Sage Journals: Discover world-class research

Abstract

Gaslighting, a form of psychological abuse, has received increasing attention in recent years. The current research provides reliability and validity evidence for a new 11-item measure of exposure to gaslighting by a romantic partner: The Gaslighting Relationship Exposure Inventory (GREI). Two studies were conducted with Israeli (N = 509) and American (N = 395) community samples. Participants were involved in a romantic relationship and completed an online questionnaire that included measures of intimate partner violence (IPV) victimization, depression, and relationship satisfaction. In both studies, the GREI items loaded onto a single factor of relationship gaslighting exposure. The scale demonstrated metric gender invariance and was reliable in both men and women. Relationship gaslighting exposure was strongly linked with psychological abuse victimization and weakly linked with other forms of IPV victimization, thus showing strong convergent and discriminant validity. Moreover, relationship gaslighting exposure was associated with greater depression and lower relationship quality, above and beyond other forms of IPV victimization. The results support the reliability and validity of the GREI in both men and women, as well as its utility for the study of gaslighting exposure within relationships. The current research highlights that gaslighting is a distinct form of psychological abuse with unique implications for personal and relational well-being.

Keywords

Gaslighting gender intimate partner violence romantic relationships scale validation

Introduction

Gaslighting is a behavior in which one person undermines another person’s confidence and stability by causing them to doubt their memories, thoughts, and perception of reality (Calef & Weinshel, 1981; Stern, 2007). The recognition of gaslighting in the academic literature dates back over 50 years, initially within the field of psychiatry and medicine (e.g., Barton & Whitehead, 1969). However, in recent years, there has been a remarkable resurgence of interest in gaslighting, resulting in a more comprehensive comprehension of gaslighting within a variety of social contexts. Moreover, there has been a growing public discourse on gaslighting evidenced by Merriam-Webster’s selection of ‘gaslighting' as its Word of the Year in 2022 (Merriam-Webster, 2022).

Gaslighting is frequently discussed in the context of intimate relationships (Graves & Samp, 2021; Klein et al., 2023; Sweet, 2019). In this context, gaslighting involves a variety of manipulative tactics employed by an intimate partner to deny, invalidate, minimize, or ignore the subjective experiences of the other partner, in a way that makes them question their feelings, competence, memory, and/or view of reality. For example, the gaslighter may deny that certain events took place, blame their partner that they do not understand them, or make their partner feel that their feelings are unjustified (Stern, 2007).

The personal consequences of gaslighting can take time to unfold. Individuals who experience gaslighting may, at first, be able to preserve their sense of reality, but as the gaslighting continues, they may start feeling confused and surreal, that reality is neither clear nor logical, and that they are losing their sense of self and independence (Klein et al., 2023). For this reason, gaslighting can threaten one’s psychological well-being and may even be traumatizing (Abramson, 2014; Klein et al., 2023). Moreover, the subtle and elusive nature of gaslighting may discourage gaslightees from seeking help, causing them to remain in a state of continuous harm (Sweet, 2019).

Given the significance of intimate relationships to overall well-being, gaslighting may be particularly impactful in this context. However, the scientific literature on gaslighting in intimate relationships remains relatively scarce. This may be due to the lack of validated measures for gaslighting exposure within romantic relationships. In the current paper, we present a measure of exposure to gaslighting by a relationship partner, the Gaslighting Relationship Exposure Inventory (GREI), and provide initial evidence of its reliability and validity in men and women from two different cultures: Israel and the United States.

Perspectives on gaslighting in romantic relationships

Understanding of the gaslighting construct is still developing and varies between disciplines. In romantic relationships, gaslighting is often considered part of the wider phenomenon of intimate partner violence (IPV) and particularly a specific type of psychological abuse (Abramson, 2014; Stern, 2007). Contrary to direct forms of psychological IPV such as verbal aggression, threats, and humiliation (Straus et al., 1996), gaslighting mainly targets the gaslightees’s sense of what is real and what is not (Miano et al., 2021). As such, some scholars consider gaslighting as a type of micro-aggression (Johnson et al., 2021), an offensive behavior that can be implicit, subtle, and difficult to identify against disadvantaged individuals. Such behaviors can make individuals feel that something is amiss, even though they cannot precisely identify what. Thus, while direct forms of psychological abuse can result in psychological distress and reduced well-being (Potter et al., 2020), gaslighting may be especially detrimental to personal and relational well-being.

Some scholars view gaslighting as one of many tactics of coercive control or “intimate terrorism” (Johnson, 2008; Klein et al., 2023; Sweet, 2019). Coercive control includes controlling, forcible, or threatening behavior that could involve violence or psychological abuse, such as minimizing, denying, and blaming (Stark & Hester, 2019). These features align with the core definition of gaslighting, but the elusive quality of gaslighting makes it difficult to detect, prevent, and address (Stern, 2007). This feature is emphasized by the notion that gaslighting is a dysfunctional communication dynamic in which one party exerts control over another’s knowledge claims, regardless of their validity or truth, making them doubt their sense of reality, which may undermine their ability to make future claims toward the gaslighter (Graves & Spencer, 2021).

Despite the variations in how gaslighting in relationships is understood and classified, several key gaslighting features can be extracted from the existing literature, including (a) denying previous actions or words, (b) eliciting self-doubt by questioning the target’s understanding of reality or memory of events, (c) eliciting self-doubt by casting blame on the target, (d) eliciting feelings of shame and guilt, (e) making the target doubt their abilities, and (f) making the target’s thoughts and feelings seem unjustified (Graves & Spencer, 2021; Klein et al., 2023; Stern, 2007). Exposure to these behaviors from a relationship partner would indicate that one has been subjected to gaslighting.

Gaslighting and gender

Gaslighting in intimate relationships is sometimes presented as a gendered phenomenon ensconced in power differentials. It is described as part of a set of tactics in relationships aimed at gaining control and power over a partner (Graves & Samp, 2021; Graves & Spencer, 2021). Because gaslighting is prevalent in situations of domestic violence, wherein tactics of control are more widely used by men than women (Myhill, 2015), men were theorized to more likely be the gaslighters (Johnson et al., 2021). Moreover, gaslighters can more readily wield gender stereotypes to their advantage against women, for instance, that women are being irrational (Sweet, 2019).

In contrast, Stern (2007) argued that gaslighters and gaslightees can be of any gender. However, she also noted that women are more likely to experience gaslighting in societies where they have gained increased social power. Moreover, while gaslighting is often discussed as having a greater impact on groups with lower social power such as women and LGBTQ+ individuals (Klein et al., 2023), emerging evidence suggests that individuals with lower power can also engage in gaslighting (Graves & Samp, 2021). Recent studies even demonstrated that men experience more gaslighting than women (Bates, 2020; Graves & Samp, 2021). Another possibility is that similar to other forms of IPV (Miano et al., 2021), the consequences of gaslighting are more severe for women, even if they are not exposed to more gaslighting in their intimate relationships than men.

Existing measures of gaslighting

While the construct of gaslighting is still developing, several gaslighting measures have already been created. One of the pioneering instruments was the Gaslight Questionnaire (GQ; Stern, 2007). The GQ includes 20 binary items and is focused on psychological and behavioral reactions to gaslighting as reflected in the inner thoughts and behaviors of the gaslightee (i.e., outcome-focused), rather than the explicit behaviors the gaslightee was exposed to (i.e., exposure-focused). The GQ is not specific to romantic relationships and was not published in a peer-reviewed journal. However, this scale was used in several peer-reviewed studies. In one such study, a likert-scale version of the GQ was used to assess the gaslighting experiences of LGBTQ+ individuals, and confirmatory factor analysis (CFA) indicated that it is unidimensional (Li & Samp, 2023). A more recent gaslighting scale that is specific to the romantic context was developed by Graves and Samp (2021). Designed as a measure of gaslighting experience, the 12-item scale includes three highly correlated factors extracted through exploratory factor analysis (EFA; N = 298, 72.5% female): dyadic dread, partner disbelief, and self-doubt. Similar to the GQ, this scale is mostly concerned with the interactional perceptions and feelings of the gaslightee.

Two additional recent measures are the Victim Gaslighting Questionnaire (VGQ; Bhatti et al., 2023) and the Gaslighting Behaviour Questionnaire (GBQ; Dickson et al., 2023). The VGQ was tested on a sample of 150 young women in Pakistan. The 14-item scale includes two strongly-correlated factors extracted through EFA: peer-disagreement and loss of self-trust. The GBQ includes ten items considering gaslighting scenarios across three contexts: intimate partners, work colleagues and close friends. The items can be used to assess both perpetration and experience. In the intimate partner victim form, two factors were extracted through principal components analysis (N = 386; 77.2% women): promoting self-doubt confusion and blame, and altering recollection of behavior. Both the VGQ and GBQ focus mostly on exposure to certain behaviors of the gaslighter (i.e., exposure-focused).

Our new measure, the GREI, differs from these prior measures in several important ways. Like the VGQ and the GBQ, it takes a more exposure-focused approach by assessing the exposure of individuals to gaslighting behaviors by their partners, rather than the associated outcomes of gaslighting. However, in contrast to these measures, we aim to introduce a multicultural, gender-inclusive measure with robust validity. None of the measures mentioned above were subjected to tests of construct validity. The factorial structure of these measures was determined mainly through EFA without replication, instead of more rigorous confirmatory methods and invariance tests. Moreover, the samples used in these analyses were primarily or exclusively female, which puts the applicability of these scales for men in serious question. Finally, cross-cultural validity was not investigated. The current study aims to fill these gaps.

The Current Research

The current research aimed to assess the factorial structure of the GREI and conduct initial validity tests in two cultures using community samples of individuals who are currently in a romantic relationship. Study 1 employed a sample of Israelis, while Study 2 employed a sample of Americans. In both studies, we examined the psychometric properties of the measure, including whether it is invariant between genders. Then, we gathered evidence for convergent and discriminant validity by examining how the measure correlates with other forms of IPV. We also examined whether exposure to relationship gaslighting is associated with personal (depression) and dyadic (relationship quality) well-being, above and beyond other forms of IPV, and whether these associations are moderated by gender. Finally, we examined whether the factorial structure of the scale is invariant between the Israeli and American cultures.

Study 1 - Israeli sample

Study 1 examined the factorial structure of the GREI using an Israeli sample. In addition, initial tests of construct (convergent and discriminant) validity and predictive utility were conducted. Based on the prior gaslighting measures that were either described by a single factor or by a few highly correlated factors, we expected the GREI to be best described by a single factor. We also expected that exposure to relationship gaslighting will be strongly associated with psychological abuse and weakly associated with physical abuse, economic control, and unconstructive communication. Finally, we expected that exposure to relationship gaslighting would be uniquely associated with greater depression and lower relationship quality, above and beyond other forms of IPV, and that these associations would be stronger in women.

Method

Item generation

We generated 14 items to assess exposure to gaslighting tactics by a relationship partner. The items reflected the six gaslighting features that arose from inspecting the scientific literature on gaslighting as described in the introduction, with each item reflecting at least one of these features. The items were worded as statements describing partner behaviors or tactics (e.g., “My partner often denies saying things that I remember him saying”). Participants were asked to rate how often the statements are true regarding their relationship with their current partner on a scale ranging from 1 (never) to 5 (always). A list of the initial item pool is given in the OSM (https://doi.org/10.17605/OSF.IO/QT5RE).

Participants

Five hundred and nine participants (49.1% men, 50.7% women, 0.2% other) from the Jewish population of Israel were recruited to the study via iPanel (ipanel.co.il), a large online panel service that has over 100,000 members from all walks of Israeli society. The participants were compensated with points on the iPanel website that can be converted to gift certificates. Data collection took place in August 2022. Participants had to be in a romantic relationship for at least 3 months. Due to their low representation in survey companies and language barriers, Israeli Arabs were not included in the study. The study was approved by the IRB of Yezreel Valley College (YVC EMEC, 2022-71).

The participants were 18–70 years old (M = 41.37; SD = 13.77; Median = 38.00) and in a relationship for 0.33–53 years (M = 15.31; SD = 13.42; Median = 10.00). They varied in their sexual orientation (86.4% heterosexual, 3.9% bisexual/pansexual, 2.8% gay/lesbian, 6.9% other/did not report), socioeconomic status (27.3% below average, 33.8% average, 38.8% above average), education level (8.5% elementary or some high school, 18.7% high school diploma, 22.8% post high school, 27.7% bachelor’s degree, 22.4% master’s degree or higher), marital status (76.2% married, 16.5% single (bachelor), 5.7% divorced/separated, 0.2% widowed, 1.4% other/did not report), and Jewish religious group (55.6% secular, 29.0% traditional, 12.8% religious, 2.6% ultra-orthodox)¹.

Procedure and measures

After indicating their informed consent on an electronic form, participants completed an online battery of questionnaires that included the GREI and measures of IPV victimization, relationship quality, and depression.

Exposure to IPV was assessed with the Revised Conflict Tactics Scale (CTS2; Straus et al., 1996). The full CTS2 consists of 39 items assessing victimization and perpetration of physical assault, psychological abuse, and sexual coercion in relationships, as well as negotiation (i.e., constructive communication) and injury. In the current research, we used the physical assault (α = .94), psychological abuse (α = .82), and negotiation (α = .86) subscales only, and asked about victimization, not perpetration. We also included three items assessing economic control (e.g., “My partner forbade me to use shared money or a bank account”; α = .84). Each item was rated on a 0 (never happened) to 6 (more than 20 times) scale concerning the last year.

Relationship quality was assessed with the seven-item short form of the Dyadic Adjustment Scale (DAS-7; Sharpley & Cross, 1982). DAS-7 items assess dyadic consensus, dyadic cohesion, and global dyadic satisfaction (e.g., “calmly discuss something together”), and are rated on varying 6- and 7-point scales (α = .83).

Depression severity was assessed with the 9-item Patient Health Questionnaire (PHQ-9; Kroenke et al., 2001). PHQ-9 items describe common depressive symptoms. Participants rated the frequency in which they experienced each problem in the last two weeks on a 0 (not at all) to 3 (nearly every day) scale (α = .86).

Results

Data analysis strategy

Data analysis was conducted in three stages. First, we examined the factorial validity and measurement invariance of the GREI with a series of confirmatory factor analyses (CFA). Then, we assessed the construct (convergent and discriminant) validity of the GREI by examining its associations with related constructs. Finally, we examined the incremental predictive value of relationship gaslighting exposure (the GREI total score) by examining whether it is associated with depression and relationship quality above and beyond other forms of IPV victimization. The analyses were conducted using Mplus version 8.3 (Muthén & Muthén, 1998-2019) and IBM SPSS version 28. The data and code are given in the OSM. (https://doi.org/10.17605/OSF.IO/QT5RE).

Factor analysis

We expected the 14 items of the GREI to load onto a single factor of relationship gaslighting exposure. However, a single-factor CFA provided only a mediocre fit to the data (see Table 1). Inspection of the modification indices suggested that the residuals of three item pairs were correlated. These item pairs showed substantial content similarity. Therefore, to reduce redundancy, we decided to exclude one item from each pair, based on their loading coefficients. The revised 11-item scale had better fit indices. However, the modification indices for this model pointed to another pair of correlated residuals between two consecutive items (“My partner makes me think that I’m distorting things” and “My partner makes me think that I’m blowing things out of proportion”). These items described somewhat distinct behaviors and feelings. Thus, we opted to retain both of them and allow their residuals to covary. The final model, with one residual covariance, fit the data well (see Table 1). All item loadings were above .60 (see Table S1 in the OSM).

Table 1.

Fit indices: Confirmatory factor analyses; study 1 – Israelis.

Model	χ²	DF	CFI/TLI	RMSEA [90% CI]	SRMR
Model 1: 14 items	465.97***	77	.918/.903	.100 [.091, .109]	.042
Model 2: 11 items	204.10***	44	.952/.940	.085 [.073, .097]	.033
Model 3: 11 items, residual cov^a	164.90***	43	.963/.953	.075 [.063, .087]	.030
Model 3 (MLR estimation)	119.11***	43	.967/.958	.059 [.047, .059]	.030
Model 3 (WLSMV estimation)	222.11***	43	.980/.975	.091 [.079, .103]	.028

Note. CFI = Comparative Fit Index, TLI = Tucker–Lewis index, RMSEA = Root Mean Square Error of Approximation, SRMR = Standardized Root Mean Square Residual.

^aThe residuals of items 9 and 10 are free to covary.

***p < .001.

The CFAs were conducted using Maximum Likelihood (ML) estimation, which assumes multivariate normality. To make sure that our results are valid even when this assumption is eased, we run two sensitivity analyses. First, we re-estimated the final model using ML with robust standard errors (MLR). Second, we ran the final model using robust weighted least squares estimation (WLSMV), which is appropriate for ordered categorical variables. The model still fit the data well by most indices when these alternative estimation methods were used (see Table 1).²

Measurement invariance by gender

We ran the final model separately for men and women to establish baseline models for each gender. These analyses showed that the baseline model for the full sample (i.e., with one residual covariance) is appropriate for both genders. We proceeded to examine measurement invariance between genders using multiple-group CFA. This was done by gradually testing three hierarchical models: configural, metric, and scaler (Little, 1997). Fit indices and comparative tests for these models are given in Table 2. The configural model implies a similar factor structure in both genders (Horn & McArdle, 1992). The good fit of this model supports configural invariance in the current data. The metric model implies that item loadings are equivalent between genders. This model is nested within the configural model, with item loadings constrained to be equal between genders as opposed to being freely estimated. In the current data, the metric model had good fit indices. Moreover, fit did not decrease significantly as a result of the constraints placed on the model (i.e., in comparison to the configural model), thus supporting metric invariance [Δχ²(10) = 11.10, p = .350]. The scalar model is nested within the metric model, with additional gender equality constraints placed on the intercepts of each item. Results showed that the fit of the scalar model was significantly reduced when compared to the metric model [Δχ²(10) = 53.18, p < .001]. Moreover, the CFI of the scalar model was more than .01 smaller than that of the metric model (Cheung & Rensvold, 2002). Thus, full scalar invariance by gender was not supported. This suggests that direct mean comparisons between genders may be suspect because some of the mean differences in the shared variance of the items are not captured by the latent factor.

Table 2.

Fit indices: Measurement invariance by gender, model 3; study 1 – Israelis.

Model	χ²	DF	CFI/TLI	RMSEA [90% CI]	SRMR	Δχ²_(DF)
Men (n = 247)	100.78***	43	.961/.951	.074 [.055, .093]	.035
Women (n = 256)	114.32***	43	.961/.950	.080 [.063, .099]	.035
Configural model	215.10***	86	.961/.950	.077 [.064, .090]	.035
Metric model	226.19***	96	.961/.955	.073 [.061, .086]	.045	11.10₍₁₀₎
Scalar model	279.37***	106	.948/.946	.081 [.069, .092]	.050	53.18₍₁₀₎***
Partial scalar model^a	238.52***	104	.959/.957	.072 [.060, .084]	.045	12.33₍₈₎^b
						23.42₍₁₈₎^c

Note. CFI = Comparative Fit Index, TLI = Tucker–Lewis index, RMSEA = Root Mean Square Error of Approximation, SRMR = Standardized Root Mean Square Residual.

^aThe intercepts of items 9 and 10 are free to vary between genders.

^bCompared with metric model.

^cCompared with configural model.

***p < .001.

To better understand this result, we examined whether our data support partial scalar invariance by gender, meaning that equality of intercepts between genders exists for some, but not all items. For this purpose, we gradually released the gender equality constraints placed on the intercepts in the scalar model, based on modification indices. Our stopping rule was achieving model fit that is not significantly worse than that of the metric model. We reached this goal after freeing up only two intercepts to vary between genders [Δχ²(8) = 12.33, p = .137; see Table 2). These intercepts were associated with the items “My partner makes me think that I’m distorting things” and “My partner makes me think that I’m blowing things out of proportion” (the same two items whose residuals were correlated). Inspection of the gender-specific intercepts showed that these item intercepts were higher in women. This result suggests that there are gender differences in these items that are not related to gender differences in the latent construct of relationship gaslighting exposure.

Reliability and scoring

A total score for the GREI was created by averaging all 11 items. This total score was internally consistent (Cronbach’s α = .93; α = .94 for women, α = .92 for men). Higher scores indicate greater exposure to gaslighting by a relationship partner.

Convergent and discriminant validity

Descriptive statistics and zero-order correlations between all study variables are given in Table 3. Gaslighting exposure was most strongly related to psychological abuse exposure (r = .60, p < .001). Given that gaslighting is considered a form of psychological abuse, this association supports the convergent validity of the GREI. Support for discriminant validity comes from the weak correlations of gaslighting exposure with non-psychological forms of IPV (r = .25, p < .001 & r = .21, p < .001 for physical and economic IPV, respectively) and less constructive communication (CTS2 negotiation; r = −.13, p < .01)³. A very weak correlation was also found with relationship duration, with individuals in longer relationships slightly more likely to report gaslighting by their partners (r = .10, p < .05). Gaslighting exposure was also correlated with greater depression (r = .41, p < .001) and lower relationship quality (r = −.57, p < .001). Overall, these correlations support the construct validity of the GREI.

Table 3.

Descriptive statistics and zero-order correlations between study variables; study 1 – Israelis.

Variable/Statistic	1	2	3	4	5	6	7	8	9
1. Gaslighting exposure
2. CTS Negotiation	−.13**
3. CTS Psychological abuse	.60***	.13**
4. CTS Physical assault	.25***	−.004	.58***
5. CTS Economic control	.21***	.02	.46***	.62***
6. Depressive symptoms	.41***	−.01	.38***	.26***	.30***
7. Relationship quality	−.57***	.35***	−.39***	−.22***	−.21***	−.36***
8. Relationship duration	.10*	−.20***	.003	.02	.01	−.08	−.14**
9. Age	.08	−.22***	−.04	.02	.03	−.06	−.14**	.83***
10. Gender	.18***	−.10*	.16***	.11*	.11*	−.06	−.06	−.02	.03
Mean	1.95	3.34	0.76	0.12	0.12	0.63	22.83	15.31	41.37
SD	0.73	1.39	0.89	0.46	0.50	0.54	5.94	13.42	13.77
Min	1.00	0.00	0.00	0.00	0.00	0.00	3.00	0.33	18.00
Max	4.91	6.00	4.88	4.25	4.00	2.67	36.00	53.00	70.00
N	504	504	504	504	504	505	502	509	509

Note. Gender: Woman = −1, Man = 1.

*p < .05; **p < .01; ***p < .001.

Gender differences

Interestingly, men reported greater gaslighting exposure than women (r = .18, p < .001; Table 3). The lack of full gender scalar invariance, however, means that this finding should be interpreted with caution. We, therefore, conducted an alternative gender comparison by placing a gender equality constraint on the gender-specific means of the latent gaslighting exposure factor in the partial scalar invariance model. This resulted in a significant reduction in fit [χ²₍₁₎ = 24.00, p < .001], suggesting a significant gender difference, with men reporting more gaslighting exposure than women (d = 0.46).

Incremental predictive value

To assess the incremental predictive value of exposure to relationship gaslighting, we examined whether it is associated with depression and relationship quality above and beyond demographic variables and the CTS2 scales. Specifically, we conducted two hierarchical regressions with depression and relationship quality as the outcomes. In each regression, we entered gender, age, and relationship duration as demographic covariates in Step 1. In Step 2, we entered the four CTS2 scales (negotiation, psychological abuse, physical assault, and economic control). In step 3, we entered the GREI score. Finally, in Step 4, we entered the interaction between gender and the GREI score to explore possible gender moderation. Standardized coefficients for the two regressions are reported in Table 4. Tolerance values were all above .30, indicating there was no multicollinearity.

Table 4.

Standardized hierarchical regression coefficients and 95% confidence Intervals; study 1 – Israelis.

Step/Predictor	Depressive symptoms		Relationship quality
Step/Predictor	β [CI]/ΔR²	p	β [CI]/ΔR²	p
Step 1	.011	.154	.025	.006
Gender	−.06 [−.15, .03]	.190	−.06 [−.15, .03]	.177
Age	.04 [−.12, .19]	.658	−.08 [−.23, .08]	.339
Relationship duration	−.11 [−.27, .04]	.160	−.07 [−.23, .08]	.354
Step 2	.175	<.001	.296	<.001
Gender	−.14 [−.22, −.06]	<.001	.05 [−.02, .13]	.155
Age	.08 [−.07, .22]	.297	−.08 [−.22, .05]	.213
Relationship duration	−.17 [−.31, −.02]	.023	.02 [−.11, .15]	.791
CTS Negotiation	−.09 [−.17, −.00]	.044	.40 [.32, .48]	<.001
CTS Psychological abuse	.36 [.26, .46]	<.001	−.46 [−.56, −.37]	<.001
CTS Physical assault	−.04 [−.16, .07]	.479	.07 [−.03, .18]	.187
CTS Economic control	.17 [.07, .28]	.001	−.06 [−.16, .03]	.198
Step 3	.070	<.001	.106	<.001
Gender	−.16 [−.24, −.09]	<.001	.08 [.01, .15]	.018
Age	.06 [−.08, .19]	.420	−.06 [−.18, .06]	.333
Relationship duration	−.17 [−.31, −.04]	.014	.03 [−.09, .15]	.671
CTS Negotiation	−.02 [−.10, .07]	.718	.31 [.24, .39]	<.001
CTS Psychological abuse	.10 [−.02, .22]	.096	−.15 [−.26, −.04]	.006
CTS Physical assault	.02 [−.10, .13]	.777	.00 [−.10, .10]	.996
CTS Economic control	.17 [.08, .28]	<.001	−.07 [−.16, .01]	.102
Gaslighting exposure	.35 [.25, .45]	<.001	−.43 [−.52, −.34]	<.001
Step 4	.000	.918	.000	.789
Gender	−.16 [−.24, −.09]	<.001	.08 [.01, .15]	.018
Age	.06 [−.08, .19]	.423	−.06 [−.18, .06]	.327
Relationship duration	−.17 [−.31, −.03]	.014	.03 [−.09, .15]	.658
CTS Negotiation	−.02 [−.10, .07]	.717	.31 [.24, .39]	<.001
CTS Psychological abuse	.10 [−.02, .23]	.096	−.15 [−.26, −.04]	.006
CTS Physical assault	.02 [−.10, .13]	.772	.00 [−.10, .10]	.988
CTS Economic control	.17 [.08, .28]	<.001	−.07 [−.16, .01]	.099
Gaslighting exposure	.35 [.25, .45]	<.001	−.43 [−.52, −.34]	<.001
Gaslighting exposure × Gender	.00 [−.07, .08]	.918	.01 [−.06, .08]	.789
R ²	.255	<.001	.427	<.001

Note. Gender: Woman = −1, Man = 1.

In the regression predicting depression, the CTS2 scales, entered in Step 2, explained a significant portion of the variance (ΔR² = .175, p < .001), with negotiation (β = −.09, p = .044, 95% CI [−.17, −.00]), psychological abuse (β = .36, p < .001, 95% CI [.26, .46]), and economic control (β = .17, p = .001, 95% CI [.07, .28]) displaying unique associations with depression in the expected directions. The associations of negotiation and psychological abuse with depression, however, diminished to non-significance when relationship gaslighting exposure was entered into the equation in Step 3 (see Table 4). Moreover, relationship gaslighting exposure was significantly associated with greater depression, above and beyond the demographic variables and the CTS2 scales, suggesting incremental predictive value (β = .35, p < .001, 95% CI [.25, .45]). This association was not moderated by gender (β = .00, p = .918, 95% CI [-.07, .08]).

Similar results were found for the regression predicting relationship quality. Step 2, which included the CTS2 scales, explained a significant portion of the variance (ΔR² = .296, p < .001), with negotiation displaying a significant positive association with relationship quality (β = .40, p < .001, 95% CI [.32, .48]) and psychological abuse displaying a significant negative association (β = −.46, p < .001, 95% CI [−.56, −.37]). These associations were somewhat diminished, though remained significant, when relationship gaslighting exposure was entered into the equation in Step 3 (see Table 4). Relationship gaslighting exposure was significantly associated with lower relationship quality, above and beyond the demographic variables and the CTS2 scales, suggesting incremental predictive value (β = −.43, p < .001, 95% CI [−.52, −.34]). As in the prior regression, this association was not moderated by gender (β = .01, p = .789, 95% CI [−.06, .08]).

Discussion

Study 1 indicated that all GREI items load on a single factor of relationship gaslighting exposure, which holds for both men and women. The lack of full scalar invariance indicates that direct gender comparisons on the scale mean may be slightly biased, but the scale’s metric gender invariance and high internal consistency indicate that it could be used in regression and correlation analysis with both men and women. As expected, relationship gaslighting exposure was more strongly associated with psychological IPV than with physical and economic IPV and relationship communication, providing initial support for convergent and discriminant validity. Moreover, relationship gaslighting exposure was associated with greater depression and lower relationship quality above and beyond demographic variables and other forms of IPV and was the strongest predictor of depression and relationship quality out of all IPV types assessed in Study 1.

Interestingly, while the correlates of gaslighting exposure were similar in men and women, men reported being exposed to more gaslighting in their relationships. Research on other forms of IPV produced inconsistent findings regarding gender differences (Ahmadabadi et al., 2021; Chan, 2011). Moreover, given the cultural differences in IPV (White & Satyen, 2015), it was important to cross-validate our results in an additional culture. Thus, we attempted to replicate the findings of Study 1 in a second study with an American sample.

Study 2 - American sample

Study 2 was conducted to replicate and extend the findings of Study 1 in an American sample. We expected the single-factor structure of the GREI to be replicated, as well as the associations of relationship gaslighting exposure with other forms of IPV and its unique associations with depression and relationship quality. Finally, we examined whether the structure of the GREI is invariant between the Israeli and American cultures.

Method

Participants

Three hundred and ninety-five participants (44.6% men, 54.9% women, 0.5% non-binary) were recruited to the study via Cint® (https://www.cint.com/), a survey research firm. Data collection took place in September 2022. Participants had to be in a romantic relationship for at least 3 months. The study was approved by the IRB of Yezreel Valley College (YVC EMEC, 2022-71).

The participants were 19–83 years old (M = 47.21; SD = 15.55; Median = 45.00) and in a relationship for 0.25–60 years (M = 17.55; SD = 15.10; Median = 11.00). They varied in their sexual orientation (90.1% heterosexual, 6.6% bisexual, 2.0% gay/lesbian, 1.3% other), socioeconomic status (28.4% below average, 38.0% average, 33.6% above average), ethnicity (81.0% White or European American, 8.6% Black or African American, 4.8% Hispanic or Latinx, 3.5% Asian, 1.0% American Indian or Alaska Native, 1.0% other), education level (3.0% some high school, 32.4% high school diploma, 15.9% trade school, 31.4% bachelor’s degree, 11.9% master’s degree or higher, 5.3% other), and marital status (73.2% married, 19.0% single, 5.6% divorced/separated, 1.0% widowed, 1.3% other).

Procedure and measures

After signing an electronic informed consent form, participants completed an online battery of questionnaires that included the same measures as in Study 1, including the 11-item GREI, CTS2 (physical assault: α = .96, psychological abuse: α = .90, negotiation: α = .87, economic control α = .88), DAS-7 (α = .87), and PHQ-9 (α = .92). The GREI was translated from Hebrew to English using back-translation (see OSM).

Results

Data analysis strategy

The data analysis strategy was similar to that of Study 1. The data and code are given in the OSM (https://doi.org/10.17605/OSF.IO/QT5RE).

Factor analysis

Fit indices for all models and gender invariance tests are reported in Table 5. Based on the results of Study 1, we expected the 11 items of the GREI to load onto a single factor of relationship gaslighting exposure, with the residuals of items 9 and 10 allowed to covary (Study 1’s Model 3). This model fit the data well by most indices, but the RMSEA indicated only marginal fit (RMSEA = .083). We decided not to try to achieve better fit through post-hoc modification to avoid over-fitting the model, given that our primary goal was to examine whether the factorial structure found with the Israeli sample in Study 1 can reasonably be extended to the American sample in Study 2. Thus, we retained this model for further analysis.

Table 5.

Fit indices: Confirmatory factor analyses and measurement invariance by gender; study 2 – Americans.

Model	χ²	DF	CFI/TLI	RMSEA [90% CI]	SRMR	Δχ²_(DF)
Full sample^a	160.63***	43	.970/.961	.083 [.070, .097]	.024
Men (n = 176)	137.38***	43	.945/.930	.112 [.091, .133]	.036
Women (n = 217)	101.65***	43	.972/.964	.079 [.059, .099]	.025
Configural model	239.03***	86	.960/.949	.095 [.081, .110]	.031
Metric model	253.94***	96	.959/.953	.092 [.078, .105]	.043	14.91₍₁₀₎
Scalar model	287.64***	106	.953/.951	.093 [.081, .106]	.044	33.71₍₁₀₎***
Partial scalar model^b	265.99***	103	.958/.955	.090 [.077, .103]	.043	12.05₍₇₎^c
						26.96₍₁₇₎^d

Note. CFI = Comparative Fit Index, TLI = Tucker–Lewis index, RMSEA = Root Mean Square Error of Approximation, SRMR = Standardized Root Mean Square Residual.

^aThe residuals of items 9 and 10 are free to covary.

^bThe intercepts of items 3, 7 and 10 are free to vary between genders.

^cCompared with metric model.

^dCompared with configural model.

***p < .001.

Measurement invariance by gender

We ran the model separately for men and women to establish baseline models for each gender. Model fit was good for women, but somewhat less so for men, perhaps because of the smaller sample size⁴. We proceeded to examine measurement invariance between genders using multiple-group CFA. As in Study 1, the results supported metric invariance [Δχ²(10) = 14.91. p = .135], but not scalar invariance [Δχ²(10) = 33.71, p < .001]. Thus, we examined whether our data support partial scalar invariance, by gradually releasing the gender equality constraints placed on the intercepts based on modification indices. Our stopping rule was achieving model fit that is not significantly worse than that of the metric model. We reached this goal after freeing up three intercepts to vary between genders [Δχ²(7) = 12.05. p = .099]. These intercepts, which were higher in women, were associated with the items “My partner often denies saying things that I remember him saying”, “My partner makes me feel like I’m impossible to get along with”, and “My partner makes me think that I’m blowing things out of proportion”.

Reliability

The total score of the English GREI (the mean of all items) was internally consistent (Cronbach’s α = .96 for the full sample and women and men separately).

Convergent and discriminant validity

Descriptive statistics and zero-order correlations for all study variables are given in Table 6. Similar to study 1, gaslighting exposure was most strongly related to psychological abuse, supporting convergent validity (r = .72, p < .001). Weaker associations were found with non-psychological forms of IPV (r = .46, p < .001 & r = .44, p < .001 for physical and economic IPV, respectively) and less constructive communication (CTS2 negotiation; r = −.13, p < .05), supporting discriminant validity. Contrary to Study 1, individuals in longer relationships were slightly less likely to report gaslighting by their partners (r = −.11, p < .05). The correlation with age was nonsignificant (r = .09, ns). As in Study 1, gaslighting exposure was associated with greater depression (r = .56, p < .001) and lower relationship quality (r = −.43, p < .001). Overall, these correlations support the construct validity of the GREI.

Table 6.

Descriptive statistics and zero-order correlations between study variables; study 2 – Americans.

Variable/Statistic	1	2	3	4	5	6	7	8	9
1. Gaslighting exposure
2. CTS Negotiation	−.13*
3. CTS Psychological abuse	.72***	−.02
4. CTS Physical assault	.46***	−.04	.67***
5. CTS Economic control	.44***	−.06	.56***	.79***
6. Depressive symptoms	.56***	−.05	.48***	.42***	.39***
7. Relationship quality	−.43***	.34***	−.43***	−.08	−.13*	−.28***
8. Relationship duration	−.11*	−.003	−.17***	−.20***	−.14**	−.30***	−.03
9. Age	.09	−.08	−.13*	−.19***	−.12*	−.28***	−.10*	.75***
10. Gender	.23***	.08	.11*	.15**	.18***	−.02	.08	−.005	.05
Mean	2.08	3.18	0.96	0.26	0.27	0.75	24.88	17.55	47.21
SD	0.95	1.49	1.21	0.73	0.83	0.73	6.57	15.10	15.55
Min	1.00	0.00	0.00	0.00	0.00	0.00	1.00	0.25	19.00
Max	5.00	6.00	5.75	5.33	5.67	2.89	36.00	60.00	83.00
N	395	386	386	386	386	395	394	395	395

Note. Gender: Woman = −1, Man = 1.

*p < .05; **p < .01; ***p < .001.

Gender differences

Similar to Study 1, we conducted a gender comparison by placing an equality constraint on the gender-specific means of the latent gaslighting exposure factor. This resulted in a significant reduction in fit [χ²₍₁₎ = 26.10, p < .001], suggesting a significant gender difference. As in Study 1, men reported greater gaslighting exposure than women (d = 0.55).

Incremental predictive value

To assess the incremental predictive value of relationship gaslighting exposure in the American sample, we examined whether it is associated with depression and relationship quality above and beyond demographic variables and other IPV forms. For this purpose, we ran similar hierarchical regressions as in Study 1 (Table 7). Tolerance values were all above .30, indicating there was no multicollinearity.

Table 7.

Standardized hierarchical regression coefficients and 95% confidence Intervals; study 2 – Americans.

Step/Predictor	Depressive symptoms		Relationship quality
Step/Predictor	β [CI]/ΔR²	p	β [CI]/ΔR²	p
Step 1	.088	<.001	.02	.053
Gender	−.03 [−.13, .06]	.493	.08 [−.02, .18]	.138
Age	−.11 [−.25, .03]	.124	−.19 [−.34, −.04]	.015
Relationship duration	−.19 [−.33, −.05]	.008	.13 [−.03, .28]	.104
Step 2	.233	<.001	.371	<.001
Gender	−.10 [−.18, −.02]	.018	.07 [−.01, .15]	.084
Age	−.10 [−.22, .02]	.113	−.10 [−.22, .02]	.112
Relationship duration	−.11 [−.24, .01]	.069	.02 [−.11, .14]	.807
CTS Negotiation	−.03 [−.12, .05]	.417	.33 [.25, .41]	<.001
CTS Psychological abuse	.35 [.24, .46]	<.001	−.67 [−.78, −.56]	<.001
CTS Physical assault	.03 [−.11, .18]	.646	.42 [.28, .56]	<.001
CTS Economic control	.15 [.02, .29]	.022	−.09 [−.22, .04]	.174
Step 3	.089	<.001	.038	<.001
Gender	−.16 [−.24, −.08]	<.001	.12 [.03, .20]	.005
Age	−.08 [−.20, .03]	.168	−.11 [−.23, .01]	.064
Relationship duration	−.14 [−.25, −.02]	.018	.03 [−.09, .15]	.594
CTS Negotiation	.02 [−.06, .10]	.581	.29 [.21, .37]	<.001
CTS Psychological abuse	.03 [−.10, .16]	.661	−.45 [−.59, −.32]	<.001
CTS Physical assault	.09 [−.04, .23]	.180	.38 [.24, .52]	<.001
CTS Economic control	.11 [−.02, .23]	.087	−.06 [−.19, .07]	.365
Gaslighting exposure	.45 [.33, .56]	<.001	−.30 [−.42, −.19]	<.001
Step 4	.000	.929	.004	.114
Gender	−.16 [−.24, −.08]	<.001	.12 [.04, .20]	.004
Age	−.08 [−.20, .03]	.167	−.11 [−.23, .00]	.057
Relationship duration	−.14 [−.25, −.02]	.018	.03 [−.08, .15]	.573
CTS Negotiation	.02 [−.06, .10]	.590	.28 [.20, .36]	<.001
CTS Psychological abuse	.03 [−.10, .17]	.655	−.43 [−.57, −.29]	<.001
CTS Physical assault	.09 [−.05, .23]	.189	.36 [.22, .51]	<.001
CTS Economic control	.11 [−.02, .23]	.091	−.07 [−.20, .06]	.290
Gaslighting exposure	.45 [.33, .56]	<.001	−.31 [−.43, −.19]	<.001
Gaslighting exposure × Gender	.00 [−.08, .09]	.929	.07 [−.02, .15]	.114
R ²	.410	<.001	.433	<.001

Note. Gender: Woman = −1, Man = 1.

In the regression predicting depression, Steps 1-3 each explained significant portions of the variance. Similar to Study 1, psychological abuse (β = .35, p < .001, 95% CI [24, .46]) and economic control (β = .15, p = .022, 95% CI [.02, .29]) were uniquely associated with greater depression when entered in Step 2, but these associations diminished to non-significance when relationship gaslighting exposure was entered in Step 3 (see Table 7). Relationship gaslighting exposure was significantly associated with greater depression, above and beyond the demographic variables and the CTS2 scales, supporting its incremental predictive value (β = .45, p < .001, 95% CI [.33, .56]). This association was not moderated by gender (β = .00, p = .929, 95% CI [-.08, .09]).

Partly similar results were found for the regression predicting relationship quality. Step 2, which included the CTS2 scales, explained a significant portion of the variance (ΔR² = .371, p < .001), with negotiation displaying a significant positive association (β = .33, p < .001, 95% CI [.25, .41]) and psychological abuse displaying a significant negative association (β = −.67, p < .001, 95% CI [-.78, −.56]) with relationship quality. Surprisingly, a significant positive association was also found between physical assault and relationship quality (β = .42, p < .001, 95% CI [.28, .56]). This association probably reflects a suppression effect, given that in this sample, physical assault had a nonsignificant negative zero-order correlation with relationship quality (r = −.08, ns), and strong positive zero-order correlations with the other CTS2 scales (see Table 6). All of these associations remained significant when relationship gaslighting exposure was entered into the equation in Step 3, but relationship gaslighting exposure was uniquely associated with lower relationship quality above and beyond the demographic variables and the CTS2 scales (β = −.30, p < .001, 95% CI [−.42, −.19]). This result supports again the incremental predictive value of relationship gaslighting exposure. As in all prior regressions, this association was not moderated by gender (β = .07, p = .114, 95% CI [−.02, .15]).

Measurement invariance by culture

To examine whether the factorial structure of the GREI is invariant between the Israeli and American cultures, we conducted an additional multiple-group CFA with the combined data from both samples. The results supported metric invariance, but not scalar invariance (see Table S2 in the OSM). Thus, we examined whether our data support partial scalar invariance, by gradually releasing the culture equality constraints placed on the intercepts. Only after freeing up six intercepts to vary between cultures, we achieved a partial scalar invariance model that was not significantly worse than that of the metric model [Δχ²(4) = 8.77. p = .067]. Specifically, the intercepts of items 1, 4, 8, 9, 10, and 11 were higher in the Israeli sample, whereas the mean of the latent gaslighting exposure factor was greater in the American sample [χ²₍₁₎ = 17.27, p < .001; d = 0.31]. This suggests that some of the behaviors described in the GREI vary between the Israeli and American cultures in ways that are unrelated to differences in relationship gaslighting exposure between these cultures. Thus, direct mean comparisons between American and Israeli samples may be biased⁵. It is also possible that some of the scalar variance was due to language differences.

Discussion

Study 2 replicated and extended the results of Study 2 in an American sample and the English language. It confirmed the single-factor structure of the GREI and its internal consistency in both men and women, despite somewhat lower fit indices for American men. As in Study 1, the scale demonstrated full metric invariance but only partial scalar invariance by gender. The findings concerning construct validity and predictive utility were also replicated, as was the finding that men report more relationship gaslighting exposure than women. Overall, the English GREI performed reasonably well.

General discussion

Gaslighting in intimate relationships has attracted growing attention in recent years (Graves & Samp, 2021; Klein et al., 2023). The current research had two goals: to examine the factorial structure of a new scale for measuring exposure to gaslighting by a relationship partner (GREI) and to provide initial evidence supporting the scale’s validity. In two studies, with Israeli and American samples, the GREI items loaded on a single factor of relationship gaslighting exposure, and the scale was reliable in both men and women. The total score of the GREI had strong links with psychological IPV victimization, as well as strong unique associations with depressive symptoms and low relationship quality. These results lend support for the scale’s validity and utility.

The GREI is the first measure of gaslighting exposure in relationships whose structure and validity were examined on a large number of women and men using confirmatory methods in two cultures. As such, it promises to contribute to the study of gaslighting within relationships.

Gaslighting as a form of psychological abuse

In our research, relationship gaslighting exposure was significantly related to measures of IPV, especially psychological abuse. This is consistent with the notion that gaslighting is a psychologically abusive behavior (Abramson, 2014; Stern, 2007). However, gaslighting differs from overt forms of psychological abuse like threats, curses, humiliations, and insults, where the behavior of the aggressor is perceived clearly as hurtful and painful. Unlike such behaviors, gaslighting can be hard to detect and easy to deny. The gaslighter does not inflict direct psychological pain (e.g., through insults), but rather sows doubt and confusion, thereby inflicting pain in an indirect way. The gaslighter can then benefit from their partner’s doubt and confusion to achieve external goals (e.g., power, protection from criticism, etc.). Thus, gaslighting can be considered an insidious control tactic that is enacted to benefit the gaslighter at the expense of the gaslightee (Bates, 2020; Klein et al., 2023; Stark & Hester, 2019).

The potential of gaslighting to cause unique harm is clearly suggested by our findings. Specifically, relationship gaslighting exposure was associated with greater depression and lower relationship quality above and beyond other forms of IPV, including psychological abuse. This indicates that even actions that do not necessarily look or sound like violence, but sow doubt and confusion, can have deleterious effects on the well-being of partners. Self-doubt may be an important mediating mechanism to consider in future gaslighting research. It may be that the ambiguous nature of gaslighting and the invalidation of the partner’s perceptions and feelings cause self-doubt, which in turn leads to lower well-being (Mirels et al., 2002).

Gaslighting and gender

Across both samples, men reported being exposed to more gaslighting than women. Small departures from scalar invariance did not change this overall picture. This gender effect is consistent with research showing that men report experiencing more emotional abuse (Ahmadabadi et al., 2021) and gaslighting (Bates, 2020; Graves & Samp, 2021) than women. Importantly, however, this finding relies on the subjective reports of the participants and could reflect greater sensitivity to gaslighting behaviors by men. Future research should attempt to replicate this finding by using less subjective methods such as behavioral observations and multi-informant surveys.

One way of contextualizing the gender effect is by considering the community samples of coupled individuals used in the current research. In non-clinical samples, especially ones that include individuals in existing relationships, gaslighting is more likely to be mild. Indeed, in both our studies, the mean level of gaslighting exposure was around 2 on a 1-5 scale. Mild gaslighting may be best understood as part of a broader pattern of situational couple violence, which is characterized by conflictual situations that can escalate into violence (Johnson, 2008). This type of violence tends to be gender symmetrical, although its impact on women tends to be greater (Sillito, 2012). More severe forms of gaslighting, however, which are more likely to be found in clinical samples, may be better understood as coercive control tactics that are more commonly employed by men against women (Johnson, 2008). For example, in a sample that included women who called a domestic violence prevention hotline, most of the women reported that they experienced gaslighting (73.8% reported that their current or former partner did things on purpose to make them feel like they were going crazy) (Warshaw et al., 2014).

Another possible explanation is that women may use gaslighting because of the difficulty of recognizing it as a violent act. Gaslighting may serve as a safer form of aggression that allows women to express violence in a way that does not endanger them, while still maintaining a non-violent image. Using gaslighting may also allow women to achieve personal and relational goals that could not be achieved using direct methods of violence that are more available to men, such as physical and sexual violence (Centers for Disease Control and Prevention, 2022). These possibilities should be investigated in future research.

Limitations and contributions

The current research has several limitations that should be considered. First, it used a correlational, cross-sectional design. Therefore, we cannot draw causal conclusions. It is possible, for example, that preexisting mental difficulties put individuals at a greater risk for IPV victimization, including gaslighting exposure (Zamir et al., 2018). Second, the GREI was validated for quantitative use with community samples. Therefore, it currently does not have cutoff scores for clinically significant gaslighting. Third, the study is based on participants’ self-reports. Some gaslighting tactics, however, may be hard to perceive by the gaslightee (Stern, 2007). This could have suppressed the measured frequency of gaslighting exposure.6 Fourth, certain minority populations were underrepresented in our samples, such as people with disabilities and Israeli Arabs. Studies show that IPV victimization is significantly higher among Arab Israeli women compared with Jewish Israeli women (e.g., Daoud et al., 2020), though no studies to date have examined gaslighting in the Arab community in Israel. Finally, while we examined the association between gaslighting and depression, gaslighting experiences may impact other aspects of mental health. Future studies should examine the diverse mental-health consequences of gaslighting.

Notwithstanding the limitations, the current research is an important step forward in the study of gaslighting in relationships. The GREI is a reliable measure of gaslighting exposure by a relationship partner that can be administered to both men and women. Initial evidence points to its construct validity and predictive utility. Thus, it should facilitate more research on this important phenomenon.

Supplemental Material

Supplemental Material - The gaslighting relationship exposure inventory: Reliability and validity in two cultures

Supplemental Material for The gaslighting relationship exposure inventory: Reliability and validity in two cultures by Tair Tager-Shafrir, Ohad Szepsenwol, Maayan Dvir and Osnat Zamir in Journal of Social and Personal Relationships.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This research was partially funded by Israel Science Foundation Grant 1197/21.

Open research statement

As part of IARR’s encouragement of open research practices,the authors have provided the following information: This research was not pre-registered. The data used in the research are available. The data can be obtained at:

. The materials used in the research are available. The materials can be obtained at: https://osf.io/qt5re/?view_only=bd80f62635f642d99000bf1abed5a98c .

ORCID iDs

Tair Tager-Shafrir

Ohad Szepsenwol

Maayan Dvir

Supplemental Material

Supplemental material for this article is available online.

Notes

References

Abramson

(2014). Turning up the lights on gaslighting. Philosophical Perspectives, 28(1), 1–30. ‏https://doi.org/10.1111/phpe.12046

Ahmadabadi

Najman

J. M.

Williams

G. M.

Clavarino

A. M.

d’Abbs

(2021). Gender differences in intimate partner violence in current and prior relationships. Journal of Interpersonal Violence, 36(1-2), 915–937. ‏https://doi.org/10.1177/0886260517730563

Barton

Whitehead

J. A.

(1969). The gas-light phenomena. The Lancet, 293(7608), 1258–1260. https://doi.org/10.1016/S0140-6736(69)92133-3

Bates

E. A.

(2020). “Walking on egg shells”: A qualitative examination of men’s experiences of intimate partner violence. Psychology of Men & Masculinities, 21(1), 13–24. https://doi.org/10.1037/men0000203

Bhatti

M. M.

Shuja

K. H.

Aqeel

Bokhari

Gulzar

S. N.

Fatima

Sama

(2023). Psychometric development and validation of victim gaslighting questionnaire (VGQ): Across female sample from Pakistan. International Journal of Human Rights in Healthcare, 16(1), 4–18. ‏https://doi.org/10.1108/IJHRH-12-2020-0119

Calef

Weinshel

E. M.

(1981). Some clinical consequences of introjection: Gaslighting. Psychoanalytic Quarterly, 50(1), 44–66. https://doi.org/10.1080/21674086.1981.11926942

Centers for Disease Control and Prevention . (2022). Fast facts: Preventing intimate partner violence. https://www.cdc.gov/violenceprevention/intimatepartnerviolence/fastfact.html#print

Chan

K. L.

(2011). Gender differences in self-reports of intimate partner violence: A review. Aggression and Violent Behavior, 16(2), 167–175. ‏https://doi.org/10.1016/j.avb.2011.02.008

Cheung

G. W.

Rensvold

R. B.

(2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 9(2), 233–255. https://doi.org/10.1207/S15328007SEM0902_5

10.

Daoud

Sergienko

Shoham-Vardi

(2020). Intimate partner violence prevalence, recurrence, types, and risk factors among Arab, and Jewish immigrant and nonimmigrant women of childbearing age in Israel. Journal of Interpersonal Violence, 35(15-16), 2869–2896. https://doi.org/10.1177/0886260517705665

11.

Dickson

Ireland

J. L.

Birch

(2023). Gaslighting and its application to interpersonal violence. Journal of Criminological Research, Policy and Practice, 9(1), 31–46. ‏.https://doi.org/10.1108/JCRPP-07-2022-0029

12.

Graves

C. G.

Samp

J. A.

(2021). The power to gaslight. Journal of Social and Personal Relationships, 38(11), 3378–3386. https://doi.org/10.1177/02654075211026975

13.

Graves

C. G.

Spencer

L. G.

(2021). Rethinking the rhetorical epistemics of gaslighting. Communication Theory, 32(1), 48–67. https://doi.org/10.1093/ct/qtab013

14.

Horn

J. L.

McArdle

J. J.

(1992). A practical and theoretical guide to measurement invariance in aging research. Experimental Aging Research, 18(3-4), 117–144. https://doi.org/10.1080/03610739208253916

15.

Johnson

M. P.

(2008). A typology of domestic violence: Intimate terrorism, violent resistance, and situational couple violence. Northeastern University Press.

16.

Johnson

V. E.

Nadal

K. L.

Sissoko

D. R. G.

King

(2021). “It’s not in your head”: Gaslighting, ‘splaining, victim blaming, and other harmful reactions to microaggressions. Perspectives on Psychological Science, 16(5), 1024–1036. https://doi.org/10.1177/17456916211011963

17.

Kellermann

(1992). Communication: Inherently strategic and primarily automatic. Communication Monographs, 59(3), 288–300. https://doi.org/10.1080/03637759209376270

18.

Klein

Wood

(2023). A qualitative analysis of gaslighting in romantic relationships. Personal Relationships, 30(4), 1316–1340. https://doi.org/10.1111/pere.12510

19.

Kroenke

Spitzer

R. L.

Williams

J. B. W.

(2001). The PHQ-9: Validity of a brief depression severity measure. Journal of General Internal Medicine, 16(9), 606–613. https://doi.org/10.1046/j.1525-1497.2001.016009606.x

20.

Samp

J. A.

(2023). Gaslighting LGBTQ+ individuals: Correlates of gaslighting experiences, gaslighters’ characteristics, and gaslighting techniques. Communication Studies, 74(4), 356–376. ‏https://doi.org/10.1080/10510974.2023.2209605

21.

Little

T. D.

(1997). Mean and covariance structures (MACS) analyses of cross-cultural data: Practical and theoretical issues. Multivariate Behavioral Research, 32(1), 53–76. https://doi.org/10.1207/s15327906mbr3201_3

22.

Merriam-Webster . (2022). Word of the year 2022. In Merriam-Webster.com dictionary. Retrieved March 31, 2023, from. https://www.merriam-webster.com/words-at-play/word-of-the-year

23.

Miano

Bellomare

Genova

V. G.

(2021). Personality correlates of gaslighting behaviours in young adults. Journal of Sexual Aggression, 27(3), 285–298. ‏https://doi.org/10.1080/13552600.2020.1850893

24.

Mirels

H. L.

Greblo

Dean

J. B.

(2002). Judgmental self-doubt: Beliefs about one’s judgmental prowess. Personality and Individual Differences, 33(5), 741–758. ‏https://doi.org/10.1016/S0191-8869(01)00189-1

25.

Muthén

L. K.

Muthén

B. O.

(1998-2019). Mplus user’s guide. 8th ed. Muthen & Muthen.

26.

Myhill

(2015). Measuring coercive control: What can we learn from national population surveys? Violence Against Women, 21(3), 355–375. ‏https://doi.org/10.1177/1077801214568032

27.

Potter

L. C.

Morris

Hegarty

García-Moreno

Feder

(2020). Categories and health impacts of intimate partner violence in the World Health Organization multi-country study on women’s health and domestic violence. International Journal of Epidemiology, 50(2), 652–662. https://doi.org/10.1093/ije/dyaa220

28.

Sharpley

C. F.

Cross

D. G.

(1982). A psychometric evaluation of the Spanier Dyadic Adjustment scale. Journal of Marriage and Family, 44(3), 739–741. https://doi.org/10.2307/351594

29.

Sillito

(2012). Gendered physical and emotional health consequences of situational couple violence for heterosexual married and cohabiting couples. Feminist Criminology, 7(4), 255–281. ‏.https://doi.org/10.1177/1557085111431695

30.

Stark

Hester

(2019). Coercive control: Update and review. Violence Against Women, 25(1), 81–104. https://doi.org/10.1177/1077801218816191

31.

Stern

(2007). The gaslight effect: How to spot and survive the hidden manipulation others use to control your life. Harmony Books.

32.

Straus

M. A.

Hamby

S. L.

Boney-Mccoy

Sugarman

D. B.

(1996). The revised Conflict tactics scales (CTS2). Journal of Family Issues, 17(3), 283–316. https://doi.org/10.1177/019251396017003001

33.

Sweet

P. L.

(2019). The sociology of gaslighting. American Sociological Review, 84(5), 851–875. https://doi.org/10.1177/0003122419874843

34.

Warshaw

Lyon

Bland

P. J.

Phillips

Hooper

(2014). Mental health and substance use coercion surveys: Report from the National Center on Domestic Violence, Trauma & Mental Health and the National Domestic Violence Hotline. Retrieved from. https://ncdvtmh.org/wp-content/uploads/2022/10/NCDVTMH_NDVH_MHSUCoercionSurveyReport_2014-2.pdf

35.

White

M. E.

Satyen

(2015). Cross-cultural differences in intimate partner violence and depression: A systematic review. Aggression and Violent Behavior, 24(10), 120–130. ‏.https://doi.org/10.1016/j.avb.2015.05.005

36.

Zamir

Szepsenwol

Englund

M. M.

Simpson

J. A.

(2018). The role of dissociation in revictimization across the lifespan: A 32-year prospective study. Child Abuse & Neglect, 79, 144–153. https://doi.org/10.1016/j.chiabu.2018.02.001

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.

0.11 MB

0.00 MB