The Influence of Deny,Attack,Reverse Victim and Offender and Insincere Apologies on Perceptions of Sexual Assault

Abstract

DARVO (deny, attack, reverse victim and offender) is a response exhibited by perpetrators to deflect blame and responsibility. When using DARVO, perpetrators deny their involvement in wrongdoing, attack their victims’ credibility, and argue that they are the real victims. The purpose of this study was to measure the influence of DARVO and another manipulative tactic—insincere perpetrator apologies—on observers’ judgments of a victim and perpetrator in a fictional sexual violence scenario. Perpetrator DARVO was experimentally manipulated via fictional vignettes to measure their impact on perceived perpetrator and victim abusiveness, responsibility, and believability. Data from 230 undergraduate students revealed that participants who were exposed to perpetrator DARVO rated the perpetrator as less abusive ( $η_{p}^{2} = . 09$ , 90% CI [0.04, 0.15]), less responsible for the sexual assault ( $η_{p}^{2} = . 02$ , [0.001, 0.06]), and more believable compared ( $η_{p}^{2} = . 03$ , [0.002, 0.07]) to participants who were exposed to a perpetrator who did not use DARVO. DARVO-exposed participants rated the victim as more abusive ( $η_{p}^{2} = . 09$ , [0.04, 0.14]) and less believable ( $η_{p}^{2} = . 08$ , [0.03, 0.14]), and also expressed less willingness to punish the perpetrator and greater willingness to punish the victim. Insincere apologies had minimal impact on ratings. By promoting distrust in victims and less punitive views of perpetrators, DARVO might contribute to rape-supporting outcomes such as victim blaming, greater victim distress, and low rates of rape reporting and perpetrator prosecution.

Keywords

sexual assault offenders reporting/disclosure sexual harassment

Introduction

People who have experienced sexual violence are often subject to greater scrutiny and doubt than victims of other offenses and traumas (Bieneck & Krahé, 2011; Brems & Wagner, 1994; Cromer & Freyd, 2009). Following a sexual assault, many victims who tell others about their assault receive responses containing victim blaming, disbelief, and minimization (Ahrens, 2006; Filipas & Ullman, 2001). Negative responses like these following disclosure are associated with greater self-blame and more pronounced psychological difficulties such as symptoms of posttraumatic stress disorder and depression (Dworkin et al., 2019; Ullman, 2023). Unsupportive and blaming responses can also discourage victims from engaging in future disclosures, reporting the assault, or seeking help from professional services (Ahrens, 2006; Ullman, 2010). In some cases, the prospect alone of possibly being blamed or not being believed prevents victims from reporting or seeking formal help (Alaggia & Wang, 2020; Sabina & Ho, 2014; Walsh et al., 2010). Whether experienced or feared, negative responses to disclosures of sexual assault not only inflict psychological harm but can also create a chilling effect. Victim blaming, expressing doubt about victims’ experiences, and other unsupportive responses effectively silence some victims of sexual violence.

The likelihood of sexual assault victims being disbelieved, blamed, or otherwise held responsible for their assault depends on a variety of factors. Using typically fictional sexual assault scenarios, experimental studies have investigated four categories of factors that influence individuals’ perceptions of victims: characteristics of the victim, characteristics of the perpetrator, characteristics of the sexual assault and other contextual factors, and characteristics relating to the individuals themselves (in-depth reviews of this expansive literature have been completed by Ferrão & Gonçalves, 2015; Gravelin et al., 2018; Persson & Dhingra, 2022; van der Bruggen & Grubb, 2014). One relatively unexamined factor in this line of research is perpetrators’ active attempts to shift perceptions of victims, themselves, and their abusive behavior. Although studies have identified that perpetrators present skewed, victim-blaming narratives (Henning et al., 2005; Lila et al., 2008), little research has investigated how perpetrators might be able to influence observers’ judgments.

Deny, Attack, Reverse Victim and Offender

DARVO (deny, attack, reverse victim and offender) is one way perpetrators might impact perceptions of sexual violence victims and perpetrators. DARVO is an acronym that describes a pattern of responses used by perpetrators of wrongdoing when they are confronted or held accountable for their behavior. First identified by Freyd (1997), perpetrators, particularly those who have committed sexual misconduct, use this tactic to deflect blame and responsibility by denying their involvement in any wrongdoing, attacking their victims’ credibility, and assuming a victimized role, thereby casting the person making the accusation as the offender. For victims, DARVO may be a particularly confusing response to receive from a perpetrator; such confusion surrounding an incident of sexual violence may lead victims to remain silent about their experiences (Veldhuis & Freyd, 1999). For non-victims, DARVO may serve as a way for perpetrators to actively manipulate individuals’ perceptions and beliefs to produce more favorable outcomes for the perpetrators.

Research on DARVO, while still in its preliminary phase, has revealed insights about this response strategy based on findings from college samples. Harsey et al. (2017) asked undergraduates to report the level of DARVO responses they experienced from another person during a confrontation. Although the reasons for the confrontation ranged from milder interpersonal harms (e.g., having a secret betrayed by a close friend or family member) to serious abuses (e.g., sexual assault), DARVO was commonly experienced. Approximately 72% of participants reported experiencing all three components of DARVO (denials, personal attacks, and reversal of victim and offender) from the person they confronted. Individuals who received a greater degree of DARVO responses also indicated experiencing greater self-blame, suggesting that DARVO might be effective in generating confusion regarding culpability. A recent study by Rosenthal and Freyd (2022) adds further evidence of DARVO in the context of sexual violence. In a qualitative analysis of data from 89 undergraduate sexual assault victims who had experienced post-assault contact with their perpetrators, the researchers found that over half of the victims (51%) reported hearing elements of DARVO from their perpetrators. Specifically, these victims indicated that their perpetrators had used denials, attacks, and reversals during post-assault contact. One participant in this study who experienced denial and attacks from her perpetrator shared that she “tried to confront the person about it and they denied it and told me I was wrong and I wasn’t remembering it right . . . he became enraged and said very hurtful things about me so I cut off our connection” (Rosenthal & Freyd, 2022, p. 469). Doubt and self-blame were identified as prominent feelings expressed by the victims who heard DARVO from their perpetrators, suggesting that DARVO responses may play a role in instilling negative feelings among victims whose perpetrators engage in DARVO responses.

While the previous DARVO studies describe the prevalence of this tactic and its associations with victim self-blame (Harsey et al., 2017; Rosenthal & Freyd, 2022), only one study has evaluated DARVO’s effect on third-party observers. Harsey and Freyd (2020) conducted an experiment measuring the influence of DARVO on perceptions of victim and perpetrator believability, responsibility, and abusiveness. In all, 316 undergraduates read fictional vignettes describing an incident of dating violence. The vignettes included post-assault statements from the victim and perpetrator that provided the characters’ point of view of the incident. While all participants read the same victim statement, half of the participants read a perpetrator statement containing DARVO and the second half read a perpetrator statement that recounted the incident without any DARVO. After reading the vignettes, participants rated the victim and perpetrator’s believability, responsibility for the dating violence incident, and abusiveness of their actions. In comparison to those in the control condition, analyses indicated that individuals who were exposed to perpetrator DARVO rated the victim as less believable, more responsible, and more abusive. DARVO also led people to rate the perpetrator as less responsible and less abusive. Contrary to predictions, DARVO penalized perpetrators’ believability, causing individuals in the DARVO condition to rate the perpetrator as less believable in comparison to the perpetrator who did not use DARVO. Still, this cost to perpetrators identified in this study may be small in exchange for damaging observers’ perceptions of victims and for reducing perceptions of perpetrator responsibility and abusiveness. Another prominent finding from this study involved participant gender: as a group, women rated the victim as more believable, less responsible, and less abusive compared to male participants; on the other hand, men’s ratings of perpetrator believability and abusiveness were higher than women’s ratings. No gender differences were found for ratings of perpetrator responsibility.

One unexplored aspect of DARVO is its use with other manipulative perpetrator strategies that seek to produce less negative outcomes for perpetrators. Apologies issued by perpetrators may serve this purpose as prior research demonstrates that apologies can prevent victims from acting aggressively (Ohbuchi et al., 1989). Even apologies containing defensive elements, such as denial, have been found to be associated with less severe perpetrator judgments (Schumann & Dragotta, 2020). In a report on sexual assault in the U.S. Air Force, researchers observed that perpetrators issued apologies for their behavior and express remorse to prevent their peers from reporting the assault (Miller et al., 2018). Apologies may also allow perpetrators to indirectly deny wrongdoing while cultivating an ostensibly sympathetic image. For instance, a wrongdoer may insincerely offer an apology by saying, “I am sorry if I offended you,” implying doubt about whether the offense actually did, in fact, occur. Actor Kevin Spacey used this type of apology in response to allegations that he sexually assaulted fellow actor Anthony Rapp, stating “But if I did behave then as [Rapp] describes, I owe him the sincerest apology. . .” (Spacey, 2017). This type of insincere apology is so ubiquitous that comedian Harry Shearer created the term “ifpology” to describe apologies that employ “if” to create ambiguity surrounding the wrongdoing for which someone appears to be apologizing. While such “ifpologies” are not DARVO, they fail to fully acknowledge that the wrongdoing has occurred and consequently allow the wrongdoer to deflect at least some responsibility. In this way, DARVO and these insincere apologies serve the interests of a wrongdoer or perpetrator who wishes to escape blame and accountability.

Current Study

The purpose of the current study is to replicate the findings from Harsey and Freyd’s (2020) previous experiment in the context of a sexual assault scenario. While Harsey and Freyd (2020) identified how DARVO impacted observers’ perceptions, they did so only in the context of a nonsexual, dating violence scenario. Investigating DARVO in the context of sexual violence is particularly important because it was originally conceptualized as a tactic used by sex offenders (Freyd, 1997). The current study also examined the influence of DARVO on observers’ judgments of victim and perpetrator punishment—should the victim be punished for their actions? Should the perpetrator? Harsey and Freyd (2020) asked their participants similar questions in their second experiment, but only examined differences in punishment judgments between individuals who learned about DARVO and those who did not. The conditions used in this comparison presented all participants with a DARVO-using perpetrator. As such, no analyses were done to measure the effect of DARVO versus No DARVO on judgments of punishment. The current study sought to address this shortcoming. In an additional expansion of Harsey and Freyd (2020), we also investigated DARVO’s effects in combination with manipulative, insincere apologies to better approximate potential real-world instances of DARVO use among perpetrators of sexual violence.

We hypothesized a few key findings using a methodological paradigm very similar to the one used in Harsey and Freyd (2020). First, mirroring findings from Harsey and Freyd (2020), we predicted that individuals exposed to perpetrator DARVO would rate the vignette victim as less believable, more responsible, and more abusive compared to ratings from individuals who were not exposed to perpetrator DARVO. We also hypothesized that DARVO-exposed participants would rate the perpetrator as more believable, less responsible, and less abusive. Despite Harsey and Freyd’s (2020) finding that DARVO perpetrators were judged as less believable than non-DARVO perpetrators, we predicted that, in the current study, perpetrators using DARVO would be rated as more believable. This prediction conforms to the original conceptualization of DARVO (Freyd, 1997), which describes DARVO as an effective tactic for perpetrators. Moreover, the findings in question from Harsey and Freyd (2020) stem from vignettes describing a dating violence incident; it is possible that sexual assault, used in the current study’s vignettes, provides a context in which perpetrators who use DARVO are afforded greater believability. We also anticipated that participants exposed to perpetrator DARVO would be more likely to endorse punishment for the victim and less likely to endorse punishment for the perpetrator.

A second aim was to explore the role of self-serving perpetrator apologies in observer perceptions. Since previous research has found that perpetrators of sexual offenses employ apologies to try and avoid accountability (Miller et al., 2018), we believed this strategy may impact perceptions of victims and perpetrators. Given that very little is known about how such apologies might influence perceptions, we conducted exploratory analyses with this variable.

Finally, given that observer gender is a critical variable in research on perceptions of sexual violence (van der Bruggen & Grubb, 2014), we hypothesized that women, in comparison to men, would rate the victim as more believable, less responsible, and less abusive. Regarding perpetrator judgments, we predicted that men would rate the perpetrator as more believable, less responsible, and less abusive. These predictions cohere with previous research concluding that men tend to assign more blame to rape victims and less blame to sexually violent perpetrators than women do (van der Bruggen & Grubb, 2014).

Our hypotheses can be summarized as follows:

Individuals exposed to DARVO will rate the victim as less believable (H1a), more responsible (H1b), and more abusive (H1c), and will rate the perpetrator as more believable (H1d), less responsible (H1e), and less abusive (H1f).

Individuals exposed to DARVO will be more likely to endorse punishment for the victim (H2a) and less likely to endorse punishment for the perpetrator (H2b).

Compared to men, women will rate the victim as more believable (H3a), less responsible (H3b), and less abusive (H3c), and will rate the perpetrator as less believable (H3d), more responsible (H3e), and more abusive (H3f).

Method

Participants

Participants were 235 undergraduates (M_age = 19.63, SD_age = 2.79) attending a large public university in the Western United States. Most participants were women (n = 169, 71.9%) and approximately a quarter of participants were men (n = 61, 26%). Four individuals (1.7%) identified as nonbinary and one participant did not report their gender identity. Most participants in the study identified as heterosexual (n = 208, 88.5%), while approximately 10% identified as either lesbian or gay (n = 7, 3%), bisexual (n = 16, 6.8%), pansexual (n = 2, 0.9%), or queer (n = 1, 0.4%). One participant did not report their sexual orientation. Just over half of the people in the study were white or Caucasian (n = 134, 57%). The remaining participants were Latine or Chicane (n = 24, 10.2%), East Asian (n = 19, 8.1%), Black or African American (n = 15, 6.4%), biracial or multiracial (n = 15, 6.4%), Southeast Asian (n = 14, 6.0%), and Native American or Native Alaskan (n = 3, 1.2%). Three participants chose to describe their racial identity in their own words and two chose not to report their racial identity at all.

Materials

The current study’s materials were comprised of a series of vignettes describing a fictional incident of sexual abuse and several items evaluating respondents’ perceptions of the vignettes’ perpetrator and victim. The incident involved a male university professor (Professor Smith) groping a female student (Emily) in his office. Prior to reading the vignettes, participants were provided with the following description:

The following are statements about an incident of sexual abuse that occurred at a university. Professor Smith is accused of groping Emily, who is a student in his history class. Both Emily and Professor Smith were interviewed by the university and provided statements about the incident. Read the statements from Emily and Professor Smith below.

The vignettes were written as first-person statements from Emily and Professor Smith. Emily’s statement detailed the incident from her perspective, including a description of the sequence of events and her feelings about the incident. This statement was identical across all conditions. Professor Smith’s statement—the perpetrator vignette—varied in terms of DARVO use (DARVO present vs. not present) and insincere apology use (insincere apologetic statements present vs. not present). The two perpetrator vignettes containing DARVO included each of the three elements of this tactic. In these vignettes, Professor Smith denies ever touching Emily (e.g., “I never once touched her”), attacks her credibility (“. . ., she kept trying to flirt with me and asked if I would give her paper a good grade”; “I don’t know why she made that up”), and reverses victim and offender roles (“My reputation has been hurt because of this, and that’s not fair. I’ve been teaching for over 20 years and this one false accusation might end my career”). Vignettes containing manipulative use of insincere apologies included phrases ostensibly expressing some remorse about the way Emily felt (e.g., “I am very sorry if I made Emily uncomfortable during our meeting”). Importantly, however, the vignettes containing elements of insincere apologies did not admit to any wrongdoing; the use of the word “if” in the apologetic phrase indicates that the apologizer is not convinced that any wrongdoing has occurred at all. This was done to model real-world insincere apologies issued by individuals and institutions that offer regret about victims’ perceived experiences but fail to acknowledge committing any intentional harm.

Emily’s statement was 208 words in length, while Professor Smith’s statement varied in length depending on condition. In the No DARVO/No Apology condition, no statement from Professor Smith was included at all and the preceding description was altered to state, “Professor Smith has chosen not to make a statement on the matter.” This was done to again approximate real-world examples of responses to allegations in which individuals or institutions decide to provide no comments at all. The remaining statements by Professor Smith were 210 words (DARVO/No Apology), 231 words (DARVO/Apology), and 181 words (No DARVO/Apology). These statements were created for the purpose of the current study and appear in full in Supplemental Appendix A.

After reading the statement by Emily and Professor Smith (or, in the No DARVO/No Apology condition, just the statement by Emily), participants responded to a brief series of questions about Emily and Professor Smith. Participants were asked to report the extent to which they perceived both Emily and Professor Smith to be believable (0 = Not at all believable, 3 = Very believable), the fictional characters’ responsibility for the incident (0 = Not at all responsible, 3 = Very responsible), and the abusiveness of their behavior (0 = Not at all abusive, 3 = Very abusive). Participants then reported whether they believed Emily should be disciplined or punished for her behavior (Yes, No, Not sure) and whether Professor Smith should be disciplined or punished for his behavior (Yes, No, Not sure). Responses to items about Emily represent individuals’ perceptions of the victim, while responses to items about Professor Smith represent perceptions of the perpetrator.

Procedure and Design

Study materials were compiled with several other brief measures from other psychology researchers in the department and administered to the university’s human subjects pool. This battery of measures, known as the General Survey, was approved by the university’s institutional review board. Students were able to participate in the General Survey by accessing the university’s research study sign-up webpage, where the survey was listed as an hour-long online study that could be taken in exchange for course credit. To prevent self-selection, the survey’s listing did not provide specific information about the content of the current study’s measures. Participants were therefore unaware of the study’s content when they signed up to participate in the survey, which minimizes threats to external validity caused by self-selection (Freyd, 2012). The survey was conducted online.

Participants who elected to take the survey were randomly assigned to one of the four conditions. Conditions were defined by the fictional perpetrator’s (i.e., Professor Smith’s) use or absence of DARVO and insincere apologetic statements. This resulted in a 2 × 2 experimental design: No DARVO/No Apology (n = 55), DARVO/No Apology (n = 63), No DARVO/Apology (n = 58), and DARVO/Apology (n = 59). In total, 122 students were randomly assigned to a condition containing perpetrator DARVO and 113 were randomly assigned to a condition with no perpetrator DARVO.

Data Analysis Plan

Data were first evaluated for missing cases and normality. Following these steps, a two-way multivariate analysis of variance (MANOVA) was computed to test for the effects of experimental condition and participant gender on victim and perpetrator believability, responsibility, and abusiveness. A one-way MANOVA was then conducted to specifically compare the DARVO and No DARVO conditions on perceptions of victim and perpetrator. The same test was then done to examine the effect of insincere apologies on participants’ perceptions. Responses on the items about victim and perpetrator punishment were then analyzed using chi-square tests and z-tests to evaluate the effect of condition and participant gender.

Results

Data Screening

Missing data were very minimal in the current study. Two of the dependent variables (i.e., the item asking whether the victim should face discipline or punishment and the item about perpetrator abusiveness) were missing a total of two datapoints. The remaining dependent variables contained either 0 or 1 missing datapoint. Missing data were excluded listwise from analyses. The continuous dependent variables (i.e., victim and perpetrator believability, responsibility, and abusiveness) were checked for distribution normality. All skewness and kurtosis values for these variables were between −2 and 2, which suggests some but not a concerning level of skew (Hair et al., 2010). As a result, no transformations were done to the data for analyses. Data were checked for multivariate outliers through the calculation of Mahalanobis’ distance. A few outliers were identified and individually examined but none appeared suspicious. Data showed no serious signs of multicollinearity. Although the Box’s M test was significant for the multivariate tests in the current study, therefore suggesting a violation of equality of covariances, MANOVA tests tend to be robust against violations of this assumption when group sizes are over 30 (Allen & Bennett, 2008). All other assumptions for the main analyses were met (e.g., independence of observation, adequate sample size).

Main Analyses

A two-way MANOVA was conducted to test the effect of experimental condition and participant gender on the continuous dependent variables. For this test, participants with complete data identifying as men or women (n = 227) were included. Nonbinary (n = 4) participants and the one participant who did not report their gender were omitted from this test as their group size was too small for adequate analysis. Results from the omnibus multivariate test indicated the dependent variables differed significantly by condition, F(18, 605.77) = 3.97, p < .001, Wilk’s Λ = .73, partial η² = .10, 90% CI [0.05, 0.12], and by participant gender, F(6, 214) = 4.0, p < .001, Wilk’s Λ = .90, partial η² = .10, [0.03, 0.15] (see Table 1). Tukey’s post hoc tests were then computed among the four conditions to determine mean differences. Statistically significant mean differences were found between the conditions for victim believability, victim abusiveness, perpetrator believability, and perpetrator abusiveness. Generally, the No DARVO/No Apology condition was less advantageous for the perpetrator, leading to the lowest perpetrator believability and highest perpetrator abusiveness scores. In contrast, the two conditions containing DARVO (DARVO/No Apology and DARVO/Apology) produced higher perpetrator believability scores and lower perpetrator abusiveness score. The No DARVO/No Apology and No DARVO/Apology conditions led to more favorable outcomes for victims, producing higher victim believability scores and lower victim abusiveness scores.

Table 1.

Means, Standard Deviations, and Between-Subject Results for Multivariate Analysis of Variance (MANOVA) Predicting Perceptions of Victim and Perpetrator from Study Condition and Participant Gender.

	Study Condition M (SD)				Multivariate Test
Dependent Variable	No DARVO, No Apology	DARVO, No Apology	No DARVO, Apology	DARVO, Apology	F-statistic	$η_{p}^{2}$	$η_{p}^{2}$ 90% CI
Victim
Believability	2.66 (0.55)^a	2.37 (0.71)^b	2.75 (0.47)^bc	2.30 (0.73)^ac	8.02***	.10	[0.04, 0.16]
Responsibility	0.45 (0.97)	0.75 (1.04)	0.49 (1.0)	0.58 (0.89)	1.60	.02	[0, 0.05]
Abusiveness	0.26 (0.59)^a	0.88 (1.03)^ab	0.25 (0.63)^b	0.58 (0.75)	9.81***	.12	[0.05, 0.18]
Perpetrator
Believability	0.79 (0.84)^ab	1.22 (0.83)^a	1.09 (0.79)	1.21 (0.70)^b	3.87*	.05	[0.01, 0.09]
Responsibility	2.49 (1.03)	2.37 (0.80)	2.67 (0.64)	2.31 (0.78)	3.37*	.04	[0.004, 0.09]
Abusiveness	2.79 (0.45)^ab	2.20 (0.86)^a	2.47 (0.78)^c	2.01 (0.97)^bc	9.88***	.12	[0.05, 0.18]
	Participant Gender M (SD)		Multivariate Test
	Men	Women	F-statistic	$η_{p}^{2}$	$η_{p}^{2}$ 90% CI
Victim
Believability	2.30 (0.69)	2.60 (0.62)	11.07***	.05	[0.01, 0.10]
Responsibility	0.89 (1.08)	0.46 (0.92)	9.56**	.04	[0.01, 0.09]
Abusiveness	0.77 (0.90)	0.40 (0.76)	11.89**	.05	[0.01, 0.11]
Perpetrator
Believability	1.34 (0.81)	0.99 (0.78)	9.43**	.04	[0.01, 0.09]
Responsibility	2.13 (0.90)	2.58 (0.76)	14.87***	.06	[0.02, 0.12]
Abusiveness	2.05 (0.96)	2.48 (0.77)	13.63***	.06	[0.02, 0.12]

Note. *p < .05, **p < .01, ***p < .001. DARVO = deny, attack, reverse victim and offender.

The interaction term between condition and participant gender was not significant, F(18, 605.77) = 1.15, p = .302, Wilk’s Λ = .91, partial η² = .03, 90% CI [0, 0.03]. Tests of between-subject effects revealed that victim believability, victim abusiveness, perpetrator believability, perpetrator responsibility, and perpetrator abusiveness were significantly different by condition; only victim responsibility did not vary by condition. All dependent variables in the MANOVA differed significantly by participant gender in the expected direction (H3a–H3f). See Table 1 for the specific statistics relating to this analysis.

Since we were primarily interested in the effect of DARVO on observer perceptions, we conducted a one-way MANOVA comparing responses between the DARVO and No DARVO conditions on the continuous dependent variables. The two DARVO conditions (DARVO/No Apology and DARVO/Apology) were combined to create a single DARVO condition, while the two conditions without DARVO (No DARVO/No Apology and No DARVO/Apology) were combined to create a single No DARVO condition. Given that participant gender did not significantly interact with condition in the previous MANOVA, gender was omitted from this analysis. As such, data from participants of all genders were included. Table 2 contains the results for the test comparing the DARVO and No DARVO conditions. Scores on all variables except victim responsibility (H1b) were statistically different between DARVO and No DARVO conditions. Specifically, individuals in the DARVO condition rated the victim as less believable (H1a) and more abusive (H1c), and rated the perpetrator as more believable (H1d), less responsible (H1e), and less abusive (H1f). Figure 1 displays the results of this analysis in a bar chart format. A one-way MANOVA was computed to compare the Apology and No Apology conditions as well. Although the multivariate test was significant, F(6, 224) = 3.07, Wilk’s Λ = .92, p = .006, partial η² = .08, 90% CI [0.01, 0.11], the tests of between-subject effects concluded that only one of dependent variables differed significantly between the Apology and No Apology conditions: perpetrator abusiveness scores from participants who viewed the perpetrator’s apologies (M = 2.26, SD = 0.91) were lower than scores from those who viewed perpetrator statements without such apologies (M = 2.49, SD = 0.75), F(1, 229) = 4.34, p = .038, partial η² = .019, [0.001, 0.06].

Table 2.

Results of One-Way Multivariate Analysis of Variance (MANOVA) Comparing DARVO and No DARVO Conditions.

Dependent Variable	DARVO M (SD)	No DARVO M (SD)	F	p	$η_{p}^{2}$	$η_{p}^{2}$ 90% CI
Victim
Believability	2.34 (0.72)	2.72 (0.51)	20.13	<.001	.081	[0.03, 0.14]
Responsibility	0.66 (0.96)	0.46 (0.98)	2.25	.135	.01	[0, 0.04]
Abusiveness	0.73 (0.91)	0.25 (0.61)	21.26	<.001	.085	[0.04, 0.14]
Perpetrator
Believability	1.20 (0.77)	0.95 (0.81)	6.03	.015	.026	[0.002, 0.07]
Responsibility	2.35 (0.79)	2.59 (0.84)	4.85	.029	.021	[0.001, 0.06]
Abusiveness	2.13 (0.92)	2.64 (0.66)	23.17	<.001	.092	[0.04, 0.15]

Note. This table contains data from all participants, n = 235. DARVO = deny, attack, reverse victim and offender.

Figure 1.

Bar graphs of average victim and perpetrator believability (Panel A, favorable evaluations), responsibility, and abusiveness (Panel B, unfavorable evaluations) by study condition.

The combined DARVO (n = 122) and No DARVO (n = 113) conditions were also used to compare responses on the items asking whether the fictional victim and perpetrator should be disciplined or punished for their behavior. Among participants in the No DARVO condition, 92.9% (n = 104) indicated the victim should not be punished for her actions, while 6.3% (n = 7) were not sure; only one participant in the No DARVO condition reported that the victim should face punishment. In comparison, 66.9% of participants (n = 81) in the DARVO condition believed the victim should not be punished. Nearly 10% (n = 12) of participants in the DARVO condition reported the victim should be punished and an additional 23.1% (n = 28) were not sure if the victim should face punishment. The same question was asked about the perpetrator—should the perpetrator be disciplined or punished for his behavior? In the No DARVO condition, 82.3% (n = 93) of participants believed the perpetrator should be punished, 3.5% (n = 4) believed the perpetrator should not be punished, and 14.2% (n = 16) were not sure. Participants in the DARVO condition were slightly more reluctant to report that the perpetrator should be punished. Of these participants, 60.7% (n = 74) indicated the perpetrator should be punished, 11.5% (n = 14) believed the perpetrator should not be punished, and 27.9% (n = 34) were not sure if the perpetrator should be punished. Chi-square tests of independence revealed that beliefs about perpetrator punishment (χ² (2, N = 235) = 13.87, p = .001) and victim punishment (χ² (2, N = 233) = 24.46, p < .001) depended on study condition. Pairwise z-tests were used to identify differences in response proportions between conditions. For both victim and perpetrator punishment variables, each response option (i.e., Yes, No, Not Sure) was significantly different between conditions. The DARVO condition led participants to be more willing or unsure to punish the victim (H2a) and less willing or more unsure to punish the perpetrator (H2b). This suggests that exposure to DARVO shifts observers’ beliefs about disciplinary action following a sexual assault in a way that is more favorable toward perpetrators. See Figure 2 for the bar graph of these results and corresponding z and p values for each pairwise comparison.

Figure 2.

Bar graphs showing participant responses to victim punishment (top) and perpetrator punishment (bottom) items.

Judgments of victim and perpetrator punishment were also evaluated by participant gender. As with the two-way MANOVA, only participants who identified as woman or man were examined since the number of participants identifying outside this binary was too small for analysis. Although the chi-square test for victim punishment approached statistical significance (χ² (2, N = 228) = 5.12, p = .08), pairwise z-tests indicated a significant difference in men (68.9%) and women’s (82.6%) No responses: z = −2.26, p = .024. Neither “Yes” (men: 8.2%; women: 4.8%) nor “Not sure” (men: 23%; women: 12.6%) responses to the victim punishment item were statistically different between men and women. The chi-square test for perpetrator punishment was overall significant, χ² (2, N = 230) = 18.03, p < .001. Women (78.1%) were more likely than men (49.2%) to indicate that the perpetrator should be punished, z = −4.24, p < .001. Conversely, men (37.7%) were more likely than women (16%) to indicate that they were “Not sure” if the perpetrator should be punished, z = 3.53, p < .001. Endorsement of the “No” response to this item did not statistically differ between men (13.1%) and women (5.9%), z = 1.79, p = .07.

Responses on the victim and perpetrator punishment items were also analyzed to compare responses between the Apology/No Apology conditions. Results from a chi-square test of independence showed that responses on neither victim punishment (χ² (2, N = 233) = 4.06, p = .131) nor perpetrator punishment (χ² (2, N = 235) = 2.22, p = .329) varied between the two conditions.

Discussion

DARVO is a tactic used to undermine confidence in victims. Past research on DARVO suggests that it is commonly experienced by victims of both sexual violence (Rosenthal & Freyd, 2022) and victims of a wide array of interpersonal wrongdoings (Harsey et al., 2017). The current study aimed to build on previous findings examining how DARVO influences third-party observers’ perceptions of victims and perpetrators (Harsey & Freyd, 2020) in the context of sexual violence. In support of hypotheses H1a and H1c, we found that individuals exposed to perpetrator DARVO, compared to those who were not exposed to DARVO, rated a victim of sexual assault as less believable and more abusive. Supporting hypotheses H1d, H1f, and H1e, individuals exposed to DARVO rated the perpetrator as more believable, less responsible for the assault, and less abusive. Exposure to DARVO also influenced individuals’ beliefs about punishment, leading a greater number of DARVO-exposed participants to endorse victim punishment or to indicate greater uncertainty about whether the victim deserved punishment (H2a). Likewise, individuals exposed to DARVO were less likely to agree that the perpetrator should be punished and were more likely to be unsure if the perpetrator should be punished (H2b).

The current study’s findings largely replicate results from Harsey and Freyd (2020) in a new context. In both the current study and Harsey and Freyd’s (2020) study, DARVO-exposed individuals rated a fictional vignette victim as less believable and more abusive, and rated the perpetrator as less abusive. Two differences between the current study and Harsey and Freyd’s (2020) study emerged. While participants in Harsey and Freyd’s (2020) study rated the perpetrator who used DARVO as less believable, findings from the present experiment indicate that DARVO boosted perpetrator believability. This discrepancy may be explained by the different types of violence investigated by the two studies. It is possible that DARVO is simply more effective in boosting perceptions of perpetrator believability in the context of sexual violence but not domestic violence. Alternatively, people may be more willing to endorse sexual violence perpetrators’ innocence more generally. The second difference between the two studies involves victim responsibility. Results from the current study found no statistical difference in victim responsibility ratings between the conditions (H1b), whereas Harsey and Freyd (2020) found that DARVO increased victim responsibility. As nonsignificant findings are difficult to interpret, we can only speculate as to why this difference emerged. One reason is that average victim responsibility ratings in both conditions are near the lowest end of the item’s response scale in the current study, suggesting a floor effect on this variable. It is possible that college-aged people are relatively unwilling to implicate a sexual violence victim as responsible for their victimization, although this seems to contrast with widely held cultural myths about sexual violence that implicate victims as deserving or responsible for being assaulted (Burt, 1980). Additional DARVO research is needed to fully contextualize this null finding.

In addition to replicating most of the findings from Harsey and Freyd (2020), the current study also revealed how DARVO affects beliefs about victim and perpetrator punishment. Exposure to perpetrator DARVO made people less likely to believe that the perpetrator should be disciplined or punished. Endorsement of perpetrator punishment was 82% among individuals not exposed to DARVO but only 60.7% among individuals exposed to DARVO. The inverse effect was found for beliefs about victims. Belief that the victim should not be punished was endorsed by 92.9% of individuals who were not exposed to DARVO but endorsed by only 66.9% of individuals who were exposed to DARVO. These findings suggest DARVO undermines third-party observers’ trust in the victim’s innocence and produces greater uncertainty surrounding the perpetrator’s guilt. The argument made by DARVO—that the perpetrator not only did nothing wrong, but is also a victim of a false accusation—is clearly effective when it comes to decisions about punishment. This has potentially significant implications for victims of sexual violence who seek justice through the legal system. Perpetrators who mount a DARVO defense may be successful in, for example, convincing jury members that they are not deserving of punishment.

In the current study, we also explored the role of insincere apologies that implicitly denied any wrongdoing. Results suggested that a perpetrator’s use of such apologies is minimally effective. The only variable in which ratings differed between the apology and non-apology conditions was perpetrator abusiveness, which was lower among individuals who were exposed to the perpetrator apology.

All hypotheses relating to participant gender were supported (H3a–H3f). In line with previous research finding that men typically attribute more blame to victims of sexual violence and less blame to perpetrators (van der Bruggen & Grubb, 2014), our data revealed that men rated the fictional victim as less believable, more responsible, and more abusive than did the women in the study. Likewise, compared to women, men rated the perpetrator as more believable, less responsible, and less abusive. Although we did not make specific predictions regarding gender and judgments of victim and perpetrator punishment, we discovered an effect of gender on individuals’ willingness to endorse no victim punishment. Approximately 69% of men reported they believed the victim should not be punished, but a greater proportion of women—82.6%—indicated the victim should not be punished. In a similar vein, more women reported that the perpetrator should be punished (78.1% of women vs. 49.2% of men) and more men were unsure if the perpetrator should be punished (37.7% of men vs. 16% of women). The discrepancies between men and women in their willingness to both punish the perpetrator and willingness to not punish the victim may reflect gender differences in rape myth acceptance. Men reliably score higher on measures of rape myth acceptance than do women (Suarez & Gadalla, 2010). One rape myth that may be particularly salient in the current study is the belief that it is common for women to falsely accuse men of rape. Belief in this particular myth might explain why some men in the current study not only expressed less willingness to punish the perpetrator, but also demonstrated a greater willingness to punish the victim.

Overall, the current study indicates that DARVO shifts observers’ perceptions of both victims and perpetrators of sexual violence. By undermining trust in victims’ narratives, DARVO enables perpetrators to generate confusion about the circumstances of the assault—is the victim lying or exaggerating? Is the perpetrator a victim of a false accusation? Who is to be believed? This has clear implications for victims of sexual violence, who often do not disclose or report the assaults they experience out of fear of not being believed (Mennicke et al., 2021; Spencer et al., 2017). Findings from the current study underscore the potentially silencing effect that DARVO may have on victims.

Limitations and Future Directions

The current study meaningfully contributes to the scientific research on DARVO and offers insight into how perpetrators of sexual violence can manipulate observers’ perceptions of victims and perpetrators. It has, however, a few notable limitations. The participants in this study were undergraduate students with relatively little heterogeneity in racial identity, education, age, nationality, and other demographic variables. The findings presented here may therefore not be generalizable among certain populations. It is possible some groups may find DARVO to be less compelling while this tactic may be even more influential among others. Future research on this issue should prioritize sampling from a large, diverse community population to more thoroughly explore DARVO’s impact on people’s perceptions. In addition, a greater number of men should be purposefully sampled since the number of male-identified participants in the current study is comparatively low, which may have reduced power to find gender effects.

A second limitation of the current study is the relatively modest effect sizes found for some mean differences between the DARVO and No DARVO conditions. For instance, the partial η² for the mean difference in perpetrator responsibility between conditions was .021, a value that is generally categorized as a small effect size (Aron et al., 2009). It is possible that the brevity of the experimental exposure to DARVO—a short, single-paragraph statement containing only a few phrases representing DARVO—is responsible for the modest effect sizes found in the current study. Experimental stimuli that better approximate real-world instances of DARVO may produce larger effect sizes. For instance, future research may use visual and audio experimental stimuli containing elements of DARVO, such as a video showing an interview or conversation with a perpetrator. Related to the current study’s experimental stimuli, it is important to note that the fictional perpetrator in the vignettes belongs to a higher social status (professor) than the victim (student). We did not vary the status of the victim or perpetrator, so we are unable to conclude whether social status influenced our findings. We do know, however, that past experiments on DARVO’s influence on perceptions, very similar in design to the current study, have found similar results using fictional victims and perpetrators who were peers (Harsey & Freyd, 2020). While it is possible that a perpetrator’s higher social status renders DARVO more effective in shifting individuals’ perceptions, it seems that DARVO does not offer an advantage only to high-status perpetrators. More research is needed to explore this issue more carefully. We also note that, in the current study, the items asking participants to indicate whether the victim and perpetrator should be punished are open to broad interpretations. Future research would benefit from either specifying the type of punishment (e.g., sanctions imposed by the university, legal charges) or by including a variety of consequences ranging in severity from which participants could select the option they believe to be the most suitable.

Finally, DARVO may also meaningfully interact with individuals’ preexisting beliefs and attitudes about sexual violence, such as rape myth acceptance. It is possible that individuals already high in rape myth acceptance, for instance, are more likely to be swayed by DARVO than individuals who are low in this construct. Although the current study did not measure such attitudes, additional research on DARVO’s influence would substantially contribute to this area of study by including scales for rape myth acceptance, sexism, and other salient attitudes.

Conclusions

DARVO enables perpetrators of sexual violence to present observers with a distorted reality. The three elements of DARVO work in conjunction to argue that allegations of sexual assault are only exaggerated anecdotes or fabrications from vindictive individuals seeking to cause reputational damage. The present findings suggest that, to some extent, this manipulative tactic can be compelling among third-party observers. The ability to sway even a small number of individuals to adopt a perspective that is more skeptical of victims and more sympathetic toward perpetrators may be a perpetrator’s most effective method of deflecting blame and escaping accountability. As a consequence of cultivating disbelief in victims, DARVO likely denies victims safety, well-being, and justice. In the context of sexual violence, DARVO therefore acts as an agent of rape culture. For this reason, it is crucial to disrupt DARVO. One effective way to do this is through education: people who are informed about DARVO are less affected by the technique (Harsey & Freyd, 2020). By offering education about the function of DARVO and its impact on people’s perceptions, the influence of this insidious tactic can be reduced. Naming and interrupting DARVO could serve as a vital act of resistance against rape culture and other social forces that foster doubt in victims’ narratives. As a policy, institutions that offer care and support for sexual violence victims could incorporate learning modules about DARVO into their training practices and could additionally provide information about DARVO to victims seeking help. Mental health clinicians in particular should be familiar with DARVO and its associated research in order to help victims understand or prepare for the potentially confusing responses they may encounter following an assault.

Supplemental Material

sj-docx-1-jiv-10.1177_08862605231169751 – Supplemental material for The Influence of Deny, Attack, Reverse Victim and Offender and Insincere Apologies on Perceptions of Sexual Assault

Supplemental material, sj-docx-1-jiv-10.1177_08862605231169751 for The Influence of Deny, Attack, Reverse Victim and Offender and Insincere Apologies on Perceptions of Sexual Assault by Sarah J. Harsey and Jennifer J. Freyd in Journal of Interpersonal Violence

Footnotes

Declaration of Conflicting Interests

Jennifer J. Freyd receives book royalties, speaking fees, and consulting fees. Sarah J. Harsey receives a postdoctoral fellowship stipend from the Center for Institutional Courage. There are no other competing interests to declare.

Funding

This work was funded by the Center for Institutional Courage.

ORCID iD

Sarah J. Harsey

Supplemental Material

Supplemental material for this article is available online.

Author Biographies

Sarah J. Harsey, PhD, is a Postdoctoral Research and Education Fellow at the Center for Institutional Courage. Her research primarily investigates issues relating to sexual violence. She is especially interested in the role of DARVO (Deny, Attack, Reverse Victim, and Offender), the objectification of women, and sexual scripts in sexual violence perpetration.

Jennifer J. Freyd, PhD, is the Founder and President of the Center for Institutional Courage, Professor Emerit at the University of Oregon, and an Adjunct Professor of Psychiatry and Behavioral Sciences, Stanford University. Freyd is known for her theories of betrayal trauma, institutional betrayal, institutional courage, and DARVO.

References

Ahrens

C. E.

(2006). Being silenced: The impact of negative social reactions on the disclosure of rape. American Journal of Community Psychology, 38, 263–274. https://doi.org/10.1007/s10464-006-9069-9

Alaggia

Wang

(2020). “I never told anyone until the #metoo movement”: What can we learn from sexual abuse and sexual assault disclosures made through social media? Child Abuse & Neglect, 103, 104312. https://doi.org/10.1016/j.chiabu.2019.104312

Allen

Bennett

(2008). SPSS for the health and behavioural sciences. Thomson.

Aron

E. N.

Coups

E. J.

(2009). Statistics for psychology (5th ed.). Pearson.

Bieneck

Krahé

(2011). Blaming the victim and exonerating the perpetrator in cases of rape and robbery: Is there a double standard? Journal of Interpersonal Violence, 26(9), 1785–1797. https://doi.org/10.1177%2F0886260510372945

Brems

Wagner

(1994). Blame of victim and perpetrator in rape versus theft. The Journal of Social Psychology, 134(3), 363–374. https://doi.org/10.1080/00224545.1994.9711741

Burt

M. R.

(1980). Cultural myths and supports for rape. Journal of Personality and Social Psychology, 38, 217–230. https://doi.org/10.1037//0022-3514.38.2.217

Cromer

L. D.

Freyd

J. J.

(2009). Hear no evil, see no evil? Associations of gender, trauma history, and values with believing trauma vignettes. Analyses of Social Issues and Public Policy, 9(1), 85–96. https://doi.org/10.1111/j.1530-2415.2009.01185.x

Dworkin

E. R.

Brill

C. D.

Ullman

S. E.

(2019). Social reactions to disclosure of interpersonal violence and psychopathology: A systematic review and meta-analysis. Clinical Psychology Review, 72, 101750. https://doi.org/10.1016/j.cpr.2019.101750

10.

Ferrão

M. C.

Gonçalves

(2015). Rape crimes reviewed: The role of observer variables in female victim blaming. Psychological Thought, 8(1), 47–67. https://doi.org/10.5964/psyct.v8i1.131

11.

Filipas

H. H.

Ullman

S. E.

(2001). Social reactions to sexual assault victims from various support sources. Violence and Victims, 16(6), 673–692.

12.

Freyd

J. J.

(1997). II. Violations of power, adaptive blindness and betrayal trauma theory. Feminism & Psychology, 7(1), 22–32. https://doi.org/10.1177/0959353597071004

13.

Freyd

J. J.

(2012). A plea to university researchers. Journal of Trauma & Dissociation, 13(5), 497–508. https://doi.org/10.1080/15299732.2012.708613

14.

Gravelin

C. R.

Biernat

Bucher

C. E.

(2018). Blaming the victim of acquaintance rape: Individual, situational, and sociocultural factors. Frontiers in Psychology, 9, 2422. https://doi.org/10.3389/fpsyg.2018.02422

15.

Hair

Black

W. C.

Babin

B. J.

Anderson

R. E.

(2010). Multivariate data analysis (7th ed.). Pearson Education.

16.

Harsey

S. J.

Freyd

J. J.

(2020). Deny, Attack, Reverse Victim and Offender (DARVO): What is the influence on perceived perpetrator and victim credibility? Journal of Aggression Maltreatment & Trauma, 29, 897–916. https://doi.org/10.1080/10926771.2020.1774695

17.

Harsey

S. J.

Zurbriggen

E. L.

Freyd

J. J.

(2017). Perpetrator responses to victim confrontation: DARVO and victim self-blame. Journal of Aggression Maltreatment & Trauma, 26, 644–663. https://doi.org/10.1080/10926771.2017.1320777

18.

Henning

Jones

A. R.

Holdford

(2005). “I didn’t do it, but if I did I had a good reason”: Minimization, denial, and attributions of blame among male and female domestic violence offenders. Journal of Family Violence, 20, 131–139. https://doi.org/10.1007/s10896-005-3647-8

19.

Lila

Herrero

Gracia

(2008). Evaluating attribution of responsibility and minimization by male batterers: Implications for batterer programs. The Open Criminology Journal, 1, 4–11. https://doi.org/10.2174/1874917800801010004

20.

Mennicke

Bowling

Gromer

Ryan

(2021). Factors associated with and barriers to disclosure of a sexual assault to formal on-campus resources among college students. Violence Against Women, 27(2), 255–273. https://doi.org/10.1177/1077801219889173

21.

Miller

L. L.

Keller

K. M.

Wagner

Marcellino

Donohue

A. G.

Greathouse

S. M.

Matthews

Ayer

(2018). Air force sexual assault situations, settings, and offender behaviors. RAND Corporation.

22.

Ohbuchi

Kameda

Agarie

(1989). Apology as aggression control: Its role in mediating appraisal of and response to harm. Journal of Personality and Social Psychology, 56(2), 219–227. https://doi.org/10.1037//0022-3514.56.2.219

23.

Persson

Dhingra

(2022). Attributions of blame in stranger and acquaintance rape: A multilevel meta-analysis and systematic review. Trauma Violence & Abuse, 23, 795–809. https://doi.org/10.1177/1524838020977146

24.

Rosenthal

M. N.

Freyd

J. J.

(2022). From DARVO to distress: College women’s contact with their perpetrators after sexual assault. Journal of Aggression Maltreatment & Trauma, 31(4), 459–477. https://doi.org/10.1080/10926771.2022.2055512

25.

Sabina

L. Y.

(2014). Campus and college victim responses to sexual assault and dating violence: Disclosure, service utilization, and service provision. Trauma Violence & Abuse, 15(3), 201–226. https://doi.org/10.1177%2F1524838014521322

26.

Schumann

Dragotta

(2020). Is moral redemption possible? The effectiveness of public apologies for sexual misconduct. Journal of Experimental Social Psychology, 90, 104002. https://doi.org/10.1016/j.jesp.2020.104002

27.

Spacey

(2017, October 29). [Image attached] [Tweet]. Twitter. https://twitter.com/KevinSpacey/status/924848412842971136

28.

Spencer

Mallory

Toews

Stith

Wood

(2017). Why sexual assault survivors do not report to universities: A feminist analysis. Family Relations, 66, 166–179. https://doi.org/10.1111/fare.12241

29.

Suarez

Gadalla

T. M.

(2010). Stop blaming the victim: A meta-analysis on rape myths. Journal of Interpersonal Violence, 25(11), 2010–2035. https://doi.org/10.1177%2F0886260509354503

30.

Ullman

S. E.

(2010). Talking about sexual assault: Society’s response to survivors (1st ed.). American Psychological Association.

31.

Ullman

S. E.

(2023). Correlates of social reactions to victims’ disclosures of sexual assault and intimate partner violence: A systematic review. Trauma Violence & Abuse, 24, 29–43. https://doi.org/10.1177/15248380211016013

32.

van der Bruggen

Grubb

. (2014). A review of the literature relating to rape victim blaming: An analysis of the impact of observer and victim characteristics on attribution of blame in rape cases. Aggression and Violent Behavior, 19(5), 523–531. https://doi.org/10.1016/j.avb.2014.07.008

33.

Veldhuis

C. B.

Freyd

J. J.

(1999). Groomed for silence, groomed for betrayal. In Rivera

(Ed.), Fragment by fragment: Feminist perspectives on memory and child sexual abuse (pp. 253–282). Gynergy Books.

34.

Walsh

W. A.

Banyard

V. L.

Moynihan

M. M.

Ward

Cohn

E. S.

(2010). Disclosure and service use on a college campus after an unwanted sexual experience. Journal of Trauma & Dissociation, 11(2), 134–151. https://doi.org/10.1080/15299730903502912

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.01 MB