A proposed model for evaluating the impact of participating in trauma-focused research

Abstract

This article describes the development of a model for assessing levels of risk in trauma-focused research, with evidence for the empirical validity of the model being provided by findings from a study of 438 undergraduate students who participated in a survey regarding past-year exposure to intimate partner violence. A minor increase in risk over minimal risk was found to constitute the modal risk category, with evidence for the empirical validity of the model being provided by the fact that levels of risk assessed by the model were associated with anticipated decreases in the proportion of positive cost–benefit ratios, with different levels of risk being predicted by a unique constellation of risk factors. These findings are discussed with respect to their implications for further research and in relation to the way in which the impact of trauma-focused research is conceptualised.

Keywords

Distress ethics participation-related distress research risk trauma

Introduction

Contemporary perspectives on the impact of trauma-focused research have been characterised by a clash of perspectives: with the views of some clinicians, researchers, and institutional review boards (IRBs) tending to reflect a risk bias, in terms of which the primary concern is with the avoidance of harm, lest research participants become distressed, destabilised, or possibly even retraumatised (Appollis, Lund, deVries, & Mathews, 2015; Rossiter, 2012), and with the views of others tending to reflect a benefit bias in terms of which the primary concern is with the potential benefits of research participation, lest research participants be deprived of such benefits (see, for example, Goossens, Nicholls, Torchalla, Brink, & de Ruiter, 2016; Hebenstreit & DePrince, 2012).

A risk-focused perspective on trauma research has tended to be informed by the tacit assumption that trauma survivors constitute a particularly vulnerable and emotionally fragile population who are likely to experience extreme psychological distress, or to feel stigmatised, if asked to recount past traumatic experiences (Becker-Blease & Freyd, 2006; Legerski & Bunnell, 2010; Rossiter, 2012; Seedat, Pienaar, Williams, & Stein, 2004), with such assumptions having found to be particularly evident in studies of IRB decision-making processes (e.g., Ceci, Peters, & Plotkin, 1985; Jaffe, DiLillo, Hoffman, Haikalis, & Dykstra, 2015). For example, in a recent survey of 114 US trauma researchers (Jaffe et al., 2015), 61% indicated that IRBs had expressed concerns regarding their research protocols, with the most common concern being that asking research participants about prior traumatic experiences may ‘cause harm by triggering negative emotions that could “shatter” participants . . . [and possibly] lead to the need for long-term therapy or even cause participants to become suicidal’ (p. 41). Based upon these concerns, IRBs had either refused to approve protocols (13% of cases) or requested modifications to protocols, including removing questions relating to a trauma history or having a qualified counsellor present in the room to deal with any adverse reactions.

Such assumptions of extreme harm have, however, been challenged by a growing body of evidence, which suggests that participating in trauma-focused research is (a) relatively well tolerated by participants, with reported distress levels tending to be low to moderate (Jaffe et al., 2015), (b) no more distressing for participants than being asked to participate in research on ostensibly non-sensitive topics (e.g., questions relating to body image or parental income; Cromer, Freyd, Binder, DePrince, & Becker-Blease, 2006); (c) not associated with any significant change in reported levels of posttraumatic stress disorder (PTSD) symptomatology (Bassa & Collings, 2012); and (d) unlikely to have stigmatising effects (Becker-Blease & Freyd, 2006).

The emerging view that participation in trauma-focused research may not be as harmful as has previously been suspected (Jaffe et al., 2015) has been supplemented by an increased focus on the benefits of participating in trauma research, with these research efforts suggesting that trauma-focused research tends to produce positive overall risk–benefit ratios (Appollis et al., 2015; Goossens et al., 2016; Hebenstreit & DePrince, 2012; Newman & Kaloupek, 2004), with participants reporting a number of positive gains from participation, including enhanced self-esteem (Kassam-Adams & Newman, 2002; Ruzek & Zatzick, 2000) and satisfaction with having been provided with the opportunity to reflect on, or to gain new insights regarding, trauma experiences (Dyregrov, Dyregrov, & Raundalen, 2000).

Although available studies suggest that participation in trauma-focused research is generally well tolerated, it is clear that this is not inevitably the case, with a subset of participants in trauma-focused research having consistently been found to report high levels of distress and negative cost–benefit ratios (Appollis et al., 2015; Newman & Kaloupek, 2004; Seedat et al., 2004). However, research findings regarding the extent of participation-related distress vary widely – available estimates varying from 4% to 50% across studies (Appollis et al., 2015) – with attempts to identify factors associated with increased perceptions of distress having produced few consistent findings. Thus, while some studies have found that females are likely to experience more distress than males (Black, Kresnow, Simon, Arias, & Shelley, 2006; Kuyper, de Wit, Adam, & Woertman, 2012), other studies have found that males experience higher levels of distress than females (DePrince & Chu, 2008; Sikweyiya & Jewkes, 2012). Similarly, while a number of studies have found that the presence of PTSD symptoms is associated with higher levels of participant-related distress (e.g., Carter-Visscher, Naugle, Bell, & Suvak, 2007; DePrince & Chu, 2008; Johnson & Benight, 2003), others have found no association between the presence of PTSD symptoms and reported levels of distress (Kassam-Adams & Newman, 2005; Newman, Willard, Sinclair, & Kaloupek, 2001).

Moreover, attempts to obtain agreement on the extent of participation-related distress have been confounded by the fact that researchers and IRB members tend to have quite separate conceptual models of research risk. For most researchers, risk has tended to be conceptualised as a dichotomous variable (i.e., either present or absent), while for IRB members, risk has tended to be conceptualised as an ordinal variable, with the lowest level of risk being ‘minimal risk’ which means that ‘the probability and magnitude of harm or discomfort anticipated in the research are not greater in and of themselves than those ordinarily encountered in daily life’ (Miller, 2003), and with progressively higher levels of risk involving graduated increases of risk above minimal risk (minor increase, more than minor increase, etc.).

To align these different conceptualisations of research risk, it would be helpful if researchers employed definitions of risk that corresponded to the levels of risk considered by IRB members. However, as Newman and Kaloupek (2004) point out, the concepts involved in such an exercise (minimal risk, minor increase over minimum risk, etc.) are likely to be difficult to operationalise.

The present study

In an attempt to address some of the issues raised in the above review, this study proposes a model for defining levels of risk in trauma-focused research (i.e., minimal, minor increase over minimal, and more than minor increase over minimal) and conducted a preliminary validation of this model in a sample of undergraduate university students (N = 438) who were surveyed regarding their recent experiences of exposure to intimate partner violence (IPV).

The risk model tested in the study was derived from scores obtained on the Discomfort factor of the Response to Research Participation Questionnaire (RRPQ; Newman & Kaloupek, 2004; Newman et al., 2001), which is scored using a 5-point Likert-type scale (1 = strongly disagree, 2 = disagree, 3 = uncertain, 4 = agree, 5 = strongly agree). First, mean item scores for the Discomfort factor were calculated for each participant (range: 1–5). Second, mean item scores were used to define three levels of risk: minimal risk (mean item score: ≤2; that is, disagree/strongly disagree), minor increase over minimal risk (mean item scores: >2 and <4; that is, the intermediate ‘neutral/maybe’ zone), and more than minor increase over minimal risk (mean item scores: ≥4; that is, agree/strongly agree).

The model validation phase of the research was informed by a number of hypotheses:

Hypothesis 1. There are those who might criticise the proposed model because of the assumption that the intermediate zone of uncertainty reflects a minor increase in risk over minimal risk, with the contrary argument being that the zone of uncertainty simply reflects a degree of uncertainty which, in itself, does not necessarily imply any increase in perceived risk. To test this key assumption, it was predicted that perceptions of minimal risk would be associated with positive cost–benefit ratios (Hypothesis 1a) and that there would be a progressive decrease in the proportion of positive cost–benefit ratios as risk levels increased – that is, a minor increase over minimal risk would be associated with a minor decrease in the proportion of positive cost–benefit ratios, with a more than minor increase over minimal risk being associated with a substantial decrease in the proportion of positive cost–benefit ratios (Hypothesis 1b).

Hypothesis 2. With respect to the proportion of participants who are likely to fall into different risk categories, it is something of a moot point as to whether minimal risk (a benefit-focused perspective) or a higher level of risk (a risk-focused perspective) is likely to constitute the modal risk category. As such, predictions regarding the modal risk category were made using a non-directional hypothesis, in terms of which it was predicted that participants’ perceptions of risk would not be evenly distributed across risk categories.

Hypothesis 3. To test whether the three levels of risk specified by the model were empirically distinct, it was predicted that different levels of risk would be predicted by a unique combination of the independent variables considered in the study (demographic variables, clinical status, and history of exposure to IPV).

Method

Participants

In total, 840 questionnaires were distributed to undergraduate social science classes at the University of KwaZulu-Natal (South Africa) in 2016. Of the 756 students (90%) who submitted completed questionnaires, 438 (57.9%) met the inclusion criterion of having been in an intimate relationship in the past 12 months. The mean age of these 438 participants was 19.5 years (range = 18–25 years, standard deviation [SD] = 1.70). From Table 1, it is evident that participants were predominantly female (76.7%), Black African (83.6%), and raised in homes where there was no poverty (66.4%); 57 participants (13.0%) met the study criterion for a diagnosis of depression, 57 participants (13%) for a diagnosis of an anxiety disorder, and 39 participants (8.9%) for a diagnosis of PTSD.

Table 1.

Sample characteristics (N = 438).

Sample characteristic	%	n
Gender
Male	23.3	102
Female	76.7	336
Ethnicity
Asian	6.2	27
Black	83.6	366
Caucasian	4.1	18
Other	6.1	27
Poverty in family home
No	66.4	291
Yes	33.6	147
Presence of clinical syndromes
Depression	13.0	57
Anxiety	13.0	57
Posttraumatic stress	8.9	39

Instruments

Demographic information was obtained regarding age, gender, and ethnicity, with the presence of poverty in the home being assessed using the poverty subscale of the Developmental Trauma Inventory (DTI; Collings, Valjee, & Penning, 2013). The poverty subscale of the DTI assesses poverty in the family home using three items (e.g., ‘My family was so poor that we sometimes did not have enough food to eat’), which are scored as being either false (score = 0) or true (score = 1), with item scores being summed to produce a possible score range of 0 to 3. Work with the poverty subscale indicates that it is characterised by acceptable levels of convergent validity (Valjee & Collings, 2015) and adequate levels of internal consistency (Cronbach’s α = .72; Collings et al., 2013). In this study, Cronbach’s alpha for the poverty subscale was .72.

The RRPQ (Newman & Kaloupek, 2004) is a 23-item scale designed to assess participant’s reactions to research participation, which is scored using five subscales (Participation, Personal Benefit, Global Evaluation, Drawbacks, and Emotional Reactions), each of which is scored using a 5-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). However, given that this five-factor structure has not been replicated in studies conducted in developing countries (Bassa & Collings, 2012), it was decided to subject RRPQ scores to an exploratory factor analysis that employed principal components analysis and factor rotation using the varimax procedure. This analysis (Kaiser-Meyer-Olkin test = .811, Bartlett’s test = .000) yielded three factors that met the study criteria for retention: an eigenvalue greater than 1, a scree plot supporting retention, a factor loading of at least .40, and the meaningfulness of identified factors. The first of these factors, which was labelled Benefit (23.5% of variance, α = .72, RRPQ items: 1, 2, 4, 7, 8, 15, 16, 17, and 22), contained items such as ‘I found participating in this study personally meaningful’. The second factor, which was labelled Satisfaction (7.4% of variance, α = .73, RRPQ items: 9, 11, 12, 13, 14, 21, 23), contained items such as ‘I was treated with respect and dignity by the researchers’; while the third factor, which was labelled Discomfort (15.3% of variance, α = .79, RRPQ items: 3, 5, 6, 10, 18, 19, 20), contained items such as ‘The research made me think about things I didn’t want to think about’.

Symptoms of depression and anxiety were assessed using the Depression (11 items) and Anxiety (six items) subscales of the Hopkins Symptom Checklist (HSCL; Derogatis, Lipman, Rickels, Uhlenhuth, & Covey, 1974), both of which have been extensively validated in psychiatric and non-psychiatric samples (Bass et al., 2013). Items on both subscales are scored using a 4-point Likert-type scale ranging from 1 (not at all) to 4 (extremely). The cut-off score used in this study to classify participants as having clinically significant levels of anxiety or depression (mean item score for a subscale of 1.75 or higher) has been found to be associated with high levels of efficiency in a number of studies (cf. Bass et al., 2013). In this study, high levels of internal consistency were obtained for both the Depression (α = .81) and the Anxiety (α = .85) subscales.

PTSD symptoms were assessed using the Trauma Screening Questionnaire (TSQ; Brewin et al., 2002). Items on the 10-point scale are scored on a 3-point Likert-type scale ranging from 0 (not present) to 2 (present at least twice a week). The TSQ has been found to be characterised by high levels of sensitivity and specificity, with the performance of the TSQ having been found to be equivalent to agreement achieved between two full clinical interviews (Brewin et al., 2002). The cut-off score used in this study to classify participants as having clinically significant levels of PTSD (a score of 2 on at least six scale items) has been found to be associated with a sensitivity of .86 and a specificity of .93 (Brewin et al., 2002). In this study, Cronbach’s alpha for the TSQ was .83.

Physical and sexual forms of IPV were assessed using the Physical Assault and Sexual Coercion subscales of the short form of the Revised Conflict Tactics Scale (CTS; Straus & Douglas, 2004), which was standardised on an undergraduate student sample and which has been shown to have sound psychometric properties in a number of studies (Shore, Brasfield, Febres, Cornelius, & Stuart, 2012; Straus & Douglas, 2004). Each subscale contains two items designed to assess for different levels of severity. For example, (a) moderate physical assault is assessed using the item, ‘My partner pushed, shoved, or slapped me’; while severe physical assault is assessed using the item, ‘My partner punched, kicked, or beat me up’; and (b) moderate sexual coercion is assessed using the item, ‘My partner insisted on sex when I did not want to or insisted on sex without a condom (but did not use physical force)’; while severe sexual coercion is assessed using the item, ‘My partner used force (like hitting, holding down, or using a weapon) to make me have sex’. In this study, a particular form of IPV was deemed to be present if participants endorsed any of the items (moderate or severe) relating to that form of IPV.

Emotional battering was assessed using the Woman’s Experience of Battering (WEB; Smith, Earp, & DeVellis, 1995) scale, with the term battering having being defined as a process whereby one member of an intimate partner relationship experiences ‘psychological vulnerability, loss of power and control, and entrapment’ as a result of a partner’s ‘exercise of power through the patterned use of physical, sexual, psychological and/or moral force’ (Smith, Danis, & Helmick, 1998, p. 2). Rather than focusing on the behaviours of abusive partners, the WEB is designed to assess for an individual’s perceptions of their vulnerability to physical danger and/or loss of power in a relationship. The WEB contains 10 items that are scored on a 6-point Likert-type scale ranging from 1 (disagree strongly) to 6 (agree strongly). The WEB has been found to have high discriminant validity (accurately distinguishing battered from non-battered women) and strong internal consistency (α = .91 in the validation sample; cf. Smith et al., 1995). The cut-off score (>19) used to classify participants as battered in this study has been found to be associated with a sensitivity of 94.6% and a specificity of 96.1% (Smith, Thornton, DeVellis, Earp, & Coker, 2002). Cronbach’s alpha for the WEB in this study was .88.

Procedure

Ethical clearance for the research was obtained from the University of KwaZulu-Natal Humanities and Social Science Research Ethics Committee in 2016. The study was introduced to potential participants by the researcher during normal lecture periods. Students were informed that the study involved ‘a survey of student’s experiences of intimate partner violence’. Potential participants were informed that information obtained from the survey would be used to obtain a better understanding of the nature and scope of IPV experienced by students and to assist the university to improve student counselling services. Students were also informed that the survey contained questions relating to psychological health as well as stressful life events such as exposure to IPV. Students were then provided with a copy of the study information sheet – which outlined the aims of the study, the study procedure, voluntariness of participation, assurances of confidentiality and anonymity, as well as an offer of free counselling services for all research participants – and encouraged to read it.

After 5 min of reading time, any questions students had regarding the study were answered, and students were invited to take a research pack which contained a consent documentation form (to be completed before leaving the lecture venue) and a copy of the research questionnaire (to be completed anonymously away from lectures and returned by posting the questionnaire in a sealed box placed outside the lecture theatre each day for a week).

Ethical considerations

Ethical clearance for the research was obtained from the University of KwaZulu-Natal Humanities and Social Science Research Ethics Committee in 2016.

Results

Preliminary analyses

Tests of normality

Tests of normality (Kolmogorov–Smirnov and Shapiro–Wilk) indicated that the distribution of scores for all variables considered in the study deviated significantly from what would be expected under the normal curve. Subsequent calculations were therefore conducted using non-parametric procedures.

Prevalence of IPV exposure

At least one form of IPV was reported by 141 participants (32.2%), with exposure to: physical assault being reported by 57 participants (13.0%), sexual coercion by 81 participants (18.5%), and emotional battering by 60 participants (13.7%). From Table 2, it is evident that male and female participants were equally likely to report exposure to emotional battering, with females being significantly more likely to report exposure to physical assault, sexual coercion, and at least one form of IPV.

Table 2.

Gender differences in exposure to IPV.

	Sample						Gender comparison
	Total(N = 438)		Male(n = 102)		Female(n = 336)		Gender comparison
	n	%	n	%	n	%	χ²(1)	p
Exposure to IPV
Physical assault	57	13.0	6	5.9	51	15.2	5.97	.015
Sexual coercion	81	18.5	6	5.9	75	22.3	14.03	<.001
Emotional battering	60	13.7	12	11.8	48	14.3	0.42	.517
At least one form of IPV	141	32.2	21	20.6	120	35.7	8.20	.004

IPV: intimate partner violence.

Overall RRPQ scores

An overall cost–benefit analysis for the sample was conducted by subtracting mean item scores for Discomfort from mean item score for Benefit (benefit–discomfort ratio) and Satisfaction (satisfaction–discomfort ratio). Mean item scores obtained for the sample were Benefit (3.79), Satisfaction (4.56), and Discomfort (3.58), resulting in overall positive benefit–discomfort (0.21) and satisfaction–discomfort (0.98) ratios, a pattern of findings that is consistent with findings obtained in most previous studies.

Tests of Hypothesis 1

Cost–benefit ratios for each participant were obtained by subtracting the mean item RRPQ Discomfort score obtained by each participant from their mean item Benefit and Satisfaction scores. This analysis indicated that 228 participants (52.1%) reported positive benefit–discomfort ratios and 354 participants (80.8%) reported positive satisfaction–discomfort ratios. Consistent with the predictions of Hypothesis 1a, all participants who rated their participation as minimal risk obtained positive satisfaction–distress and benefit–distress ratios (see Figure 1).

Figure 1.

Proportion of positive cost–benefit ratios by degree of perceived risk (N = 438).

From Figure 1, it is evident that a minor increase in risk over minimum risk was associated with a small but significant decrease (1.00 to 0.85) in the proportion of positive benefit–distress ratios, χ²(1) = 4.63, p = .031, and a small but non-significant decrease (1.00 to 0.95) in the proportion of positive satisfaction–distress ratios, χ²(1) = 1.14, p = .287, with a more than minor increase in risk over minimal risk being associated with substantial decreases in both the proportion of positive benefit–distress ratios (1.00 to 0.06), χ²(1) = 140.04, p < .001, and in the proportion of positive satisfaction–distress ratios (1.00 to 0.60), χ²(1) = 16.43, p < .001. Taken together, these findings provide compelling support for Hypothesis 1b, which states that a minor increase over minimal risk would be associated with a minor decrease in the proportion of positive cost–benefit ratios, with a more than minor increase over minimal risk being associated with a more substantial decrease in the proportion of positive cost–benefit ratios.

Tests of Hypothesis 2

Consistent with the predictions of Hypothesis 2, there were significant differences in the proportion of participants who fell into different risk categories (see Figure 1), χ²(2) = 149.22, p < .001. The modal risk category was a minor increase over minimal risk (51%), followed by a more than minor increase over minimal risk (43%), with minimal risk being reported by only 6% of participants. Taken together, these findings suggest that more than 9 out of 10 participants (94%) experienced their participation as more than minimal risk.

Tests of Hypothesis 3

Factors predicting risk levels were assessed using a multinomial logistic regression analysis in which risk levels (minimal risk entered as the reference category) were regressed on demographic characteristics (age, gender, ethnicity, and poverty levels), clinical status (depression, anxiety, and PTSD), and exposure to IPV (physical assault, sexual coercion, and emotional battering). The final model accounted for a significant proportion of the variance: χ²(20) = 147.98, p < .001, Nagelkerke R² = .329, f² effect size = 0.52 (see Table 3).

Table 3.

Factors associated with increased risk levels: multinomial regression analysis (N = 438).

Risk factor	Risk
	Minimal vs. minor increase over minimal			Minimal vs. more than minor increase over minimal
	Wald	OR (95% CI)	p	Wald	OR (95% CI)	p
Covariates
Younger age	0.20	0.96 [0.80, 1.15]	.652	11.24	1.50 [1.18, 1.90]	.001
Female gender	0.90	1.30 [0.76, 2.23]	.342	1.23	1.62 [0.69, 3.77]	.267
Black African	0.78	1.30 [0.73, 2.34]	.372	1.25	1.24 [0.70, 2.36]	.299
Poverty in the home	2.47	1.56 [0.90, 2.73]	.116	1.81	0.59 [0.28, 1.27]	.179
Clinical status
Depression	4.18	2.09 [1.03, 4.31]	.021	7.94	2.97 [1.25, 7.51]	.005
Anxiety	2.01	1.47 [0.86, 2.51]	.156	0.07	1.10 [0.52, 2.30]	.795
PTSD	0.01	1.03 [0.50, 2.12]	.934	8.17	3.65 [1.50, 8.85]	.004
Exposure to IPV
Physical assault	0.08	1.12 [0.51, 2.48]	.893	0.80	0.61 [0.21, 1.79]	.372
Sexual coercion	0.01	1.04 [0.51, 2.11]	.919	0.25	1.26 [0.51, 3.14]	.793
Emotional battering	3.80	3.50 [0.97, 12.34]	.055	21.51	21.73 [5.92, 76.92]	<.001

OR: odds ratio; CI: confidence interval; PTSD: posttraumatic stress disorder; IPV: intimate partner violence.

Significant main effects are presented in bold.

A minor increase over minimal risk was twice as likely, odds ratio (OR) = 2.09, p = .041, among participants who reported clinically significant levels of depression, with there being no significant main effects of either demographic variables or IPV exposure. A more than minor increase over minimal risk was most strongly predicted by a history of emotional battering (OR = 21.51, p < .001), with there also being significant main effects of PTSD (OR = 3.65, p = .004), depression (OR = 2.97, p = .006), and an older age (OR = 1.50, p = .001). Taken together, these findings provide support for the prediction that different levels of risk would be predicted by a unique combination of the independent variables considered in the study.

Discussion

Study findings provide two quite distinct perspectives regarding the risk of participating in IPV research. When cost–benefit ratios were calculated using mean group scores (the typical research paradigm), all cost–benefit ratios (satisfaction–distress and benefit–distress) were positive, suggesting that participation was generally well tolerated. However, when the proposal risk model was tested, the vast majority of participants rated their participation as more than minimal risk (94%), with 43% rating their participation as more than a minor increase over minimal risk. Moreover, cost–benefit analyses indicated that cost–benefit ratios were systematically associated with risk levels, with the highest risk category being associated with substantial decreases in the proportion of positive cost–benefit ratios. In sum, study findings suggest that, for the majority of participants, (a) the magnitude of risk was moderate to high (with ‘a minor increase over minimal risk’ being the modal risk category) and (b) the probability of risk was high with more than 9 out of 10 participants (94%) reporting more than minimal risk. Taken together, these findings suggest that research participation was not particularly well tolerated by the majority of participants.

The key distinction between these two perspectives is, of course, the way in which intermediate scores (>2 to <4) on the RRPQ Distress scale are understood: that is, should intermediate scores be interpreted as representing no increase over minimal risk (the typical research paradigm) or should intermediate scores be interpreted as representing a minor increase in risk over minimal risk (an assumption of the proposed risk model). The fact that intermediate scores (compared to lower scores) on the Distress scale were found to be associated with a minor reduction in the proportion of positive cross–benefit ratios in this study – with this trend being significant for benefit–distress ratios – would suggest that intermediate scores on the Distress scale can meaningfully be construed as constituting a minor increase in risk over minimal risk.

Study findings provide support for the prediction that different constellations of risk factors would be associated with different levels of distress. The one factor that was found to be associated with both ‘a minor’ (OR = 2.09) and ‘a more than minor’ (OR = 2.96) increase in risk over minimal risk was the presence of clinically significant levels of depression. This finding is congruent with findings reported by other researchers (Carlson et al., 2003; Johnson & Benight, 2003; Ruzek & Zatzick, 2000) and is consistent with the emerging view that depression and distress levels are likely to be highly correlated (Devarajooh & Chinna, 2017; Fisher, Gonzalez, & Polonsky, 2014; Janssens et al., 2003).

Although depression was the only variable that was found to be significantly associated with a minor increase over minimal risk, a more than minor increase over minimal risk was additionally associated with a history of emotional battering (OR = 21.51), clinically significant levels of PTSD (OR = 8.17), and a younger age (OR = 1.50). These different associations for different risk levels provide support for the empirical distinctiveness of the proposed risk categories.

There are two possible explanations for the finding that high levels of distress were found to be associated with a history of emotional battering, but not with a history of physical assault or sexual coercion. First, it is generally acknowledged that battering tends to be a chronic, continuous phenomenon that is both conceptually and empirically distinct from discrete episodes of physical and/or sexual assault (Smith et al., 2002), with chronic/continuous forms of abuse tending to be associated with more severe/complex posttraumatic reactions (Herman, 1992). Second, from a cognitive-mediational perspective (cf. Valjee & Collings, 2016), the impact of traumatic events is assumed to be mediated by an individual’s subjective perceptions of such experiences (e.g., feelings of fear, shame, guilt, or hopelessness). As such, it is the extent of negative abuse–related perceptions and appraisals (the essence of emotional battering), rather than mere exposure to abuse events (e.g., physical or sexual abuse), that is likely to be most strongly predictive of the severity of posttraumatic distress levels.

The finding that PTSD status was only associated with high levels of distress may provide an explanation for the inconsistent findings that have emerged regarding the association between PTSD and participation-related distress in previous studies (Carter-Visscher et al., 2007; DePrince & Chu, 2008; Johnson & Benight, 2003; Kassam-Adams & Newman, 2005; Newman et al., 2001). Previous studies have tended to operationalise distress in various ways (cf. Legerski & Bunnell, 2017, pp. 432–433), with some of these definitions failing to adequately capture the distinction between different distress levels. As such, differences in research findings may reflect differences in the way in which distress and other key constructs are operationalised, with further research on this issue appearing to be strongly indicated.

Conclusion

Evidence for the empirical validity of the proposed risk model is provided by the fact that levels of risk assessed by the model were associated with anticipated decreases in the proportion of positive cost–benefit ratios, with different levels of risk being predicted by a unique constellation of risk factors.

Although these findings are promising, they should at best be regarded as preliminary as they were derived from a student sample, which is unlikely to be representative of the population in general. In addition, only one form of traumatic exposure (i.e., IPV) was examined in the study, which is likely to have provided only a partial estimate of participants’ full victimisation profile, and consequently a possible overestimation of the traumagenic potential of the limited range of traumatic events surveyed (cf. Finkelhor, Ormrod, & Turner, 2007). As such, further research involving large and representative samples is indicated, with such research ideally relying on estimates of traumatic exposure that adequately capture participants’ full victimisation profile.

Finally, study findings point to the value of an ideographic (rather than a nomothetic) approach to data analysis in studies of research-induced distress, which would appear to more adequately capture and account for variations in participant reactions. In addition, the present findings suggest that the traumagenic potential of traumatic events may subsist, not so much in mere exposure to traumatic events, but rather in the ways in which such experiences are appraised by survivors. As such, a cognitive-mediational conceptualisation of the impact of traumatic events may provide a more meaningful basis for predicting distress levels in trauma-focused research.

Footnotes

Funding

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

References

Appollis

T. M.

Lund

de Vries

P. J.

Mathews

(2015). Adolescents’ and adults’ experiences of being surveyed: A systematic review of harms, benefits, and regrets. American Journal of Public Health, 105, e31–e45.

Bass

J. K.

Annan

Murray

S. M.

Kaysen

Griffiths

Cetinoglu

Bolton

P. A.

(2013). Controlled trial of psychotherapy for Congolese survivors of sexual violence. New England Journal of Medicine, 368, 2182–2191.

Bassa

Collings

S. J.

(2012). The risks and benefits of participating in research on child abuse experiences: A preliminary exploration in a non-clinical South African sample. Child Abuse Research, 13, 55–61.

Becker-Blease

K. A.

Freyd

J. J.

(2006). Research participants telling the truth about their lives: The ethics of asking and not asking about abuse. American Psychologist, 61, 218–226.

Black

M. C.

Kresnow

Simon

T. R.

Arias

Shelley

(2006). Telephone survey respondents’ reactions to questions regarding interpersonal violence. Violence and Victims, 21(4), 445–459.

Brewin

C. R.

Rose

Andrews

Green

Tata

McEvedy

Foa

E. B.

(2002). Brief screening instrument for post-traumatic stress disorder. The British Journal of Psychiatry, 181, 158–162.

Carlson

E. B.

Newman

Daniels

J. W.

Armstrong

Roth

Loewenstein

(2003). Distress in response to and perceived usefulness of trauma research interviews. Journal of Trauma & Dissociation, 4, 131–142.

Carter-Visscher

R. M.

Naugle

A. E.

Bell

K. M.

Suvak

M. K.

(2007). Ethics of asking trauma-related questions and exposing participants to arousal-inducing stimuli. Journal of Trauma & Dissociation, 8, 27–55.

Ceci

S. J.

Peters

Plotkin

(1985). Human subjects review, personal values, and the regulation of social science research. American Psychologist, 4, 994–1002.

10.

Collings

S. J.

Valjee

S. R.

Penning

S. L.

(2013). Development and preliminary validation of a screen for interpersonal childhood trauma experiences among school-going youth in Durban, South Africa. Journal of Child and Adolescent Mental Health, 25, 23–34.

11.

Cromer

L. D.

Freyd

J. F.

Binder

A. K.

DePrince

A. P.

Becker-Blease

(2006). What’s the risk in asking? Participants’ reaction to trauma history questions compared with reaction to other personal questions. Ethics & Behavior, 16, 347–362.

12.

DePrince

A. P.

Chu

A. T.

(2008). Perceived benefits in trauma research: Examining methodological and individual difference factors in responses to research participation. Journal of Experimental Research on Human Research Ethics, 3, 35–47.

13.

Derogatis

L. R.

Lipman

R. S.

Rickels

Uhlenhuth

E. H.

Covey

(1974). The Hopkins Symptom Checklist (HSCL): A self-report symptom inventory. Behavioral Science, 19, 1–15.

14.

Devarajooh

Chinna

(2017). Depression, distress and self-efficacy: The impact on diabetes self-care practices. PLoS ONE, 12, 1–16.

15.

Dyregrov

Raundalen

(2000). Refugee families’ experience of research participation. Journal of Traumatic Stress, 13, 413–426.

16.

Finkelhor

Ormrod

R. K.

Turner

H. A.

(2007). Poly-victimization: A neglected component in child victimization. Child Abuse & Neglect, 31, 7–26.

17.

Fisher

Gonzalez

J. S.

Polonsky

W. H.

(2014). The confusing tale of depression and distress in patients with diabetes: A call for greater clarity and depression. Diabetic Medicine, 31, 764–772.

18.

Goossens

Nicholls

T. L.

Torchalla

Brink

de Ruiter

(2016). The perceived impact of trauma-focused research on forensic psychiatric patients with lifetime victimization histories. Journal of Empirical Research on Human Research Ethics, 11, 334–345.

19.

Hebenstreit

C. L.

DePrince

A. P.

(2012). Perceptions of participating in longitudinal trauma research among women exposed to intimate partner abuse. Journal of Empirical Research on Human Research Ethics, 7, 60–69.

20.

Herman

(1992). Trauma and recovery: From domestic abuse to political terror. London, England: Basic Books.

21.

Jaffe

A. E.

DiLillo

Hoffman

Haikalis

Dykstra

R. E.

(2015). Does it hurt to ask? A meta-analysis of participant reactions to trauma research. Clinical Psychology Review, 40, 40–56.

22.

Janssens

A. C. J. W.

van Doorn

P. A.

de Boer

J. B.

van der Merwe

F. G. A.

Passchier

Hintzen

R. Q.

(2003). Impact of recently diagnosed multiple sclerosis on quality of life, anxiety, depression and distress of patients and partners. Acta Neurologica Scandinavica, 108, 389–395.

23.

Johnson

L. E.

Benight

C. C.

(2003). Effects of trauma-focused research on recent domestic violence survivors. Journal of Traumatic Stress, 16, 567–571.

24.

Kassam-Adams

Newman

(2002). The Reactions to Research Participation Questionnaire for Children and Parents (RRPQ-C and RRPQ-P). General Hospital Psychiatry, 24, 336–342.

25.

Kassam-Adams

Newman

(2005). Child and parent reactions to participation in clinical research. General Hospital Psychiatry, 27, 29–35.

26.

Kuyper

de Wit

J. F.

Adam

P. F.

Woertman

(2012). Doing more good than harm? The effects of participation in sex research on young people in the Netherlands. Archives of Sexual Behavior, 41, 497–506.

27.

Legerski

Bunnell

S. L.

(2010). The risks, benefits, and ethics of trauma-focused research participation. Ethics & Behavior, 20, 429–442.

28.

Miller

F. G.

(2003). Ethical issues in research with healthy volunteers: Risk-benefit assessment. Clinical Pharmacology and Therapeutics, 74, 513–515.

29.

Newman

Kaloupek

D. G.

(2004). The risks and benefits of participating in trauma-focused research studies. Journal of Traumatic Stress, 17, 383–394.

30.

Newman

Willard

Sinclair

Kaloupek

(2001). The costs and benefits of research from the participants’ view: The path to empirically informed research practice. Accountability in Research, 8, 27–47.

31.

Rossiter

K. R.

(2012). Victimization, trauma, and mental health: Women’s recovery at the interface of the criminal justice system and mental health systems (Unpublished doctoral dissertation). Simon Fraser University, Burnaby, British Columbia, Canada.

32.

Ruzek

J. I.

Zatzick

D. F.

(2000). Ethical considerations in research participation among acutely injured trauma survivors: An empirical exploration. General Hospital Psychiatry, 22, 27–36.

33.

Seedat

Pienaar

W. P.

Williams

Stein

D. J.

(2004). Ethics of research on survivors of trauma. Current Psychiatry Reports, 6, 262–267.

34.

Shore

R. C.

Brasfield

Febres

Cornelius

T. L.

Stuart

G. L.

(2012). A comparison of three different scoring methods for self-report measures of psychological aggression in a sample of college students. Violence and Victims, 27, 973–990.

35.

Sikweyiya

Y. F.

Jewkes

(2012). Perceptions and experiences of research participants on gender-based violence community based surveys: Implications for ethical guidelines. PLoS ONE, 7, e35495.

36.

Smith

P. H.

Danis

Helmick

(1998). Changing the health care response to battered women: A health education response. Family and Community Health, 20, 1–18.

37.

Smith

P. H.

Earp

J. A.

DeVellis

(1995). Measuring battering: Development of the Women’s Experience With Battering (WEB) scale. Women’s Health, 1, 273–288.

38.

Smith

P. H.

Thornton

DeVellis

Earp

Coker

(2002). A population-based study of the prevalence and distinctiveness of battering, physical assault, and sexual assault in intimate relationships. Violence Against Women, 8, 1209–1232.

39.

Straus

M. A.

Douglas

E. M.

(2004). A short form of the Revised Conflict Tactics Scales, and typologies for severity and mutuality. Violence and Victims, 19, 507–520.

40.

Valjee

S. R.

Collings

S. J.

(2015). Convergent validity assessment of a brief screen for poverty with South African adolescents. Social Behavior and Personality, 43, 875–880.

41.

Valjee

S. R.

Collings

S. J.

(2016). The meaning of developmental trauma: Validation of a brief screen for developmental trauma appraisals. South African Journal of Psychology, 46, 1–13.