware condition (M = 2.58, SD = 1.12) to the Aware condition (M = 0.95, SD = 1.00), t(61) = 9.80, p < .001, d = 1.25, r = .24. Next, a bivariate correlation analysis revealed a significant negative relationship between hypervigilance scores and change in gait, r(62) = −.23, p < .05, indicating that individuals scoring higher on hypervigilance had less change in their gait from the Unaware to Aware condition.

Prediction 2: The effect of hypervigilance on the relationship between vulnerable gait cues and victimization history will be more pronounced for participants with a history of sexual victimization compared with violent victimization.

Table 3.

Descriptive Statistics by Condition for Study 2.

Variable	Actual Range	M	SD
Vulnerable gait
Unaware	0-5	2.58	1.12
Aware	0-5	0.95	1.00
Violent victimization	0-22	4.59	5.75
Sexual victimization	0-6	0.68	1.30
Hypervigilance	15-49	29.21	9.22

Note. Total sample size is 62.

To assess differences between the two types of victimization, bivariate correlations were first conducted to examine the relationship between victimization and gait in both conditions (Aware and Unaware). Results indicated that sexual victimization had a smaller relationship to gait in the Unaware condition, r(62) = .30, p < .05, in comparison with violent victimization, r(61) = .49, p < .01. In the Aware condition, as in Study 1, these relationships decreased, demonstrating the effect of inducing hypervigilance on the relationships between violent victimization and gait, r(61) = .02, p = .89, and sexual victimization and gait, r(62) = .04, p = .77.

Two hierarchical multiple regression analyses were conducted to test our prediction that hypervigilance moderates the relationship between sexual victimization and gait but not the relationship between violent victimization and gait in the Unaware condition. All assumptions for multiple regression analysis were satisfied (e.g., linearity, independence of errors, no outliers; Field, 2013). The first analysis specifically tested whether hypervigilance moderated the relationship between sexual victimization and vulnerable gait cues in the Unaware condition. The overall regression was significant, F(3, 57) = 4.48, p < .01, with a significant interaction between sexual victimization and hypervigilance in predicting gait (t = 2.34, p = .02; see Table 4), which supports our prediction that hypervigilance moderates the relationship between sexual victimization and gait. Thus, these results suggest that the relationship between sexual victimization and gait is lessened in the presence of hypervigilance (see Figure 1).

Table 4.

Analyses of Hypervigilance as a Moderator for the Relationship Between Victimization and Gait for Both Sexual and Violent Victimization.

Step	R ²	F Change	p	Variable	b	SE	β	t	p	Lower CI	Upper CI
1	.11	3.69	.03	Sexual victimization	0.28	0.10	.33	2.68	.01	0.07	0.48
1	.11	3.69	.03	Hypervigilance	−0.01	0.02	−.10	−0.77	.44	−0.04	0.02
2	.19	5.50	.02	Sexual victimization	0.16	0.11	.19	1.44	.16	−0.06	0.38
				Hypervigilance	−0.02	0.01	−.15	−1.21	.23	−0.05	0.01
				Interaction	0.001	0.001	.32	2.34	.02	0.0001	0.001
1	.26	9.91	< .001	Violent victimization	0.10	0.02	.51	4.43	< .001	0.05	0.14
1	.26	9.91	< .001	Hypervigilance	−0.01	0.01	−.12	−1.02	.31	−0.04	0.01
2	.26	0.73	.40	Violent victimization	0.08	0.03	.42	2.82	.01	0.02	0.14
				Hypervigilance	−0.02	0.01	−.13	−1.15	.25	−0.04	0.01
				Interaction	0.001	0.001	.13	0.85	.40	0.001	0.001

Note. Sample size is 62 for models of sexual victimization; models of violent victimization are based on a sample size of 61 given missing information on the hypervigilance measure for one participant.

Figure 1.

Moderating effect of hypervigilance on the relationship between sexual victimization and gait cues.

The second hierarchical multiple regression analysis was conducted to determine whether hypervigilance moderates the relationship between violent victimization and gait in the Unaware condition. The results revealed that violent victimization significantly predicted gait in the Unaware condition; however, the interaction between violent victimization and hypervigilance was not significant (t = 0.85, p = .40; see Table 4). This suggests that hypervigilance does not moderate the relationship between violent victimization and gait in the Unaware condition.

Study 2: Summary

The results of Study 2 supported our predictions that (a) individuals exhibit less vulnerable gait cues in the Aware condition compared with the Unaware condition, (b) this difference is less pronounced for those scoring higher in hypervigilance, and (c) the effect of hypervigilance on the relationship between vulnerable gait cues and victimization history is more pronounced for individuals with a history of sexual victimization compared with individuals with a history of violent victimization. Such findings support the contention that a hypervigilant disposition is associated with less vulnerability cues in general due to a greater global awareness of threat. Results from Study 2 also specifically revealed that hypervigilance moderates the relationship between sexual victimization and gait but does not moderate the relationship between violent victimization and gait. These results are in keeping with previous research (Cortina & Kubiak, 2006), which reported the increased severity of PTSD symptoms, including hypervigilance, in women who experienced sexual victimization relative to other forms of victimization. Thus, it appears that hypervigilance functions as a protective factor for sexually victimized women by decreasing the vulnerability cues in walking style associated with previous victimization.

General Discussion

The major aim of the current research was to investigate the effect of hypervigilance on the relationship between victimization and walking style. In both studies, we established that the relationship between victimization and gait was stronger in the Unaware condition than in the Aware condition, confirming that awareness decreases vulnerability cues in walking style. In addition, Study 1 revealed that perceived impact of victimization mediated the relationship between victimization and gait in the Unaware condition. In Study 2, self-reported hypervigilance was found to be associated with less of a decrease in vulnerability cues between the two conditions, demonstrating that hypervigilant individuals exhibit less vulnerability cues in general. In addition, it was found that hypervigilance moderates the relationship between sexual victimization and walking style but does not moderate the relationship between violent victimization and walking style.

The above results support the findings of previous research (Grayson & Stein, 1981; Wheeler et al., 2009), demonstrating the association between victimization and distinct vulnerability cues in walking style. Furthermore, the finding that inducing hypervigilance through the knowledge of being watched decreases vulnerability cues in walking style is in line with the study by Johnston and colleagues (2004), which showed that inducing threat salience through imagination results in less vulnerable walking styles. Considering that hypervigilance is defined as an increased sensitivity to threat, these results suggest that hypervigilance causes less vulnerability cues in walking style. This is supported by the results of Study 2, which found that the actual measure of hypervigilance was associated with less of a change in walking style between the two conditions because hypervigilance is associated with less vulnerability cues in general. The decrease in vulnerability cues associated with hypervigilance demonstrates that it is adaptive, in accordance with evolutionary theory (e.g., Cantor, 2009). However, this adaptive function contrasts the subjective experience of hypervigilance, particularly with respect to its association with anxiety and other related symptoms. Therefore, while hypervigilance is considered a maladaptive psychological symptom, it is simultaneously a protective factor against future victimization insofar as it reduces one’s perceived vulnerability.

Given the evident connection between cognition and walking style, it was expected that cognitive variables such as self-perceived vulnerability and perceived impact of victimization would influence the display of vulnerability cues in walking style associated with previous victimization. Partially supporting our hypothesis, it was found that only perceived impact of victimization was significantly related to gait and mediated the relationship between victimization and vulnerability cues. Therefore, subjective evaluations of the impact of victimization are an important factor in the relationship between victimization history and vulnerability cues in walking style.

Study 2 indicated that the relationship between victimization and walking style was contingent on the type of victimization. Specifically, sexual victimization was associated with less vulnerability cues in walking style compared with violent victimization. In addition, the finding that hypervigilance does not mediate the relationship between violent victimization and walking style suggests that hypervigilance plays a unique role in sexual victimization. This is not surprising, given that hypervigilance is more likely to occur after sexual victimization compared to violent victimization (e.g., Cortina & Kubiak, 2006). Altogether, the results of both studies extend previous research by demonstrating the moderating effect of hypervigilance on the relationship between sexual victimization and walking style.

Limitations

While the results of both studies offer some insight into the relationship between past victimization and gait vulnerability, some limitations must be considered. First, the findings might be limited by the use of exclusively female samples. Although women do tend to report more victimization than men, and also tend to self-identify as a victim or survivor of victimization (Cloitre, Tardiff, Marzuk, Leon, & Portera, 1996), victimization can and does occur among men. Future research should examine sex differences in the expression of gait vulnerability related to past victimizations. Furthermore, both studies recruited undergraduate university students and may be biased toward a specific demographic. Thus, it would be informative to replicate these results using other populations, such as community samples. We also had a relatively small sample size in both studies, which may have limited our ability to find existing relationships.

Another limitation is that the difference in walking style between the two conditions may be due to other variables apart from hypervigilance, such as demand characteristics. In particular, the salience of being recorded may have caused participants to walk properly because they wanted to appear presentable for the camera. Indeed, previous research has demonstrated the phenomenon of social facilitation, in which the mere presence of others causes impression management behavior (Uziel, 2010; Zajonc, 1965). Hence, using a more direct manipulation of hypervigilance could provide a more accurate indication of its effects. For example, Johnston and colleagues (2004) induced threat salience directly by instructing participants to imagine walking alone in a dark alleyway. Thus, in order to further corroborate the role of hypervigilance in reducing vulnerability cues in walking style, future research should use more direct methods of manipulation.

Finally, there is also the possibility of measurement bias, particularly for assessing victimization history. Considering that victimization is a sensitive topic, it is possible that victims are unwilling to report these experiences. In line with this idea, it has been suggested that social desirability bias is still a prevalent issue in victimization research (Tourangeau & Yan, 2007). In addition, victimization was assessed using retrospective recall, which is prone to memory biases, both deliberate and accidental (Lalande & Bonanno, 2011). Such methodological issues should therefore be taken into consideration when interpreting the results of the present study.

Implications

The results of both studies extend previous literature documenting the association between victimization history and walking style. Our findings further demonstrated the importance of considering the type of victimization as well as its perceived impact when examining effects on gait. This adds a new dimension of complexity in the process of decoding nonverbal behavior, which future studies in the victim selection literature should take into account. In addition, considering the role of cognitive factors such as perceived impact of victimization on vulnerable gait cues may improve trauma interventions aimed at reducing future victimization. Further research is needed to determine whether reducing the perceived impact of traumatic events through intervention reduces vulnerability cues in walking style.

Our finding that hypervigilance decreases vulnerability cues in walking style suggests that it could be a protective factor against future victimization. The implication of this finding is that it may be advantageous to induce hypervigilant feelings to reduce one’s perceived vulnerability in a situation of potential threat. This contradicts the general understanding of hypervigilance as a maladaptive trait that negatively impacts the mental well-being of PTSD sufferers. However, it is important to distinguish between mild and extreme manifestations of hypervigilance. In particular, extreme forms hypervigilance may cause more severe anxiety-related symptoms while milder forms simply enhance an awareness of threat. Such a distinction suggests that it is possible to acquire the benefits of hypervigilance without having to jeopardize one’s mental health.

An alternative state-of-mind akin to hypervigilance is mindfulness, which specifically refers to an impartial awareness of the present situation and is found to positively impact one’s mental well-being (Bishop et al., 2004). One of the major facets of mindfulness-based therapy is enhancing bodily awareness through practices such as breathing exercises, postural movements, and becoming aware of specific sections on the body (Michalak, Troje, & Heidenreich, 2011). Research on the efficacy of mindfulness-based therapy has indicated that such practices significantly reduce vulnerable gait cues found in depressive patients (Michalak et al., 2011). In a similar vein, our results suggest that enhancing one’s awareness of threat reduces vulnerability cues in walking style. Thus, the present study corroborates the benefits of practicing mindfulness by demonstrating its effects on walking style and one’s likelihood of being selected as a victim. Future research is needed to provide further grounding for the protective function of mindfulness against victimization.

Although this research suggests a protective strategy against future victimization, it does not imply that the victim is to blame for their victimization. On the contrary, a major goal of this study is to understand the various strategies perpetrators use in order to select their victims. The finding that individuals display certain cues to vulnerability that observers may be able to detect supports the notion that social predators can use these cues to assess vulnerability. By investigating the process of victim selection from a perpetrator’s perspective, this research takes an important step toward understanding how to reduce predatory behavior. In this way, decoding walking style cues indicative of vulnerability may also serve to decode the underlying motivations and strategies of those who use them.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Lindsay Fulham, BA, is a graduate student in the Psychology Department at Carleton University (Ottawa, Canada).

Angela S. Book, PhD, is an associate professor in the Psychology Department at Brock University (St. Catharines, Canada) where she researches the relationship between psychopathy and victim selection.

Julie Blais, PhD, is an assistant professor in the Psychology Department at Carleton University (Ottawa, Canada). Her research focuses on psychopathy, dark personalities, and risk assessment.

Mary B. Ritchie, MA, is a doctoral candidate in the Psychology Department at the University of Western Ontario (London, Canada).

Nathalie Y. Gauthier, MA, is a doctoral candidate in the Psychology Department at Brock University (St. Catharines, Canada).

Kimberly Costello, MA, is a doctoral candidate in the Psychology Department at Brock University (St. Catharines, Canada).

References

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author.

Bishop

S. R.

Lau

Shaprio

Carlson

Anderson

N. D.

Carmody

. . . Devins

(2004). Mindfulness: A proposed operational definition. Clinical Psychology: Science and Practice, 11, 230-241. doi:10.1093/clipsy.bph077

Bogels

S. M.

Mansell

(2004). Attention processes in the maintenance and treatment of social phobia: Hypervigilance, avoidance, and self-focused attention. Clinical Psychology Review, 24, 827-856. doi:10.1016/j.cpr.2004.06.005

Bolstad

B. R.

Zinbarg

R. E.

(1997). Sexual victimization, generalized perception of control, and posttraumatic stress disorder symptom severity. Journal of Anxiety Disorders, 11, 523-540. doi:10.1016/S0887-6185(97)00028-5

Book

Costello

Camilleri

J. A.

(2013). Psychopathy and victim selection: The use of gait as a cue to vulnerability. Journal of Interpersonal Violence, 28, 2368-2383. doi:10.1177/0886260512475315

Brand

A. H.

Johnson

J. H.

(1982). Note on reliability of the life events checklist. Psychological Reports, 50(3), 1274.

Cantor

(2009). Post-traumatic stress disorder: Evolutionary perspectives. Australian and New Zealand Journal of Psychiatry, 43, 1038-1048. doi:10.3109/00048670903270407

Cloitre

Tardiff

Marzuk

Leon

A. C.

Portera

(1996). Child abuse and subsequent sexual victimization among female inpatients. Journal of Traumatic Stress, 9(3), 473-482.

Cortina

L. M.

Kubiak

S. P.

(2006). Gender and posttraumatic stress: Sexual violence as an explanation for women’s increased risk. Journal of Abnormal Psychology, 115, 753-759. doi:10.1037/0021-843X.115.4.753

10.

Field

(2013). Discovering statistics using IBM SPSS statistics (4th ed.). London, UK: Sage Publications.

11.

Foa

E. B.

Riggs

D. S.

Dancu

C. V.

Rothbaum

B. O.

(1993). Reliability and validity of a brief instrument for assessing post-traumatic stress disorder. Journal of Traumatic Stress, 6, 459-473. doi:10.1002/jts.2490060405

12.

Garner

M. J.

Mogg

Bradley

B. P.

(2003, September). Orienting and maintenance of attention to facial expressions in social anxiety: Eye movement study. Paper presented at the Annual Conference of the European Association of Behavioural and Cognitive Psychotherapies, Prague, Czech Republic.

13.

Grayson

Stein

M. I.

(1981). Attracting assault: Victims’ nonverbal cues. Journal of Communication, 31, 68-75. doi:10.1111/j.1460-2466.1981.tb01206.x

14.

Gunns

R. E.

Johnston

Hudson

S. M.

(2002). Victim selection and kinematics: A point-light investigation of vulnerability to attack. Journal of Nonverbal Behaviour, 26, 129-158. doi:10.1023/A:1020744915533

15.

Hayes

A. F.

(2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: The Guilford Press.

16.

Johnston

Hudson

S. M.

Richardson

M. J.

Gunns

R. E.

Garner

(2004). Changing kinematics as a means of reducing vulnerability to physical attack. Journal of Applied Social Psychology, 34, 514-537. doi:10.1111/j.1559-1816.2004.tb02559.x

17.

Kimble

Boxwala

Bean

Maletsky

Halper

Spollen

Fleming

(2014). The impact of hypervigilance: Evidence for a forward feedback loop. Journal of Anxiety Disorders, 28, 241-245. doi:10.1016/j.janxdis.2013.12.006

18.

Kimble

M. O.

Frueh

B. C.

Marks

(2009). Does the modified Stroop effect exist in PTSD? Evidence from dissertation abstracts and the peer reviewed literature. Journal of Anxiety Disorders, 23(5), 650–655. doi:10.1016/j.janxdis.2009.02.002

19.

Lalande

K. M.

Bonanno

G. A.

(2011). Retrospective memory bias for the frequency of potentially traumatic events: A prospective study. American Psychological Association, 3, 165-170. doi:10.1037/a0020847

20.

Michalak

Troje

N. F.

Heidenreich

(2011). The effects of mindfulness-based cognitive therapy on depressive gait patterns. Journal of Cognitive and Behaviouralpsychotherapies, 11(1), 13-27. doi:10.1016/j.jpsychores.2010.01.004

21.

Murzynski

Degelman

(1996). Body language of women and judgments of vulnerability to sexual assault. Journal of Applied Social Psychology, 26, 1617-1626. doi:10.1111/j.1559-1816.1996.tb00088.x

22.

Norris

F. H.

(1992). Epidemiology of trauma: Frequency and impact of different potentially traumatic events on different demographic groups. Journal of Consulting and Clinical Psychology, 60, 409-418. doi:10.1037/0022-006X.60.3.409

23.

Runeson

Frykholm

(1981). Visual perception of lifted weight. Journal of Experimental Psychology: Human Perception and Performance, 7, 733-740. doi:10.1037/0096-1523.7.4.733

24.

Runeson

Frykholm

(1983). Kinematic specification of dynamics as an informational basis for person-and-action perception: Expectation, gender recognition, and deceptive intention. Journal of Experimental Psychology: General, 112, 585-615. doi:10.1037/0096-3445.112.4.585

25.

Theriot

M. T.

Dulmus

C. N.

Sowers

K. M.

Johnson

T. K.

(2005). Factors relating to self-identification among bullying victims. Children and Youth Services Review, 27, 979-994. doi: https://doi.org/10.1016/j.childyouth.2004.12.024

26.

Tourangeau

Yan

(2007). Sensitive questions in surveys. Psychological Bulletin, 133, 859-883. doi:10.1037/0033-2909.133.5.859

27.

Uziel

(2010). Look at me, I’m happy and creative: The effect of impression management on behaviour in social presence. Personality and Social Psychology Bulletin, 36, 1591-1602. doi:10.1177/0146167210386239

28.

Van Ameringen

Mancini

Patterson

Boyle

M. H

. (2008). Post-traumatic stress disorder in Canada. CNS Neuroscience & Therapeutics, 14, 171-181. doi:10.1111/j.1755-5949.2008.00049.x

29.

Wheeler

Book

Costello

(2009). Psychopathic traits and perceptions of victim vulnerability. Criminal Justice and Behaviour, 36, 635-648. doi:10.1177/0093854809333958

30.

Zajonc

R. B.

(1965). Social facilitation. Science, 149, 269-274. doi:10.1126/science.149.3681.269