Development and Validation of the Bystander Engagement Scale for Witnessed Intimate Partner Violence: A South Korean Version

Abstract

Bystanders play a crucial role in assisting and rescuing victims of intimate partner violence (IPV) or sexual assault, thereby preventing further harm. Consequently, an increasing amount of research has focused on improving bystander interventions in such situations. While many existing scales that measure bystanders’ willingness to intervene and their behaviors are likely tailored to Western settings and primarily focused on preventing sexual assaults, pursuing diversity to make these scales applicable to a variety of contexts is necessary. This study aimed to develop and validate a scale from South Korean samples that measure the likely level of bystander engagement in witnessed IPV situations, named the Bystander Engagement Scale for Witnessed Intimate Partner Violence (BESW-IPV). Preliminary items, developed to describe various IPV situations that a bystander might witness, underwent content validity testing through expert review and cognitive focus group interviews. Afterward, the scale was administered to 311 participants from the target population in South Korea. The scale’s factor structure was assessed through exploratory factor analysis (EFA) and confirmed through confirmatory factor analysis (CFA). The reliability and validity were rigorously assessed. The final version of the BESW-IPV comprised 28 items. Through EFA, “direct IPV indicators” and “indirect IPV indicators” were identified. Together, these factors accounted for 76.10% of the total variance. CFA affirmed that the final model offers an acceptable fit. The scale’s convergent and discriminant validities were also well established. The Cronbach’s alpha and the McDonald’s omega values were 0.98 and 0.99, respectively. In contrast to many existing tools that measure bystander intervention predominantly within the confines of Western educational settings, we believe that the BESW-IPV can be applied in broader contexts, especially in patriarchal environments. Further research is required to translate and validate this tool in different cultural contexts.

Keywords

intimate partner violence bystander intervention crisis intervention measurement prevention

Introduction

Intimate partner violence (IPV) is a serious global concern as it leads to devastating consequences for victims, including physical injury, fear, anxiety (Kwon & Park, 2019), posttraumatic stress disorder (Gerber et al., 2012), and even homicide by perpetrators (Kim et al., 2021). IPV is often believed to be hidden, as it is more likely to occur in private places; however, recent studies have revealed that many incidents are witnessed by others, so-called bystanders (Hamby et al., 2016; Taylor et al., 2019). A bystander is any individual who has the potential to witness an incident of IPV or sexual assault or can be present in the situation; thus, they are capable of helping the victims and preventing further risks (Banyard et al., 2014). Despite their pivotal roles, many bystanders are reluctant to intervene in IPV situations for various reasons, such as not knowing how to intervene or being unsure of what is really happening (Park & Ko, 2020). Given that bystanders’ non-interventions could reinforce perpetrators’ behaviors and inhibit victims’ help-seeking (Gracia et al., 2018), enhancing bystanders’ intervention willingness and behavior in IPV situations is essential. Thus, there have been ongoing efforts by researchers to explore the bystander engagement mechanism and develop programs that facilitate bystanders’ involvement (Mujal et al., 2019).

Measurement of Bystander Intervention

Measuring bystander behavior is essential when exploring its mechanisms or evaluating the effectiveness of programs promoting it; however, it is considered challenging for several reasons. One of the reasons is the uncertainty that individuals have encountered instances of IPV or sexual assault in their lifetimes. Despite this, certain scales have measured respondents’ actual behaviors or actions with the assumption that they have witnessed such incidents. This approach can threaten the validity of findings. Participants might indicate they did not intervene because they have never witnessed such an event (Mennicke et al., 2022). In response to this challenge, some studies have been, alternatively, designed to measure bystander efficacy (Stojanov et al., 2021), bystander readiness to help and act (Rothman et al., 2018), and behavioral intention to intervene (Wong et al., 2019) as predictors of bystander behaviors. According to a recent review, most existing bystander intervention scales tend to measure the bystanders’ intent, willingness, and likelihood to intervene rather than their actual behaviors (Mennicke et al., 2022).

Measuring bystanders’ willingness to intervene, however, is also challenging as it is highly contextual. Bystanders’ likelihood to intervene in a situation varies depending on their gender (Casper et al., 2021), whether they know the victims or perpetrators personally or not (Katz et al., 2015), how seriously they perceive the situation (Bennett et al., 2017), or their cultural context (Kamimura et al., 2016). Thus, existing tools often present specific contexts to reliably measure bystanders’ willingness to intervene, rather than considering all these contextual factors within a single tool. For instance, some scales measure bystanders’ intentions to help people they know, such as friends (Borsky et al., 2016; Cascardi et al., 2021), while others focus on bystanders’ intentions to help strangers (Banyard et al., 2014). In addition, the measures often contextualize the nature of the violence, focusing on situations involving sexual violence (McMahon et al., 2017), dating violence (Borsky et al., 2016), teen relationship abuse (Edwards et al., 2021), or IPV against women (Gracia et al., 2018).

A Necessity for Diversity in Bystander Intervention Measurement

Existing bystander measurements lack diversity, even though bystander roles are pivotal regardless of the context. According to Mennicke et al. (2022), approximately 45.8% of existing scales have been validated using samples from educational institutions, mainly colleges, the majority of whom were white, young, heterosexual, and female. Consequently, most existing bystander scales adopted have been limited in their application to marginalized populations, including people of color; lesbian, gay, bisexual, and transgender; and people living in non-Western cultural contexts. Indeed, some items within the existing scales may not be feasible in certain cultures that are not familiar with house parties and drinking cultures (e.g., “if a friend had too much to drink, I asked them if they needed help getting home from the party” [Cascardi et al., 2021, p. NP4862]). Moreover, certain items may not be applicable in contexts where offering assistance to strangers or intervening in personal affairs is not commonly accepted or practiced (e.g., “ask a stranger who looks very upset at a party if they are okay or need help” [Banyard et al., 2014, p. 108]). Furthermore, most existing scales primarily target the prevention of sexual assault incidents and are not well-suited for addressing IPV issues outside the scope of educational institutions.

Both the IPV phenomenon and bystanders’ expected roles are likely to be shaped by culture (Kamimura et al., 2016). South Korea had long been a patriarchal society under the influence of Confucianism, which stressed the strong idea of the dominance of men over women (Park & Chesla, 2007). Although South Korean society has changed and become Westernized, men’s controlling attitudes and behaviors toward women seem to be taken for granted and even understood as a way of expressing their love and care for women (Kwon & Park, 2019; Son & Cho, 2016). A survey of 1,082 representative samples of unmarried South Korean females revealed that 62.6% of respondents had experienced some form of control exerted by their partners. This included surveilling their interactions with people (45.1%), regulating their attire (36.9%), compelling them to quit their jobs (29.5%), and intervening and controlling their personal daily schedules (25.5%). However, approximately 38.9% of respondents did not perceive these controlling behaviors as acts of abuse, and 32.1% even interpreted them as demonstrations of their partners’ love for them (Son & Cho, 2016). Such cultural perceptions can deter bystanders from intervening in others’ intimate relationships, as they too may perceive that there are no issues despite the victims’ suffering.

Moreover, literature often highlights the prevalent victim-blaming tendencies in South Korea regarding IPV issues (Park & Ko, 2020; Park & Woo, 2023). A recent study identified three types of bystander groups in South Korea: hesitant helpers needing justification, outsiders denouncing the couple, and active interveners (Park & Woo, 2023). The second group, outsiders denouncing the couple, comprised mostly young and single men who showed blaming attitudes toward the couple, especially the victims, and justified the violence against the victims (Park & Woo, 2023). In addition, according to a survey by Son and Cho (2016), approximately 51% of bystanders responded that the victim was also at fault for not leaving the perpetrator, a perspective that was the most prevalent response. Therefore, the tendency to blame the victim may be one of the most prevalent and significant barriers for the South Korean population intervening in others’ IPV situations.

Present Study

The current study developed and validated a scale, Bystander Engagement Scale for Witnessed IPV (BESW-IPV), designed specifically for the South Korean cultural context. Recognizing and understanding the situation poses significant challenges for the South Korean population in terms of executing bystander roles (Park & Ko, 2020). Hence, the scale aims to capture a range of bystanders’ intended engagement in situations with various potential and actual threats and harms of IPV.

Methods

Study Design

To develop and validate the scale, this study followed two steps: instrument development and instrument validation, referring to the methods suggested by De Vellis (2003). In the instrument development phase, preliminary items were compiled based on theories and literature reviews. The items were then modified through content validity testing and pre-testing. In the instrument validation phase, reliability and validity were tested by administering the scale to the sample population, and the final version of the BESW-IPV was confirmed.

Instrument Development

Composition of Preliminary Items

To develop the preliminary items, various IPV situations that bystanders could witness and suspect were identified. This study constructed five theoretical domains of IPV situations that could be noticed by bystanders: physical abuse, psychological abuse and threats, sexual assault, coercive control, and a victim’s reaction to a crisis. Within these domains, coercive control is defined as a systematic pattern of behavior in which the abuser deliberately seeks to exert and maintain dominance over the victim using a range of credible threats within the context of IPV, even when the victim perceives these actions negatively (Hamberger et al., 2017). Moreover, bystanders often recognize or foresee the crisis based on the victim’s reactions, which might involve seeking help, revealing the signs of abuse, or showing a wish to return to the toxic relationship (Park & Ko, 2020). Consequently, we incorporated the domain of the victim’s response to a crisis as a scenario in which a bystander might become aware of or witness IPV.

The first and corresponding author of this study reviewed the existing literature separately to identify the specific items in the five IPV domains. In all, 30 items were initially developed by the authors after several discussions. Specifically, we developed six items for the domain of physical abuse, five items for psychological abuse and threats, five for sexual assault, seven for coercive control, and seven to assess victim’s reactions to a crisis. A 5-point Likert scale was initially used to measure a level of bystander’s intended engagement in 30 situations: 1 (not intervening at all), 2 (not intervening), 3 (neutral), 4 (intervening), and 5 (completely intervening).

Content Validity Testing

Content validity, a method used to assess whether the items adequately measured the domain of the concept, was then performed (Boateng et al., 2018). The preliminary items were evaluated by both experts and target populations.

Each item was assessed using a 4-point Likert scale from 1 (not relevant) to 4 (highly relevant) by a panel of five experts in the following fields: gender health (n = 2), violence and crime (n = 1), family health (n = 1), and psychiatric mental health (n = 1). They were asked to leave feedback on the items that scored less than four points. Any content validity index (CVI) below .78 was supposed to be deleted (Polit & Beck, 2006); however, all items scored ≥.80. Several items were revised based on the feedback from the experts. For instance, vague expressions (e.g., inappropriate behavior) were revised to concrete expressions (e.g., violent behavior). Sentences described from the victim’s perspective (e.g., A person fears that the partner will be angry) were revised to sentences observable from the bystanders’ viewpoint (e.g., A person explains trivial things to their partner, fearing that the partner will be angry).

Cognitive-focused group interviews (FGIs) were conducted to determine whether the items represented the actual bystander experience of the target population. Interviewees were recruited through online flyers posted on the authors’ social networking sites. In all, 20 volunteers contacted the researchers and provided informed consent by agreeing to participate. The FGIs were conducted in four groups, with each group comprising five interviewees. The interviewees’ ages ranged from 19 to 43 years, and the mean age was 25.85 ± 7.78 years. Among the 20 interviewees, 40% (n = 8) were male and 60% (n = 12) were female. In terms of marital status, 75% (n = 15) were single and 25% (n = 5) were married. Of the interviewees, 50% (n = 10) reported that they had witnessed IPV situations involving their acquaintances, while 45% (n = 9) did not; the remaining 5% (n = 1) reported that they were not sure.

The FGI consisted of two rounds, both conducted on the same day and moderated by the corresponding author, a psychiatric nurse specialist. During the first round, the interviewees were asked to freely discuss IPV situations that they witnessed in close proximity or to share their thoughts and opinions about a bystander’s role. The main questions included the following: (1) Have you ever witnessed, heard of, or suspected situations involving IPV? (2) If yes, what made you suspect or judge that IPV occurred? (3) How did you behave at that time? Why did you behave in that manner? (4) What are the thoughts and feelings that contributed the most to the bystander’s decision to intervene? (5) What are the thoughts/feelings that contributed the most to the bystander’s decision not to intervene? All interviews were audio-recorded and transcribed near-verbatim.

By comparing the preliminary items with the transcripts of FGI discussions, we checked for any missing situations in which a bystander noticed IPV. Through this process, seven items were added to the preliminary list. Specifically, items were incorporated that described witnesses’ accounts of victims’ experiences after leaving their partners and subsequently enduring stalking (e.g., A person’s ex-partner, who had been violent, visits the person without notice, terrifying them). Additional situations were introduced where victims directly request help from those around them (e.g., A person asks me for help to get out of a situation of violence with their partner), bystanders inadvertently become catalysts for the victims’ abuse (e.g., When I am with a person, the person’s partner checks whether I am of the gender that the person is interested in; in that case, the relationship becomes suspicious), and perpetrators attempt to involve bystanders in the abuse (e.g., The person’s ex-partner, who had been violent, asks me about the person’s whereabouts). Ultimately, the total number of items increased to 37.

Pre-Testing Items

In the second round of cognitive FGI, the interviewees completed a survey comprising preliminary items, which took approximately 5 minutes on average to complete. Thereafter, they were asked several questions, including the following: (1) Is the description of this questionnaire sufficient to guide responses? (2) Are there any questions that are difficult or ambiguous to understand? (3) Are there any difficult questions to answer? (4) Are there questions that are difficult to agree with?

Through discussions, three main instances of feedback were adopted, and the scales were revised and confirmed accordingly. First, Likert response options were modified based on feedback from interviewees. They found it difficult to distinguish between point 2 (not intervening) and point 1 (not intervening at all). In addition, they mentioned that point 3 (neutral) was vague and they were uncertain about what “neutral” implied in terms of the level of engagement. Consequently, we changed the 5-point Likert scale to a 4-point scale, consisting of 1 (no engagement), 2 (a little engagement), 3 (substantial engagement), and 4 (full engagement). Second, we provided some examples in the introduction of the scale due to difficulties expressed by interviewees regarding understanding levels of engagement. For instance, there was confusion about whether “no engagement” signified ignoring the situation altogether or simply listening to the victims without taking action. Finally, the relationship between bystanders and victims was predetermined as a general acquaintance (e.g., co-workers, classmates/seniors/juniors, neighbors, and general friends) rather than family members or close friends. Interviewees mentioned that their behaviors and reactions were highly dependent on their relationships with the victims. Groups discussed each item individually and revised the grammar and expressions of some items when necessary to enhance the clarity of the items’ meanings.

Instrument Validation

Participants and Procedure

The eligibility criteria for the survey administration were as follows: (1) aged 20–64 years and (2) able to read and understand the items of the instrument and respond adequately. A stratified random sampling strategy that considered the gender and age of the participants was used. Ten strata were created for male and female groups: 20s, 30s, 40s, 50s, and 60s. The goal was to recruit 330 participants, with 33 from each stratum, assuming a 10% withdrawal rate. The data were collected online by a professional survey company, Embrain (https://embrain.com), which had over 1,600,000 South Korean sample panels. Interested panel members voluntarily opted into this study by clicking on a provided URL. This link led them to an electronic consent form, and upon providing their consent, they were redirected to the online survey. In total, 311 completed the survey, satisfying the necessary sample size of 300 for factor analysis, according to Clark and Watson (1995). The panel members who completed the survey were compensated with 5,000 Korean won (3.84 US dollars).

Data Analysis

Participant characteristics were summarized by performing descriptive statistics. Item analysis was then performed, and each item’s mean and standard deviation (SD), skewness, kurtosis, and communality were calculated. Before conducting exploratory factor analysis (EFA), Kaiser–Meyer–Olkin (KMO) statistics and Bartlett’s test of sphericity were performed to determine whether the selected items were suitable for EFA. Thereafter, for EFA, the factor structure was assessed using maximum likelihood estimation and oblique rotation strategies. In addition, parallel analysis and the scree test were conducted to validate the factor structure underlying the items. Following this, confirmatory factor analysis (CFA) was performed by utilizing bootstrapped maximum likelihood estimation, performed 5,000 times. Several model-fit indices, including the $χ^{2} / df$ , comparative fit index (CFI), Tucker–Lewis index (TLI), standardized root mean square residual (SRMR), and root mean square error of approximation (RMSEA), were calculated to determine whether the structured model fits the data well (Hair et al., 2019). The reference values were $χ$ ²/df ≤ 3.0, CFI ≥ .90, TLI ≥ .90, SRMR ≤ .08, and RMSEA ≤ .10, respectively (Kang, 2013). The model was adjusted to improve the fit indices, and the final model was subsequently confirmed. Thereafter, we assessed convergent validity within the factors using construct reliability (CR) values. Discriminant validity between factors was ascertained through the heterotrait-monotrait ratio (HTMT). The final items were validated by evaluating their internal consistency using both Cronbach’s alpha and McDonald’s omega. Three software programs—SPSS 28.0, AMOS, and JASP—were employed for the analysis.

Results

Participant Characteristics

Table 1 presents the demographic characteristics of the study participants. The mean age of the participants was 43.90 ± 13.32 years. In terms of gender, 50.5% were male and 49.5% were female. Most were married (58.8%), and 33.8% and 6.8% were single and divorced/widowed, respectively. Only 0.6% reported living with their partners without a legal relationship. Among the single participants, 43.8% answered that they had an intimate relationship during the survey period, and another 43.8% answered that they had intimate relationships before but not during the survey period. Approximately 12.4% reported that they had never had an intimate relationship in their lifetime. Furthermore, approximately 40.2% of the participants reported that they were religious, and 20.6% reported witnessing or suspecting IPV among their acquaintances.

Table 1.

Demographics of Survey Participants.

Characteristics	Categories		n (%)
Age (years)	20–29		63 (20.3)
	30–39		62 (19.9)
	40–49		60 (19.3)
	50–59		60 (19.3)
	60–64		66 (21.2)
Gender	Male		157 (50.5)
	Female		154 (49.5)
Marital status	Single		105 (33.8)
	Married		183 (58.8)
	Divorced/widowed		21 (6.8)
	Living together		2 (0.6)
(If single), history of having intimate partner relationships	Yes	Currently ongoing	46 (43.8)
		Before but not now	46 (43.8)
	No		13 (12.4)
Having a religion	Yes		125 (40.2)
	No		186 (59.8)
Experience of witnessing/suspecting IPV	Yes		64 (20.6)
	No		247 (79.4)

IPV = intimate partner violence.

Item Analysis

Table 2 lists each item’s mean, SD, and communality. The skewness and kurtosis values of each item ranged from .39 to .70 and from −1.45 to .01, respectively. As the absolute values of skewness and kurtosis did not exceed 2 and 7, respectively, the normality of the data was assumed (Kim, 2013). Items with communalities below .30 are considered to have low explanatory power (Holm et al., 2019). However, all items in our study exhibited communality values exceeding .30.

Table 2.

Analysis of the Items (n = 311).

Items	Mean ± SD	Communality
1. A person constantly informs their partner about their whereabouts and whom they meet	2.01 ± .76	.33
2. When I am with a person, the person’s partner checks whether I am of the gender that the person is interested in; in that case, the relationship becomes suspicious	1.89 ± .79	.46
3. A person’s partner does not allow the person to communicate privately with other people around them, including me	1.93 ± .86	.52
4. A person’s partner wants to know every detail about my relationship with that person	1.91 ± .82	.53
5. A person explains trivial things to their partner, fearing that the partner will be angry	1.94 ± .85	.59
6. A person’s partner jokes about the person in public to the point where the person feels humiliated	1.90 ± .98	.67
7. A person’s partner publicly insults and uses offensive language toward the person	1.94 ± 1.06	.70
8. After starting a relationship (or after marriage), a person becomes less reachable	1.74 ± .77	.54
9. After starting a relationship (or after marriage), a person becomes notably less present at social events	1.81 ± .75	.58
10. After starting a relationship (or after marriage), a person shows emotional instability toward their partner	1.95 ± .85	.75
11. A person confides to me that they feel too scared to initiate a breakup with their partner	1.97 ± .97	.81
12. A person confides to me that their partner is obsessed with them, making their daily life difficult	2.03 ± 1.00	.80
13. When I am with a person, the person cannot let go of their cell phone due to excessive contact from the partner	1.93 ± .91	.71
14. A person’s partner asks me details about a meeting with the person (e.g., meeting time/place, people gathered)	1.98 ± .94	.73
15. A person’s partner publicly shouts at the person in anger	2.05 ± 1.08	.82
16. A person’s partner, who is angry during an argument, says they will kill the person	2.09 ± 1.13	.89
17. A person’s partner, who is angry during an argument, drags the person out by the collar	2.15 ± 1.20	.89
18. A person’s partner physically pushes the person in my presence	2.08 ± 1.16	.90
19. I discover a wound on the face or body of a person who has just met with their partner	2.06 ± 1.12	.87
20. A person’s partner hurls objects at the person in my presence	2.15 ± 1.17	.88
21. A person’s partner pinches the person so hard that they visibly wince in pain	1.98 ± 1.05	.81
22. A person confesses to being beaten by the partner	2.12 ± 1.14	.87
23. A person’s partner slaps the person in my presence	2.13 ± 1.24	.93
24. A person’s partner kicks the person in my presence	2.18 ± 1.24	.92
25. A person’s partner threatens the person with a weapon in my presence	2.20 ± 1.31	.89
26. A person’s partner brags about sexual matters with the person	2.02 ± 1.11	.81
27. A person confides in me about their difficulty with unwanted sexual demands from their partner	1.93 ± 1.08	.80
28. A person’s partner releases their sex videos or photos without the consent of the person	2.20 ± 1.29	.92
29. A person’s partner touches a part of the person’s body in my presence despite their evident dislike for it	2.06 ± 1.17	.86
30. A person confesses to being raped by their partner	2.18 ± 1.28	.90
31. A person defends their partner’s abusive behaviors toward them	2.02 ± 1.13	.87
32. A person admits to being abused but does not want to leave their partner, expressing love for them	1.93 ± 1.01	.80
33. After breaking up with an abusive partner, a person wants to be reunited with them again	1.92 ± 1.03	.81
34. A person asks me for help to get out of a situation of violence with their partner	2.18 ± 1.23	.88
35. A person’s ex-partner, who had been violent, visits the person without notice, terrifying them	2.15 ± 1.22	.90
36. A person’s ex-partner, who had been violent, asks me about the person’s whereabouts	2.02 ± 1.18	.81
37. A person’s ex-partner, who had been violent, persuaded the person to reconcile	2.01 ± 1.11	.84

Exploratory Factor Analysis

Construct Validity

KMO value was .98, and the $χ$ ² of Bartlett’s test of sphericity was 17101.32 (df = 630, p < .001). Therefore, the items were determined to be suitable for EFA. According to the rule of thumb for eigenvalues, which suggests selecting factors with eigenvalues greater than 1.0, a three-factor model might be possible. However, based on the results of the scree test and parallel analysis, only the first two factors from the actual dataset exhibited eigenvalues surpassing those in the simulated random samples (Supplemental Appendix A). Consequently, a two-factor model was selected for this study.

Items with factor loadings below .30 or with cross-loading differences greater than .10 were individually excluded (Seo et al., 2018). After each exclusion, the data were subsequently reanalyzed. Consequently, 5 items (Q6, Q7, Q11, Q12, and Q14) were excluded, leaving a total of 32 items, as presented in Table 3. The first and second factors consisted of 23 and 9 items, explaining 65.21% and 10.90% of the variance, respectively. The total explained variance was 76.10%.

Table 3.

Factor Loading Analysis of the Items (n = 311).

Factor Domains		Factor Loading
Factor Domains	Items	1	2	Total
Factor 1: Direct IPV indicators	28. A person’s partner releases their sex videos or photos without the consent of the person	1.013
	34. A person asks me for help to get out of a situation of violence with their partner	.990
	30. A person confesses to being raped by their partner	.987
	35. A person’s ex-partner, who had been violent, visits the person without notice, terrifying them	.986
	25. A person’s partner threatens the person with a weapon in my presence	.952
	29. A person’s partner touches a part of the person’s body in my presence despite their evident dislike for it	.944
	23. A person’s partner slaps the person in my presence	.941
	24. A person’s partner kicks the person in my presence	.928
	37. A person’s ex-partner, who had been violent, persuaded the person to reconcile	.925
	31. A person defends their partner’s abusive behaviors toward them	.921
	36. A person’s ex-partner, who had been violent, asks me about the person’s whereabouts	.901
	19. I discover a wound on the face or body of a person who has just met with their partner	.889
	33. After breaking up with an abusive partner, a person wants to be reunited with them again	.870
	22. A person confesses to being beaten by the partner	.864
	20. A person’s partner hurls objects at the person in my presence	.860
	27. A person confides in me about their difficulty with unwanted sexual demands from their partner	.859
	26. A person’s partner brags about sexual matters with the person	.848
	32. A person admits to being abused but does not want to leave their partner, expressing love for them	.833
	21. A person’s partner pinches the person so hard that they visibly winces in pain	.806
	17. A person’s partner, who is angry during an argument, drags the person out by the collar	.753
	18. A person’s partner physically pushes the person in my presence	.751	.292
	16. A person’s partner, who is angry during an argument, says they will kill the person	.706	.346
	15. A person’s partner publicly shouts at the person in anger	.634	.399
Factor 2: Indirect IPV indicators	9. After starting a relationship (or after marriage), a person becomes notably less present at social events		.744
	4. A person’s partner wants to know every detail about my relationship with that person		.685
	3. A person’s partner does not allow the person to communicate privately with other people around them, including me		.654
	8. After starting a relationship (or after marriage), a person becomes less reachable		.641
	5. A person explains trivial things to their partner, fearing that the partner will be angry		.629
	2. When I am with a person, the person’s partner checks whether I am of the gender that the person is interested in; in that case, the relationship becomes suspicious		.628
	10. After starting a relationship (or after marriage), a person shows emotional instability toward their partner		.620
	13. When I am with a person, the person cannot let go of their cell phone due to excessive contact from the partner		.563
	1. A person constantly informs their partner about their whereabouts and whom they meet		.519
Number of items		23	9	32
Eigenvalue		24.13	4.03
Explanatory power		65.21	10.90	76.10
Cronbach’s α [95% CI]		.990 [.988, .991]	.911 [.895, .925]	.982 [.979, .984]
McDonald’s ω [95% CI]		.990 [.989, 992]	.913 [.899, .928]	.986 [.984, .988]

IPV = intimate partner violence.

Factor Labeling

The two authors named each factor based on the patterns of the items assigned to those factors. The items in the first factor pointed to overt and clear acts or indications of IPV. We labeled this factor “Direct IPV Indicators.” These items detail situations where IPV has already taken place, with visible evidence of the violence. Examples include scenarios where abuse is directly witnessed or overheard, or confessions of such acts are made. The items in the second factor portrayed signs or behaviors that might hint at possible IPV but are not overt acts or confessions of violence, leading us to name this factor “Indirect IPV Indicators.” This includes situations that are not direct revelations or observations of abuse but rather behaviors or patterns that might suggest a risk.

Reliability

The Cronbach’s alpha and the McDonald’s omega values for the instrument with 32 items were .98 and .99, respectively. Both Cronbach’s alphas and McDonald’s omegas for factors 1 and 2 were .99 and .91, respectively.

Confirmatory Factor Analysis

Assessing the Model Fit

The EFA’s two-factor solution was subsequently validated using bootstrapped maximum likelihood estimation. However, this validation indicated a disparity between the proposed model and the actual data. Specifically, the model-fit indices were as follows: $χ$ ²/df = 6.925, CFI = .818, TLI = .805, SRMR = .097, and RMSEA = .138.

To enhance the model fit, we referred to Nizar et al. (2019) and Lee (2023). Items with standardized regression weights greater than .70 indicated satisfactory factor loading; hence, items falling below this threshold were removed. In addition, we closely examined items with a modification index (MI) exceeding 15 to assess potential theoretical redundancy. For pairs of items linked by a high MI value, a covariance path was established between their error terms. Items that had multiple connections to other items or another factor with a high MI value were deemed problematic. Their removal was judiciously determined through discussions among the authors, taking into consideration factor loading values and expert CVI. Consequently, four items (Q1, Q10, Q13, and Q31) were identified as problematic and subsequently removed. Following these modifications, the model fit was significantly improved: $χ$ ²/df = 3.713, CFI = .934, TLI = .924, SRMR = .049, and RMSEA = .094. The finalized model and the instrument, which consists of 28 items, are presented in Supplemental Appendix B and Appendix C, respectively.

Convergent/Discriminant Validity and Internal Consistency

The CR values for the two factors were .99 and .93, respectively, both of which exceed .70. This demonstrates that the convergent validity of each factor is well established (Choi & You, 2017). The HTMT value was .47. As this value is below .85 (Lee, 2023), discriminant validity between the two reflective constructs is also confirmed. Finally, the internal consistency of the final scale is outstanding, with Cronbach’s alpha and McDonald’s omega values at .98 and .99, respectively. For factor 1 (22 items), both Cronbach’s alpha and McDonald’s omega were .99, while for factor 2 (6 items), both measures were .90.

Discussion

This study was conducted to develop and validate a scale, named BESW-IPV, to measure the level of bystander’s intended engagement in witnessed IPV situations. The final scale consisting of 28 items demonstrates robust internal consistency, along with well-established convergent and discriminant validity.

The BESW-IPV is designed to measure the depth or intensity of a bystander’s intended intervention (e.g., to what extent they would intervene) rather than the mere probability of them intervening (e.g., whether they might intervene or not). This was because the degree to which the interviewees wanted to intervene in situations varied even though they all showed a positive willingness to intervene in accordance with our second round of FGIs. For instance, among the participants who expressed a desire to intervene, some felt compelled to report the situation to the police, whereas others chose to provide emotional support to the victims without officially reporting the circumstances. This was consistent with McMahon et al. (2017), who stressed the need for a wider realm of possible bystander responses. Furthermore, we provided straightforward examples to describe the bystander’s degree of engagement based on the feedback from FGIs, such as, a bystander might ignore the situation (no engagement), or they may simply keep an eye on the situation or provide advice (a little– substantial engagement), or they may also directly intervene in the situation or report it to the police (full engagement). We believe that this approach will allow respondents to comprehend bystander behaviors, even if they are not culturally familiar with assisting or intervening in others’ IPV issues. This understanding encompasses the depth of engagement and the spectrum of response choices. However, our examples are not intended to judge the right or wrong ways of intervening in a situation, as we do not assume that providing advice to victims or perpetrators is inherently less effective than reporting incidents to the police. To utilize this tool in future studies, researchers should consider the depth or intensity of engagement itself.

McMahon et al. (2017) also asserted that bystander behavior should be measured by considering the “relational distance” (p. 18) between bystanders and IPV parties. Therefore, many existing measures set the distance in advance; for instance, the 9-item scale developed by Borsky et al. (2016) is targeted to measure bystander behavior intentions toward friends, while Banyard et al. (2014) developed the Intent to Help Strangers Scale to measure bystander help intentions toward strangers. Many of our South Korean interviewees expressed that they would intervene in most situations if the victims were close friends or family members. While relational intimacy has been found to increase bystanders’ intention to help, this association may be even stronger in Asian cultures, which value family and community due to the influence of collectivism (Kamimura et al., 2016). Conversely, meddling in someone else’s familial issues or intimate relationships has been culturally deemed inappropriate behavior. Therefore, in South Korea, the promotion of bystander interventions needs to be focused more on general relationships than close ones. The BESW-IPV, thus, identifies the possible victims as bystanders’ general acquaintances rather than family members or close friends.

One of the practical implications of the BESW-IPV is its potential application in previously overlooked settings, particularly in societies characterized by Confucian-influenced patriarchal and sexually conservative norms. Given that culture shapes the IPV phenomenon, IPV is often employed in a patriarchal society to uphold cultural values under the guise of women’s protection (Kwon & Park, 2019), gender role sustenance (Cao et al., 2022), or enhancement of the community’s reputation and honor (Brown et al., 2018). Hence, the prevalent forms of IPV in such contexts, including South Korea, are likely to be isolating and coercive controlling behaviors, and these tendencies can escalate over time into more overt and severe manifestations, such as physical assaults and rape (Kwon & Park, 2019). In addition, shame has been identified as a major negative consequence for both male and female victims who disclose IPV (Park et al., 2021). Victims often worry about being stigmatized and blamed, which has been recognized as one of the main barriers to victims seeking help (Park & Ko, 2020). This phenomenon is commonly found in many countries influenced by Confucianism, such as Japan and Taiwan (Shen, 2011). Shen (2011) also reported that female victims in Taiwan often face negative reactions from others regarding the IPV they have experienced, leading to feelings of fear and shame.

Concerning this, BESW-IPV can be adopted as a new tool in patriarchal settings, as it captures not only the perpetrator’s typical IPV behaviors published throughout the literature but also their control and isolation tactics toward the victim, as well as the victim’s responses and coping strategies during crises. This scale is also expected to be effective in a setting where there is a prevalent tendency to blame IPV victims and justify the perpetrator’s controlling behaviors. In addition, the scale items were designed in a gender-neutral tone and validated using samples from a wide range of age groups (from their 20s to their 60s) to enhance generalizability.

Nonetheless, this study has several limitations that must be addressed and considered in future studies. First, the social desirability bias issue is unavoidable in BESW-IPV, as it is a self-report-based scale. Second, all participants were volunteers; thus, they might have had more active and motivated characteristics than the general population, which might have influenced their responses. Third, the model-fit index of $χ$ ²/df did not align with the commonly accepted model-fit criteria. Although other fit indices were considered, it is important to interpret the findings with caution considering this particular misfit. Efforts should be made to re-evaluate the model with larger or different samples in future studies. Fourth, we provided examples of response options based on FGIs; however, the examples might be vague for some respondents and may affect their responses. Finally, the BESW-IPV was developed in South Korea and thus validated only on South Korean samples. Hence, further research is needed to translate the BESW-IPV and validate it in other cultural contexts.

Supplemental Material

sj-docx-1-jiv-10.1177_08862605231222902 – Supplemental material for Development and Validation of the Bystander Engagement Scale for Witnessed Intimate Partner Violence: A South Korean Version

Supplemental material, sj-docx-1-jiv-10.1177_08862605231222902 for Development and Validation of the Bystander Engagement Scale for Witnessed Intimate Partner Violence: A South Korean Version by Sihyun Park and Yejung Ko in Journal of Interpersonal Violence

Supplemental Material

sj-docx-2-jiv-10.1177_08862605231222902 – Supplemental material for Development and Validation of the Bystander Engagement Scale for Witnessed Intimate Partner Violence: A South Korean Version

Supplemental material, sj-docx-2-jiv-10.1177_08862605231222902 for Development and Validation of the Bystander Engagement Scale for Witnessed Intimate Partner Violence: A South Korean Version by Sihyun Park and Yejung Ko in Journal of Interpersonal Violence

Supplemental Material

sj-docx-3-jiv-10.1177_08862605231222902 – Supplemental material for Development and Validation of the Bystander Engagement Scale for Witnessed Intimate Partner Violence: A South Korean Version

Supplemental material, sj-docx-3-jiv-10.1177_08862605231222902 for Development and Validation of the Bystander Engagement Scale for Witnessed Intimate Partner Violence: A South Korean Version by Sihyun Park and Yejung Ko in Journal of Interpersonal Violence

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research and/or authorship of this article: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) under Grant number 2021R1F1A1048142.

ORCID iDs

Sihyun Park

Yejung Ko

Supplemental Material

Supplemental material for this article is available online.

Author Biographies

Sihyun Park, PhD serves as an associate professor in the Department of Nursing at Chung-Ang University, Seoul, South Korea. She earned her PhD from the University of Washington School of Nursing in 2015. Her research primarily revolves around interpersonal violence, with a particular focus on intimate partner violence and violence against women. Dr. Park holds a keen interest in scrutinizing these issues through both cultural and systematic lenses.

Yejung Ko, PhD is an associate professor in the College of Nursing and Health at Kongju National University, as well as a registered psychiatric nurse specialist in South Korea. She earned her PhD in Nursing from Chung-Ang University in 2014, and her research interests include mental health issues, with a particular emphasis on evidence-based interventions pertaining to stress and trauma.

References

Banyard

V. L.

Moynihan

M. M.

Cares

A. C.

Warner

(2014). How do we know if it works? Measuring outcomes in bystander-focused abuse prevention on campuses. Psychology of Violence, 4(1), 101–115. https://doi.org/10.1037/a0033470

Bennett

Banyard

V. L.

Edwards

K. M.

(2017). The impact of the bystander’s relationship with the victim and the perpetrator on intent to help in situations involving sexual violence. Journal of Interpersonal Violence, 32(5), 682–702. https://doi.org/10.1177/0886260515586373

Boateng

G. O.

Neilands

T. B.

Frongillo

E. A.

Melgar-Quiñonez

H. R.

Young

S. L.

(2018). Best practices for developing and validating scales for health, social, and behavioral research: A primer. Frontiers in Public Health, 6, 149. https://doi.org/10.3389/fpubh.2018.00149

Borsky

A. E.

McDonnell

Rimal

R. N.

Turner

(2016). Assessing bystander behavior intentions toward friends to prevent dating violence: Development of the bystander behavior intentions-friends scale through concept mapping and exploratory factor analysis. Violence and Victims, 31(2), 215–234. https://doi.org/10.1891/0886-6708.VV-D-13-00196

Brown

R. P.

Baughman

Carvallo

(2018). Culture, masculine honor, and violence toward women. Personality and Social Psychology Bulletin, 44(4), 538–549. https://doi.org/10.1177/0146167217744195

Cao

Liu

Gonzalez-Guarda

R. M.

(2022). Trapped in my roles as a woman with no help: Experiences of intimate partner violence against Chinese women. Violence Against Women, 29(5), 964–986. https://doi.org/10.1177/10778012221104504

Cascardi

Krauss

O’Leary

K. D.

Loatman

K. L.

Sargent

Grych

Jouriles

E. N.

(2021). The bystander behavior (for friends) scale: Factor structure and correlation with prior victimization. Journal of Interpersonal Violence, 36(9–10), NP4850–NP4873. https://doi.org/10.1177/0886260518794011

Casper

D. M.

Witte

Stanfield

M. H.

(2021). “A person I cared about was involved”: Exploring bystander motivation to help in incidents of potential sexual assault and dating violence. Journal of Interpersonal Violence, 36(9–10), 4406–4430. https://doi.org/10.1177/0886260518791232

Choi

H. C.

You

Y. Y.

(2017). The study on the comparative analysis of EFA and CFA. Journal of Digital Convergence, 15(10), 103–111. https://doi.org/10.14400/JDC.2017.15.10.103

10.

Clark

L. A.

Watson

(1995). Constructing validity: Basic issues in objective scale development. Psychological Assessment, 7(3), 309–319. https://doi.org/10.1037/1040-3590.7.3.309

11.

De Vellis

R. F.

(2003). Scale development: Theory and applications (2nd ed., Vol. 26). Sage Publications.

12.

Edwards

K. M.

Sessarego

S. N.

Stanley

L. R.

Mitchell

K. J.

Eckstein

R. P.

Rodenhizer

K. A. E.

Leyva

P. C.

Banyard

V. L.

(2021). Development and psychometrics of instruments to assess school personnel’s bystander action in situations of teen relationship abuse and sexual assault. Journal of Interpersonal Violence, 36(3–4), NP1586–1606NP. https://doi.org/10.1177/0886260517746946

13.

Gerber

M. R.

Fried

L. E.

Pineles

S. L.

Shipherd

J. C.

Bernstein

C. A.

(2012). Posttraumatic stress disorder and intimate partner violence in a women’s headache center. Women & Health, 52(5), 454–471. https://doi.org/10.1080/03630242.2012.684088

14.

Gracia

Martín-Fernández

Marco

Santirso

F. A.

Vargas

Lila

(2018). The willingness to intervene in cases of intimate partner violence against women (WI-IPVAW) scale: Development and validation of the long and short versions. Frontiers in Psychology, 9, 1146. https://doi.org/10.3389/fpsyg.2018.01146

15.

Hair

J. F.

Black

W. C.

Babin

B. J.

Anderson

R. E.

(2019). Multivariate data analysis (8th ed.). Cengage.

16.

Hamberger

L. K.

Larsen

S. E.

Lehrner

(2017). Coercive control in intimate partner violence. Aggression and Violent Behavior, 37, 1–11.

17.

Hamby

Weber

M. C.

Grych

Banyard

(2016). What difference do bystanders make? The association of bystander involvement with victim outcomes in a community sample. Psychology of Violence, 6(1), 91–102. https://doi.org/10.1037/a0039073

18.

Holm

Alvariza

Fürst

C. J.

Öhlen

Årestedt

(2019). Psychometric evaluation of the anticipatory grief scale in a sample of family caregivers in the context of palliative care. Health and Quality of Life Outcomes, 17(1), 42. https://doi.org/10.1186/s12955-019-1110-4

19.

Kamimura

Trinh

H. N.

Nguyen

Yamawaki

Bhattacharya

Birkholz

Makomenaw

Olson

L. M.

(2016). Bystander attitudes to prevent sexual assault: A study of college students in the United States, Japan, India, Vietnam, and China. Journal of Sex Research, 53(9), 1131–1138. https://doi.org/10.1080/00224499.2016.1184605

20.

Kang

(2013). Discussions on the suitable interpretation of model fit indices and the strategies to fit model in structural equation modeling. Journal of the Korean Data Analysis Society, 15(2), 653–668.

21.

Katz

Pazienza

Olin

Rich

(2015). That’s what friends are for: Bystander responses to friends or strangers at risk for party rape victimization. Journal of Interpersonal Violence, 30(16), 2775–2792. https://doi.org/10.1177/0886260514554290

22.

Kim

H.-Y.

(2013). Statistical notes for clinical researchers: Assessing normal distribution using skewness and kurtosis. Restorative Dentistry & Endodontics, 38(1), 52–54. https://doi.org/10.5395/rde.2013.38.1.52

23.

Kim

Park

Champion

J. D.

(2021). Intimate partner femicide in South Korea: An integrative review of the literature. Journal of Forensic Nursing, 18(2), 117–123. https://doi.org/10.1097/jfn.0000000000000353

24.

Kwon

Park

(2019). Experiencing coercive control in female victims of dating violence. Journal of Korean Academy of Nursing, 49(1), 46–58. https://doi.org/10.4040/jkan.2019.49.1.46

25.

Lee

E. H.

(2023). Application and cases of structural equation modeling in nursing research. Research Seminar, 2023, 3–58.

26.

McMahon

Palmer

J. E.

Banyard

Murphy

Gidycz

C. A.

(2017). Measuring bystander behavior in the context of sexual violence prevention: Lessons learned and new directions. Journal of Interpersonal Violence, 32(16), 2396–2418. https://doi.org/10.1177/0886260515591979

27.

Mennicke

Montanaro

Bowling

Tirunagari

Williams

Jules

B. N.

Campbell

Carlson

Farris

McClare

Kissler

McGonagle

Pruneda

Haley

McMahon

Correia

C. J.

Meehan

Benson

J. K.

Willard

. . . Sotiroff

(2022). A systematic review of validated measures of U.S.-based bystander intervention-related constructs. Trauma, Violence, and Abuse, 24(5), 3732–3747. https://doi.org/10.1177/15248380221137067

28.

Mujal

G. N.

Taylor

M. E.

Fry

J. L.

Gochez-Kerr

T. H.

Weaver

N. L.

(2019). A systematic review of bystander interventions for the prevention of sexual violence. Trauma, Violence, and Abuse, 22(2), 381–396. https://doi.org/10.1177/1524838019849587

29.

Nizar

Ali

M. W.

Tuan Amran

T. A.

Alias

Chik

(2019). Assessment model for construct occupational accident using confirmatory factor analysis. E3S Web of Conferences, 90, 03004. https://doi.org/10.1051/e3sconf/20199003004

30.

Park

Chesla

(2007). Revisiting confucianism as a conceptual framework for Asian family study. Journal of Family Nursing, 13(3), 293–311. https://doi.org/10.1177/1074840707304400

31.

Park

Bang

S. H.

Jeon

(2021). “This society ignores our victimization”: Understanding the experiences of Korean male victims of intimate partner violence. Journal of Interpersonal Violence, 36(23–24), 11658–11680. https://doi.org/10.1177/0886260519900966

32.

Park

(2020). Victims of intimate partner violence in South Korea: Experiences in seeking help based on support selection. Violence Against Women, 27(3–4), 320–338. https://doi.org/10.1177/1077801220905638

33.

Park

Woo

(2023). Types of bystanders among South Koreans based on their subjective reactions to intimate partner violence. Journal of Interpersonal Violence, 38(15–16), 8969–8990. https://doi.org/10.1177/08862605231162652

34.

Polit

D. F.

Beck

C. T.

(2006). The content validity index: Are you sure you know what’s being reported? Critique and recommendations. Research in Nursing and Health, 29(5), 489–497. https://doi.org/10.1002/nur.20147

35.

Rothman

E. F.

Paruk

Banyard

(2018). The escalation dating abuse workshop for college students: Results of an efficacy RCT. Journal of American College Health, 66(6), 519–528. https://doi.org/10.1080/07448481.2018.1431909

36.

Seo

Lee

Kim

(2018). Exploratory factor analysis in psychological research: Current status and suggestions for methodological improvements. Journal of Social Science, 29(1), 177–193. https://doi.org/10.16881/jss.2018.01.29.1.177

37.

Shen

A. C.

(2011). Cultural barriers to help-seeking among Taiwanese female victims of dating violence. Journal of Interpersonal Violence, 26(7), 1343–1365. https://doi.org/10.1177/0886260510369130

38.

Son

M. S.

Cho

J. Y.

(2016). Dating violence victims support measures. Korean Women’s Hot Line. http://hotline.or.kr/board_statistics/28328

39.

Stojanov

Treharne

G. J.

Graham

Liebergreen

Shaw

Hayward

Beres

(2021). Pro-social bystander sexual violence prevention workshops for first year university students: Perspectives of students and staff of residential colleges. Kotuitui: New Zealand Journal of Social Science Online, 16(2), 432–447. https://doi.org/10.1080/1177083X.2021.1906712

40.

Taylor

Banyard

Grych

Hamby

(2019). Not all behind closed doors: Examining bystander involvement in intimate partner violence. Journal of Interpersonal Violence, 34(18), 3915–3935. https://doi.org/10.1177/0886260516673629

41.

Wong

J. Y.-H.

Tang

N. R.

Yau

J. H.-Y.

Choi

A. W.-M.

Fong

D. Y.-T.

(2019). Dating CAFE ambassador programme: Chinese college students to help peers in dating violence. Health Education & Behavior, 46(6), 981–990. https://doi.org/10.1177/1090198119867736

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

0.07 MB

0.43 MB

0.03 MB