Generally, studies have revealed that only a minority of people are bothered by participation in research on traumatic stress. Severity of traumatic events and subsequent responses are typically unrelated to negative reactions. We included 386 family members and caregivers (respondents) of people with intellectual and developmental disabilities (focus people). Focus people (ages 4–82) had a wide range of physical and intellectual disabilities, medical and behavioral problems, and exposure to potentially traumatic events. The measures of impact of research participation (based on J. I. Ruzek & D. F. Zatzick's  Reactions to Research Participation Questionnaire [RRPQ]; S. Folkman and R. S. Lazarus's [1986, 1988] Emotional Responses to Participation Scale) showed good psychometric properties. Response to participation was highly skewed toward good understanding of informed consent, valuing participation, and minimal negative reactions. Number of traumatic events was related, positively, to only one RRPQ subscale: Valuing Participation. Implications for research and clinical work are discussed.
Research in the field of trauma can have considerable benefits and provide useful information; however, a significant concern when conducting trauma research is the health and well-being of those who participate. Institutional review boards are often concerned that participants in trauma research are in a fragile and unique state and that the research protocol will somehow traumatize them again (Newman & Kaloupek, 2004; Ruzek & Zatzick, 2000). One way to address these concerns is to include an assessment of participants' responses to the research protocol. An instrument for accomplishing this is the Reactions to Research Participation Questionnaire (RRPQ), an initial version of which was developed and evaluated by Newman, Willard, Sinclair, and Kaloupek (2001). They began with a 60-item RRPQ, based on multiple constructs gleaned from the research ethics literature, and then suggested an 18-item short form, with a recommendation of five additional items to field test. Various versions of the RRPQ items have been used in several studies concerning the reactions of trauma survivors to research participation (Kassam-Adams & Newman, 2002; Ruzek & Zatzick, 2000; Widom & Czaja, 2005), as have related, independently developed evaluations of research response (Carlson et al., 2003; Griffin, Resick, Waldrop, & Mechanic, 2003; see Newman & Kaloupek, 2004, for a comprehensive review). Reviews and discussions of the literature on participation in trauma research have consistently pointed to the perceived personal and global benefits of, as well as personal satisfaction with, research participation. Moreover, and most critically, participants have reported minimal negative impact from research protocols; such impact is typically related to specific participant vulnerabilities, such as type of trauma, gender, socioeconomic status, and diagnostic status (Newman & Kaloupek, 2004; Newman, Risch, & Kassam-Adams, 2006; Widom & Czaja, 2005).
Most studies to date have included adult participants who have been exposed to specific traumatic events, such as sexual assault, domestic violence, motor vehicle accidents, war, and terrorism; other adult studies have covered events occurring throughout the participants' lives (see Newman & Kaloupek, 2004). Kassam-Adams and Newman (2002, 2005) studied the responses of children and their parents to research on the traffic-related or traumatic injury of the child. Their results showed that only 5% of children and parents reported distress from participation, with most children and parents having been glad they participated and feeling good about helping others.
Information concerning the response to research participation by members of underserved populations is notably lacking. One such group is individuals with intellectual and developmental disabilities (IDD). Other than a handful of studies that are largely record reviews concerning sexual and physical assault (see Bowman, Scotti, & Morris, 2010), a dearth of research exists on the multiple types of traumatic events occurring in the lives of people with IDD. To determine the full range of traumatic events and their impact on people with IDD, Scotti et al. (2007) conducted a mail survey completed by the caretakers (primarily parents) of children and adults with IDD. Similar to Kassam-Adams and Newman's (2002, 2005) focus on parents of trauma survivors, our focus was on assessing the response of parents and caregivers to participation in that traumatic stress survey on their children and those in their care. We sought to document how research concerning trauma on the underserved and vulnerable population of people with IDD might affect the participating parents and caregivers (i.e., is participation in such research distressing?)
Parents, guardians, or care providers (i.e., respondents) completed the survey packets about their family member or care recipient (regardless of age) with an IDD (i.e., focus person) as part of a larger survey concerning traumatic events and their impact on the lives of people with IDD (Scotti et al., 2007). In the larger study, we evaluated the relation between levels of IDD, history of traumatic experiences, and severity of behavior problems, with the goal of understanding the parallels between behavior problems and symptoms of posttraumatic stress disorder (PTSD). In this article, we present only those survey components relevant to the current research question.
Respondent and focus person characteristics
The respondents were first asked to provide the following information: (a) demographics on the respondent and focus person (e.g., age, gender, education, income, living arrangements), (b) intellectual disability of the focus person (e.g., autism, mental retardation), and (c) the level of daily living skills and behavioral problems of the focus person (e.g., level of independence in eating, dressing, toileting, and communication; rating of severity of self-injurious behavior, aggression, and property destruction). Most items were completed by entering information (e.g., age), circling an option (e.g., gender), or indicating yes or no (e.g., mental retardation). Level of independence in daily living skills was rated on a 5-point scale (1 = needs complete assistance, 5 = totally independent), and occurrence of behavioral problems was indicated as no or yes; if the response was yes, a severity rating was provided (1 = not at all severe, 5 = very severe).
History of Psychosocial Stressors
We used the History of Psychosocial Stressors (HPS; Scotti, 1992) to evaluate the exposure of the focus person to 35 potentially traumatic events (e.g., motor vehicle crashes, floods, fires, physical or sexual abuse, witnessing domestic violence). For each event, the respondent reported whether the focus person was ever exposed to that event, resulting in a total score ranging from 0 to 35 (respondents had the option to provide details of the event). The HPS was originally developed by Scotti (1992); several studies have evaluated its psychometric properties. One-week test–retest reliability was established by Scotti, r(64) = .90, and by Ruggiero, Del Ben, Scotti, and Rabalais (2003), rs(90) = .82–.90. Total number of events endorsed on the HPS has been shown to be positively correlated with measures of depression, PTSD, and general anxiety (Ruggiero et al., 2003; Scotti, 1992; Scotti et al., 2000), demonstrating construct validity.
Reactions to research participation questionnaire
As noted earlier, multiple versions of the RRPQ exist. We used a version based on that used by Ruzek and Zatzick (2000), adding additional items concerning the informed consent process and asking about specific components of the study that might be bothersome. The questions focused on three aspects of research participation (see Table 1). Six questions covered understanding of the informed consent process (e.g., “The consent form clearly described this study,” “I felt I could stop participating in this study at any time”). Another six items concerned the experience and value of participating in the survey (e.g., “I gained something positive from participating,” “I would participate in future studies like this one.”). An additional five questions asked how bothered the respondent was in general by the survey, as well as by specific parts of it (e.g., psychological problems, personal characteristics and daily life, traumatic events). All items were rated on a scale ranging from 1 (strongly disagree) to 5 (strongly agree); five items were then reverse scored so that higher RRPQ scores reflected a more positive participation experience (possible range = 17–85).
Using the data from this study, we conducted an initial principal-components analysis (varimax rotation) for the purpose of data reduction (Bandalos & Boehm-Kaufman, 2009), which revealed that all RRPQ items should be retained because they related to one of three factors (56.2% of variance accounted for). As displayed in Table 1, a subsequent principal-axis factor analysis (varimax rotation) resulted in the same three factors (clearly reflecting the face validity of the items), accounting for a total of 47.1% of the variance: Factor 1, Understanding of the Consent Process (Items 1–6; 19.1%); Factor 2, Personal Value of Participation (Items 7–12; 17%); and Factor 3, Being Bothered by Participation (Items 13–17; 11%). The Cronbach's alphas for the three factors were (a) for Factor 1, .86 (corrected item-total rs = .62–.72), (b) for Factor 2, .83 (corrected item-total rs = .50–.71); and (c) for Factor 3, .73 (corrected item-total rs = .42–.57). These psychometric properties support the use of the expanded RRPQ in this study.
Emotional responses to participation scale
Folkman and Lazarus (1986, 1988), in their extensive research on stress and coping, established the psychometric properties of varying sets of emotional words and their relation to levels of depression and coping with stressful situations. We selected the most commonly used words from their research to evaluate respondents' emotional reactions while completing the survey. The Negative Emotions subscale (NES) of the Emotional Responses to Participation Scale (ERPS) included feeling angry, anxious, disappointed, disgusted, fearful, guilty, sad, and worried. The Positive Emotions subscale (PES) included feeling confident, eager, excited, happy, hopeful, pleased, and relieved. Respondents rated these words on a Likert scale on which 1 = not at all and 5 = a great deal. We derived total scores by separately summing the ratings for negative and positive emotions. The data from this study had Cronbach's alphas of .82 for PES (corrected item-total rs = .34–.74) and .84 for NES (corrected item-total rs = .46–.70).
In collaboration with the West Virginia Department of Health and Human Resources, Bureau for Behavioral Health, Division of Mental Retardation, surveys were mailed to the parent or guardian of 2,000 individuals identified as having IDD (i.e., mental retardation, autism). These individuals were randomly selected from the West Virginia Medicaid Waiver List (which included approximately 4,300 people). To keep all names and addresses confidential, we provided 2,000 survey packets to Department of Health and Human Resources staff; the staff printed labels from the list and conducted the mailings. Three mailings were performed, as recommended by Dillman (2000). First, a postcard was mailed to inform the individuals that a survey was coming. One week later, the survey was mailed, followed one week later by a reminder postcard. A total of 420 surveys were returned. Of these, 405 surveys (20% return rate) were sufficiently complete to include in the main study; however, 19 were dropped for purposes of these analyses because of failure to complete the RRPQ. The final sample size was 386 respondents.
Respondent and Focus Person Characteristics
Of the respondents, 317 (82%) were a parent of the focus person, another 33 (9%) were a relative of the focus person (sibling, aunt or uncle, grandparent), 8 were guardians (2%), and the remainder were caseworkers or direct care staff (n = 27; 7%). The reported educational level of the respondents was 59% high school or GED, 15% some college, 18% bachelor's degree, and 8% postgraduate degree. The average age of the respondents completing the surveys was 50 years (n = 362, SD = 11.8, range = 23–86 years), with 86% being female. The average age of the focus people was 22.1 years (n = 385, SD = 14.9, range = 4–82 years, 64% younger than age 22, quartiles of 11, 18, and 27 years), with 60% being male. Most people in both groups were Caucasian (94% of focus people, 96% of respondents), reflecting the demographics of West Virginia.
Mental retardation was reported by the respondents as the primary diagnosis for 94% of the focus people; 6% were reported to have a primary diagnosis of autism. The mean score for independence in daily living skills was 2.9 (SD = 1.1, Mdn = 3.1, range = 1–5), which was the midpoint of the scale (ranging from needs total assistance to totally independent) and consistent with the functional deficits associated with IDD. Respondents reported that focus people exhibited a mean of 6.5 different behavioral problems (SD = 5.4, Mdn = 5.0, range = 0–25); 92% reported one or more behavioral problems, with these people receiving a mean behavior severity rating of 2.7 (SD = 0.9, Mdn = 2.7, range = 0–5), just below the midpoint of the two ends of the scale (ranging from not at all severe to very severe). The behavior problems most commonly reported were tantrums (54%), irritability (42%), and yelling (39%).
Respondents reported that the focus people were exposed to 0–14 potentially traumatic events (M = 2.8, SD = 2.7, Mdn = 2.0), with 21% reporting no exposure to these events. The most frequently reported events were death of a family member (44%), life-threatening illnesses (33%), motor vehicle crashes (22%), and emotional abuse (18%). We found a positive correlation between number of potentially traumatic events and the number of behavioral problems, r(354) = .20, p < .001.
Reactions to Research Participation Questionnaire
The mean total score on the RRPQ was 70.6 out of a maximum score of 85 (SD = 9.9, range = 34–85, Mdn and mode = 73), with a higher score meaning more positive reaction to research participation. The distribution of scores was skewed (−1.04) toward the high end of the scale, with 25% of respondents scoring 78 or higher and 25% scoring lower than 67. Factor 1 (Understanding the Consent Process) had a mean score of 27.0 (SD = 5.0) of a possible 30, indicating that respondents somewhat to strongly agreed that they understood the components of the consent form. For Factor 2 (Personal Value of Participation), the mean score was 23.8 (SD = 5.2) of a possible 30, suggesting that respondents somewhat agreed that they found personal value in participating in this research. Finally, the mean score on Factor 3 (Being Bothered by Participation) was 19.8 (SD = 4.8) of a possible 25, indicating that respondents somewhat agreed that they were not bothered either in general by the survey or by the particular types of questions on the survey. Total scores for Factors 1 and 2 were significantly correlated with each other, r(386) = .46, p < .001; Factor 3 was not correlated with the other two factors.
Emotional Responses to Participation Scale
The respondents (n = 383) rated seven positive and eight negative emotions they may have felt while completing the survey, with higher scores indicating feeling more of an emotion. PES had a mean rating of 16.0 (SD = 6.2, range = 7–35, skew = 0.49) of a possible score of 35 (with 75% of respondents falling below the midpoint of 21). For NES, the mean rating was 11.4 (SD = 5.0, range = 8–40, skew = 1.8) of a possible score of 40 (with 75% of respondents falling below 14). We found no statistically significant correlation between the two scales. These scores indicate that the respondents, for the most part, experienced little to no emotion, positive or negative, while completing the survey. PES had a small positive correlation with number of traumatic events, r(383) = .10, p < .05. NES had a small correlation with number of behavior problems, r(351) = .12, p < .05, and a more meaningful correlation with number of traumatic events, r(383) = .21, p < .001. Examining NES, those in the highest quartile (score > 16) reported more traumatic events (M = 3.6, SD = 3.1, n = 98) than the remainder of the sample (M = 2.6, SD = 2.6, n = 288), t(384) = −3.1, p < .01 (we found no differences on any of the other variables).
Correlations with Reactions to Research Participation Questionnaire
Analysis of variance revealed that gender of the focus person, gender of the respondent, and relationship between respondent and focus person were all unrelated to scores on the total RRPQ and the three subscales. The total score on the RRPQ was correlated with four variables: age of focus person, r(385) = −.13, p < .01; age of respondent, r(362) = −.13, p < .02; NES, r(383) = −.26, p < .001; and PES, r(383) = .21, p < .001. Total score was not significantly correlated with education level of the respondent or other characteristics of the focus person, including total number of traumatic events, daily living skills score, number of behavioral problems, and behavior severity rating.
Scores on Factor 1 (Understood the Consent Process) were not significantly correlated with number of traumatic events, nor with any respondent characteristics (age, education), NES, or PES. Two variables correlated with Factor 2 (Personal Value of Participation): PES, r(383) = .29, p < .001, and number of traumatic events, r(386) = .11, p < .05. Finally, Factor 3 (Being Bothered by Participation) correlated only with NES, r(383) = −.47, p < .001; number of traumatic events was unrelated to this factor. Despite those experiencing more negative emotions also being more bothered by aspects of the survey, we found no statistically significant correlation between Factor 2 (Personal Value of Participation) and Factor 3 (Being Bothered by Participation). Moreover, we found no differences on any variables between the lowest quartile on Factor 3 and the remainder of the sample.
The study included 386 respondents, primarily mothers of the focus people. The focus people (two thirds of whom were children younger than age 22) were documented to meet criteria for IDD (including functional deficits and behavioral problems) and to have been exposed to a range of potentially traumatic events. The two measures of impact of research participation, the RRPQ and ERPS, were shown to have good psychometric properties in terms of both internal consistency and concurrent validity, similar to Kassam-Adams and Newman (2002, 2005) and Newman and colleagues (2001). Our version of the RRPQ was based on that of Ruzek and Zatzick (2000); they did not report psychometrics, but the factors found here match the content they sought to evaluate.
The total score on the RRPQ, reflecting the overall experience of participating in the survey, was highly skewed toward a very favorable response to participation. Few studies have included a wide age range in their participant sample; those that have done so have found more reported distress by the youngest and oldest participants (Newman & Kaloupek, 2004); we found a similar but very weak relation, with older respondents reporting a less favorable response to participation. Kassam-Adams and Newman (2005) included children and parents as participants, both asking about the parents themselves (e.g., injury severity, acute stress disorder symptoms) and asking the parents about the children's behavioral and emotional response and trauma histories. They found no relation between parent distress and child's age, likely because of the restricted age range of the children. The age range of the focus people in this study was wide (4–82 years); focus people were family members of 93% of our respondents. We found a weak negative correlation between total score on the RRPQ and age of the focus person. These results replicate prior studies for respondent age and extend research to a wider age range for focus people (i.e., children of respondents; cf. Kassam-Adams & Newman, 2005).
As makes sense, lower RRPQ scores were related to more negative emotion, and higher scores were related to more positive emotion. Exposure to more potentially traumatic events was related to greater negative emotional response (with a weak relation to positive emotional response). In the context of a survey asking respondents to provide information about the focus people's trauma history, number of traumatic events was unrelated to total RRPQ scores. Of interest, Personal Value of Participation was positively associated with number of traumatic events. Thus, despite these difficult aspects of the focus people's lives, the respondents reported having valued participating in research on this topic, as well as having experienced more positive emotion. Also positively associated with Personal Value of Participation was Understood the Consent Process. Although Being Bothered by Participation was associated with negative emotions, it is important to note that number of traumatic events was unrelated to scores on this factor. Moreover, being bothered by aspects of the study was unrelated to valuing participation and understanding the voluntary nature of the study. Finally, educational level and age of the respondent did not relate to respondents' ability to understand aspects of the consent process (Factor 1).
A limitation of this study is the 20% response rate to the mail survey. However, although this response rate is on the low end, it is within the range of response rates to mail surveys (Dillman, 2000; Dillman & Frey, 1974; Shih & Fan, 2008). We did follow the procedure recommended by Dillman (2000) for increasing mail survey response rates, and we note that respondents to mail surveys appear more likely to use the full range of rating scales compared with telephone surveys (Dillman, 2000). The low response rate could, in part, be related to the respondents (primarily family) having the demands of caring for a person with IDD; thus, this population of respondents may have had less time to devote to a survey of this length.
Given this and other self-selection biases that are problematic for social and behavioral science research, it is important to show that the final sample is similar in key demographics to the original population. Recall that the surveys were mailed to a random sample of 2,000 parents and caregivers of individuals with IDD from a larger list of more than 4,300 people. The final sample of 386 is similar to the random sample of 2,000 on the two known demographics: (a) living with family (91% vs. 90%, respectively) and (b) focus person younger than age 22 (59% vs. 50%, respectively). Finally, our sample of 386 focus people was 60% male, which is consistent with multiple studies of gender differences in IDD (see Coutinho & Oswald, 2005). Thus, any self-selection bias was unrelated to the relationship of the respondent to the focus person and to focus person age or gender.
The data presented here indicate that participation in a detailed survey concerning the trauma history of one's child with IDD (or a person with IDD for whom one is a care provider) does not negatively affect the respondent in terms of general emotional reaction or reactions to specific aspects of the survey. In fact, participation was positively valued. Note that we found that 79% of the focus people in this study were reported to have experienced at least one potentially traumatic event (M = 2.8 events) and that number of events was positively related to number of behavior problems. Given the lack of existing data concerning exposure of people with IDD to a range of potentially traumatic events and the possibility of related disorders such as PTSD, these findings are in themselves important. We recommend that both researchers and clinicians ask about such events and evaluate their possible relations to behavioral issues. This evaluation is important when developing behavioral support plans because typical interventions for behaviors thought to have functions such as attention or escape from demands are not likely to well address behaviors that are, instead, related to trauma (e.g., PTSD, depression, phobia). Because our findings related to the RRPQ suggest that inquiries about trauma history are valued and are not related to significant levels of distress in the parents and caregivers of people with IDD, researchers and clinicians may feel more at ease addressing this topic.
Regardless of the lack of documented negative impact, we recommend, as have others, that measures of reaction to research participation be included in all research, not just that related to trauma or potentially deleterious factors. Ongoing research has suggested that the number of questions needed to evaluate impact can be significantly reduced, thus making such evaluation more efficient and making it difficult to argue against its inclusion (Scotti et al., 2008).
This research was supported by funding from the West Virginia Department of Health and Human Resources, Bureau for Behavioral Health and Health Facilities, which is gratefully acknowledged. We are particularly indebted to Faith Stuart for assistance with funding and Frank Kirkland for assistance with the process of mailing the surveys. Portions of this research were presented at the Conference on Innovations in Trauma Research Methods (2007) and the International Society for Traumatic Stress Studies (2007). We thank Ashley Cavender, Amy Morford, Steven Burkhart, Jessi Kalvitis, Susie Nicholson, and other members of the Traumatic Stress Research Group for assistance with various aspects of this project.
Editor-in-Charge: Philip Ferguson
Joseph R. Scotti (e-mail: Joseph.Scotti@mail.wvu.edu), Traumatic Stress Research Group, Department of Psychology, Box 6040, West Virginia University, Morgantown, WV 26506-6040, USA; Sarah B. Stevens, Vanessa M. Jacoby, Magdalene R. Bracken, Rachel Freed, and Elizabeth Schmidt, West Virginia University.