Development and Validation of the Prolonged Activation and Anticipatory Race-Related Stress Scale

Abstract

This article describes two separate studies that were conducted to develop and validate a measure of the prolonged stress activation and anticipatory race-related stress response in African American adults (Prolonged Activation and Anticipatory Race-Related Stress Scale [PARS]). In Study 1, an exploratory factor analytic procedure (N = 292) resulted in a17-item measure with four underlying factors: (a) Perseverative Cognition, (b) Secondary Appraisal, (c) Anticipatory Race-Related Stress Scale–Psychological, and (d) Anticipatory Race-Related Stress Scale–Physiological. In Study 2 (N = 227), a confirmatory factor analytic procedure was conducted to evaluate and compare the underlying factor structure for several competing models of the PARS. This procedure supported a four-factor oblique solution as having the best fit to the data. Study 2 also provided evidence for the convergent validity of the PARS in that its factor scores correlated, in the anticipated direction, with scores on measures of related constructs.

Keywords

prolonged activation race-related stress racism anticipatory racism

There is substantial empirical literature indicating that African Americans tend to have higher rates of stress-related diseases (e.g., hypertension, cardiovascular disease, stroke, and cancer) with more negative outcomes in comparison with other racial/ethnic groups (Brondolo, Gallo, & Myers, 2009; Davis, Liu, & Gibbons, 2003; Franks, Muennig, Lubetkin, & Jia, 2006; Geronimus, Hicken, Keene, & Bound, 2006; McCord & Freeman, 1990; Shavers & Shavers, 2006). Moreover, there is increasing evidence to suggest that these health disparities may stem from African Americans’ chronic exposure to race-related stressors (Bowen-Reid & Harrell, 2002; Gee, 2002; Karlsen & Nazroo, 2002; Kwate, Valdimarsdottir, Guevarra, & Bovbjerg, 2003; Merrit, Bennett, Williams, Edwards, & Sollers, 2006; Paradies, 2006; Pieterse & Carter, 2010; Utsey & Hook, 2007; Utsey & Ponterotto, 1996; Williams, 1999; Wyatt et al., 2003). The link between chronic exposure to race-related stress and the health of African Americans has important implications for counseling psychologists and other mental health professionals working with this population. Therefore, it is imperative for researchers to develop more precise conceptual models as well as valid and reliable instruments for assessing race-related stress.

Because of the significance of race-related stress and its potential consequences for the health and well-being of African Americans, the literature in this area has burgeoned, including conceptual models (e.g., Clark, Anderson, Clark, & Williams, 1999; Harrell, 2000; Krieger, 1999; Outlaw, 1993; Utsey, Giesbrecht, Hook, & Stanard, 2008) and assessment instruments (see Krieger, 1999, and Utsey, 1998, for a review). Most of the existing models of race-related stress are derived from Lazarus and Folkman’s (1984) classic transactional theory of stress and coping.

The Lazarus and Folkman model provides a basic framework for understanding stress and its consequences as a function of the dynamic process (transaction) between the person and his or her environment. A central component of this model is cognitive appraisal. Cognitive appraisal is the mechanism by which a person evaluates an encounter with the environment as a threat or benefit to his or her well-being. There are two types of cognitive appraisal: primary appraisal evaluates the potential risk for harm and secondary appraisal evaluates what resources can be summoned to prevent/reduce harm (Lazarus & Folkman, 1984). Outlaw (1993) would later modify the Lazarus and Folkman model to include unique racial and cultural factors that influence the appraisal of race-related stressors in African Americans. According to Outlaw, a race-related stressor is a transaction between an individual and their environment that is rooted in racism. Harrell (2000) further developed the race-related stress paradigm, rooted in the Lazarus and Folkman model, to include antecedent variables (e.g., race and socioeconomic status), family and socialization influences (e.g., familial dynamics and race-related messages received from family), sources of stress (race-related, other group membership), internal and external mediators (e.g., internal characteristics and external resources), and outcomes (e.g., physical, psychological, and social). Outlaw’s (1993) and Harrell’s (2000) conceptual contributions to the Lazarus and Folkman model demonstrate the importance of attending to racial and cultural factors throughout the stress experience, in particular in respect to race-related stress.

Despite the popularity of Lazarus and Folkman’s model, including the modifications made by Outlaw (1993) and Harrell (2000), this theoretical framework alone does not fully elucidate the dynamic process between an individual and race-related stressors in the environment. While Lazarus and Folkman (1984) provide a model of how people assess and attempt to cope with stress, their model does not venture to explain how and why some individuals remain in a state of stress-induced arousal for extended periods of time. Some scholars have suggested that the chronically elevated stress levels that are associated with the negative health outcomes observed in African American populations do not typically stem from transient stressors, but rather, from the prolonged activation of stress (Brosschot, Gerin, & Thayer, 2006; Brosschot, Pieper, & Thayer, 2005). Similarly, although models of race-related stress imply that racism tends to have lasting effects (such as hypervigilance) on those who experience it (e.g., Harrell, 2000), these models do not provide a framework for understanding the mechanism by which this occurs. Therefore, what is needed is a model that considers the degree to which prolonged activation of race-related stress increases the likelihood of negative psychological and physical health outcomes for African Americans.

In addition to solid conceptual models of race-related stress, empirically valid and reliable measures are of critical importance for assessing the degree to which race-related stress is linked to health disparities among African Americans. There are a plethora of psychometric instruments intended to measure the stress experienced as a consequence of an individual’s race and/or encounters with racism (see Utsey, 1998 for a review). In general, existing measures of race-related stress tend to focus on the number of race-related stressors an individual has encountered and the primary stress appraisal that accompanied the stressor. Such measures, however, tend to conceptualize race-related stressors as having a distinct beginning and end, and the stress response is assumed to be discrete.

Several researchers in the area of racism, discrimination, stress, and health have noted methodological limitations with existing measures of race-related stress (Krieger, 1999; Kressin, Raymond, & Manze, 2008; Landrine, Klonoff, Corral, Fernandez, & Roesch, 2006; Meyer, 2003; Smith, 2002; Utsey, 1998). Some common methodological problems among existing measures include ambiguous and overly general items (Kressin et al., 2008; Krieger, 1999), yes/no item response choices that assess only the presence or absence of a race-related event, but ignore appraisal of the stressor (Landrine et al., 2006), and closed questions that may result in over- or underreporting of events (Smith, 2002). A more problematic issue with existing measures of race-related stress is an overreliance on the checklist approach to assess specific race-related events, which also has a high likelihood for underreporting events. Most checklist measures of race-related stress have between 30 and 50 items. As such, if an individual has experienced a race-related event that is not represented on the checklist, no matter how significant and stressful the event, underreporting will result. Checklist measures have also demonstrated bias with regard to geographic region (e.g., you had difficulty catching a taxi), gender (e.g., you were treated like a criminal), socioeconomic status (e.g., you had trouble getting a loan), and occupational status (you were passed over for an important assignment at work). What is needed, then, is a measure of race-related stress that does not focus exclusively on the primary appraisal of the event, does not use ambiguous and overly general items, and is not dependent on a demographically biased checklist approach to assessing race-related events.

Psychometrically valid and reliable measures are critical to conducting meaningful research in the area of race-related stress and health. The current article describes the development and validation of a measure of race-related stress intended not only to assess the initial (primary) appraisal of stress but to also to capture the secondary appraisal process and the prolonged activation stress response, including anticipatory race-related stress. This measure is based primarily on the conceptual framework of the prolonged activation of stress model proposed by Brosschot and colleagues. However, there are some conceptual contributions from the stress and coping model posited by Lazarus and Folkman (1984). The measure is not intended to assess the cumulative effects of race-related stress, but the process of the prolonged activation stress response in the context of a specific race-related event.

Proposed Conceptual Frameworkand Measurement Model

Prolonged activation refers to the extended physiological experience of stress above and beyond the duration of the actual stressor (Brosschot et al., 2005). Most existing models of stress ignore the duration of the stress response. According to Brosschot et al., the experience of stress may begin far before the stressor is actually encountered, and last long after the stressor ends. Based on this conceptualization, perseverative cognition and anticipatory stress are codetermining variables connecting stress appraisal to the prolonged physiological stress process (Brosschot et al., 2005; Brosschot, Van Dijk, & Thayer, 2002; Monat, Averill, & Lazarus, 1972).

Perseverative cognition involves mentally retaining a stressor, keeping it cognitively active, and thereby prolonging the physiological and psychological stress response associated with the event (Brosschot et al., 2005; Brosschot et al., 2006). Brosschot and colleagues hypothesized that perseverative cognition serves to convert stressors into the prolonged physiological activation that may trigger the onset of stress-related disorders. The construct of perseverative cognition is intimately related to the cognitive processes of worry and rumination (Brosschot et al., 2005; Brosschot et al., 2006). Perseverative cognition not only prolongs the experience of stress but also increases anxiety by keeping an individual in a hypervigilant state in which he or she is thinking excessively about the stressor and possible responses to it (Brosschot et al., 2006).

Brosschot et al. (2006) suggest that the process of perseverative cognition is most common when an individual is faced with stressors that seem to be out of his or her personal control, and when coping strategies or resources are limited. Health psychologists have proposed a link between the perceived inability to control external stressors (i.e., through the process of secondary appraisal) and negative affect, anger, depression, general mistrust, hypervigilance, and increased physiological arousal (Sharpley, 2002). A study by Borders and Liang (2011) found that ethnic minorities who reported perceived discrimination were more likely to engage in angry rumination. Moreover, they found that angry rumination partially mediated the relationship between perceived discrimination, emotional distress, and aggressive behavior. Similarly, a study by Hatzenbuehler, Nolen-Hoeksema, and Dovidio (2009) found that the stigma-related stressors resulted in increased rumination in African Americans. There is a cyclical nature to this association in that stressors which are appraised as uncontrollable in the process of secondary appraisal lead to more perseverative cognition, and increased perseverative cognition leads to heightened feelings of uncontrollability and prolonged stress activation (Brosschot et al., 2006).

Secondary appraisal is an evaluative process that takes into account the nature of the stressor and the available coping resources. When faced with a stressor, individuals engage in secondary appraisal by considering their available coping options and assessing the likelihood that any given coping strategy will be effective (Lazarus & Folkman, 1984). This evaluative sequence is a central feature of the stress and coping process, and has implications for the effective regulation of the stress response. It should be noted that an individual’s assessment of their available coping resources has direct impact on their emotions and their subsequent behavior in light of the stressor.

Anticipatory stress involves the subjective experience of stress prior to actually encountering the stressor. It is closely related to the threat appraisal described by Lazarus and Folkman (1984), and creates a prolonged physiological arousal in response to the anticipated stressor (Brosschot et al., 2005). Nearly four decades ago, Monat et al. (1972) posited that it was not the stressor itself, but the uncertainty of the occurrence of the stressor that most strongly influences stress and anxiety. Their research determined that even when the actual stressor remains constant, the level of uncertainty of encountering the stressor may play a significant role in the experience of stress (Monat et al., 1972).

Anticipatory stress has not received adequate attention in the stress and coping literature (Brosschot et al., 2005; Brosschot et al., 2006), and the few studies that have been conducted have not considered how race-related stressors would affect the anticipatory responses. However, due to the insidious and omnipresent nature of racism in American society, many African Americans are socialized to maintain a state of vigilance, especially in environments that are perceived as potentially threatening (Comas-Diaz & Jacobsen, 2001). This vigilant watch for potential race-related stressors in the environment is typically unconscious and results from chronic exposure to racism and discrimination (Harrell, 2000; Outlaw, 1993). Even if an individual never personally experiences a race-related event, learning about the experiences of others (vicarious, chronic-contextual, collective, and/or transgenerational) creates an awareness or expectation that he or she may encounter racism at any time (Harrell, 2000). Consequently, this expectation of stress induces a prolonged stress response, which may lead to negative health outcomes (Brosschot et al., 2005; Brosschot et al., 2006). Hence, the conceptual application of the prolonged activation of stress model produces the process of anticipatory race-related stress, a central feature of the prolonged activation of race-related stress response.

A measure of race-related stress that captures the prolonged activation and anticipatory response dimension of the stress process would contribute much to our ability to assess the degree to which race-related stress is a contributing factor in the health disparities of African Americans. Moreover, newly developed measures of race-related stress should address the limitations inherent in existing measures (e.g., checklist approach, gender and class bias, and items that are vague and overly general). By improving the conceptual and measurement models for assessing race-related stress, researchers will be better equipped to assess the link between exposure to race-related stressors and psychological and physical well-being in African Americans. This article combines parts (i.e., secondary appraisal) of the transactional stress theory proffered by Lazarus and Folkman (1984) with the prolonged activation theory of stress posited by Brosschot et al. (2005) and Brosschot et al. (2006) to provide a more comprehensive conceptual framework for assessing race-related stress. The conceptual framework being proposed posits that the prolonged activation of race-related stress is the function of a process that is composed of four distinct components; these are as follows: perseverative cognitions about a race-related event, secondary appraisal (i.e., evaluating one’s resources for managing a race-related stressor), anticipation of future race-related stressors, and bodily alarm response to an impending race-related stressor.

Based on the expanded conceptual framework of race-related stress described above, a measure of the prolonged activation stress response to race-related stressors was developed over three separate studies. It was hypothesized that the underlying factor structure of the measure being developed would load items to factors consistent with the proposed four-component model of the prolonged activation of race-related stress described above. Furthermore, to establish some preliminary evidence of the measure’s construct validity, it was hypothesized that its component scores would correlate, in the anticipated direction, with scores on measures of related constructs. Study 1 established some preliminary evidence for the underlying factor structure of the Prolonged Activation and Anticipatory Race-Related Stress Scale (PARS, see the appendix). Study 2 examined how well the PARS factor structure fit the data, its convergent validity, and internal consistency. Study 2 established some preliminary evidence of the PARS test-retest reliability.

Initial Scale Development

Item Development

The initial pool of items for the PARS was derived from a review of existing measures of race-related stress, a review of the related literature, and a review of the prolonged activation of stress literature—this includes perseverative cognition and anticipatory stress. The justification for the development of a new measure of race-related stress was in part based on the theoretical limitations of existing instruments; therefore, the process of generating items for the PARS was conceptually driven. Based on the conceptual models proposed by Brosschot et al. (2005) and Lazarus and Folkman (1984), PARS items were developed to reflect the following conceptual components of the prolonged activation stress process: perseverative cognition of race-related stressors, secondary appraisal of race-related stressors, and anticipatory race-related stress responses. As noted, the PARS is not intended to measure the cumulative effects of multiple race-related events across time, as the majority of existing measures do, but to capture the process of the prolonged activation stress response to race-related stress in the context of a specific race-related event. The initial PARS prototype consisted of 25 items.

Content Validity

To establish the content validity of the PARS, the 25-item prototype was evaluated by six scholars who have published conceptual and/or empirical articles in the area of race-related stress. These expert judges were asked to rate each item of the PARS for its face validity, its domain appropriateness, and clarity of wording. The expert reviewers were instructed to rate each item on a scale of 1 to 5 for its face validity (1 = Item does not appear to assess the intended construct, 5 = Item appears to assess the intended construct) and domain appropriateness (1 = Not at all related to the construct, 5 = Exact fit to construct). The same instructions were given for item clarity (1 = Item not clear at all, 5 = Item is very clear). Items that received an average rating of 3.0 or less were rewritten to improve their face validity and/or clarity. Based on this process, several items were reworded to improve both their face validity and readability. All the items were judged to be domain appropriate.

Pilot Study

Following the item development and content validity process, and after receiving approval from the institutional review board, the 25-item PARS was administered to a convenience sample of 79 African Americans from a large public university in the Mid-Atlantic region of the United States. The sample consisted of 44 women and 35 men. The participants ranged in age from 22 to 70 years, with a mean age of 40.90 years (SD = 11.71). The primary purpose of the pilot study was to evaluate the PARS for its ease of administration, the time it takes to complete the measure, clarity of its instructions, and the readability of its items. In addition, the pilot study allowed for an item analysis to determine which of the instrument’s items were functioning adequately.

The pilot study’s findings resulted in the elimination of two items. Items were eliminated if they had low item-total correlations (<.30) or little discriminatory ability as indicated by extreme high or low mean scores. The PARS had a corrected item-total correlation coefficient range of .16 to .54, and Cronbach’s alpha for the total scale was .76. It should be noted that participants did not report any difficulty in completing the instrument, understanding its instructions, or interpreting the meaning of the items. The PARS, on average, required less than 10 minutes to complete.

As a consequence of the procedure described above, a 23-item self-report measure of race-related stress resulted. The PARS is intended to measure the prolonged activation and anticipatory stress response to race-related stressors. The measure was developed for use with African American adults from diverse backgrounds who read minimally at the seventh grade level.

Study 1

Study 1 was intended to establish the initial factor structure of the PARS using exploratory factor analytic methods. In addition, indexes of internal consistency (Cronbach’s alpha) were computed for each of the PARS subscales. Pearson’s product-moment correlation coefficients were calculated to determine the subscale intercorrelations for the PARS.

Method

Participants

The total sample for the current study consisted of 290 African Americans from the Mid-Atlantic region of the United States. Participants were recruited from introductory psychology courses at a large public university (n = 93), general psychology courses at a Historically Black University (n = 157), and from the community at large (n = 40). Surveys were administered in person and via an online data collection program. The sample consisted of 188 (64.8%) females and 99 (34.1%) males. The participants ranged in age from 18 to 70 years, with a mean age of 28.08 years (SD = 12.3).

Measures

The PARS is a 23-item self-report measure of the prolonged activation and anticipatory stress response to race-related events experienced by African Americans. To complete the measure, respondents are asked beforehand to think of a race-related stressor that they or someone close to them (network events) experienced within the past year or so. Kessler and McLeod (1984) coined the term network event to describe stress induced from a stressful event experienced not by the respondent but by someone intimately connected to the respondent through his or her social network (family and friends). Respondents are then asked to describe the event in detail in the space provided on the PARS form. The structure and format of the PARS also permits researchers to impose a specific race-related event (real or imagined) on respondents. Next, respondents are instructed to complete the 23-item PARS using the event they just described as their reference. The first item requires respondents to indicate, on a 7-point Likert-type scale, the degree to which they experienced the race-related event as stressful (1 = Not at all stressful to 7 = Extremely stressful). The next four items ask respondents to indicate, on a 7-point scale, the duration (1 = Did not think about it at all to 7 = I still think about it), frequency (1 = Not at all to 7 = More than 3 times a day), intensity (1 = Did not think about it at all to 7 = Could not stop thinking about it), and controllability (Thought about it when I didn’t mean to) of their thoughts about the event after it occurred. Higher scores reflect a more perseverative cognition about the race-related event. Items 7 through 11 ask respondents to indicate the degree to which they agree or disagree with the statements (1 = Strongly disagree to 7 = Strongly agree) about whether they felt they had the resources to manage/cope with the race-related event they had described. Higher scores indicate a more positive secondary appraisal of one’s coping resources. Items 12 through 16 are intended to assess, on a Likert-type scale (1 = Strongly disagree to 7 = Strongly agree), the degree to which respondents make a threat appraisal of the likelihood they will experience race-related stress in the future. Higher scores indicate a greater expectation that an individual will experience race-related stress in the future. Items 17 through 23 assess, on a Likert-type scale (1 = Strongly disagree to 7 = Strongly agree), the degree to which respondents respond physiologically in anticipation of an impending race-related stressor. Higher scores indicate a greater bodily alarm response (e.g., racing heart beat, shaking or sweaty hands) to an impending race-related event.

A demographic data questionnaire was completed by all participants in the study. The questionnaire asked respondents to identify their sex, age, and academic status.

Procedures

Following institutional review board approval, study participants were recruited from undergraduate introductory psychology classes and the community at large. Those recruited from the introductory psychology courses participated as part of a course requirement. There was no compensation for the community participants. The survey included a consent form that explained the study’s procedures and the rights of participants. Following consenting, participants completed the 23-item PARS. After completing the questionnaires, participants read a debriefing statement that included the contact information of the researcher in case they had any questions about the study. Individuals recruited from the introductory psychology courses were assigned course credit for their participation.

Results

Preliminary data analysis

Prior to conducting the exploratory factor analyses, the data were screened for outliers and missing data. The findings revealed that there were no outliers and approximately 5% missing data. Missing data were addressed using the listwise deletion procedure. According to Yuan and Bentler (2000), when there are fewer than 20% missing data listwise deletion is just as effective as the (preferred) full information maximum likelihood method.

Exploratory factor analysis

To establish the underlying factor structure of the PARS, an exploratory factor analysis using the principal axis factoring method was performed on the 23-item PARS. First, the data were evaluated using the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlett’s test of sphericity to determine if the sample was optimal for conducting an exploratory factor analysis. The results indicated that the current sample’s KMO (.87) and Bartlett’s test of sphericity (p < .001) were acceptable. Next, the Cattell’s (1966) scree test was used to determine the optimal number of factors to best represent the underlying structure of the PARS. The scree test indicated that a four-factor solution could be extracted. This finding was confirmed with the use of a parallel analysis, which was conducted based on 1,000 random data sets. Parallel analysis is a more conservative procedure for identifying the number of factors to retain. The procedure involves computing an average eigenvalue for each factor based on random data matrices. These eigenvalues are then compared with the actual eigenvalues derived from the data; factors are retained if their eigenvalues are larger than the eigenvalues produced by the parallel analysis (Brown, 2006). Based on the findings from the Cattell’s scree test, and the parallel analysis, three-, four-, and five-factor extractions using both orthogonal (varimax) and oblique (direct oblimin) solutions were forced. The three- and five-factor extractions were included for exploratory purposes to rule out the possibility of better alternative solutions for the PARS’ underlying factor structure. Of these extraction methods, the four-factor oblique solution loaded items to factors that were conceptually more compelling than the competing three- and five-factor solutions, or the orthogonal solutions (see Table 1). The three-factor oblique solution failed to converge, the three-factor orthogonal solution was less compelling, and the five-factor solutions loaded only two items to the fifth factor.

Table 1.

Means, Standard Deviations, Principal Axis Factor Loadings, and Communalities for PARS Items.

Item Content by Factor	M	SD	F1	F2	F3	F4	h ²
Factor 1: Perseverative Cognition
1. On a scale from 1 to 7, I would describe my experience with racism as . . .	3.04	1.57	.63	−.14	.30	−.31	.42
2. In the days/weeks after my experience with racism, I thought about it . . .	3.36	1.98	.80	−.21	.31	−.25	.64
3. Whenever I thought about my experience with racism, I would think about it for . . .	3.37	1.53	.77	−.19	.28	−33	.59
4. In the days/weeks after my experience with racism, I continued to think about it for . . .	3.86	2.52	.66	−.25	.26	−32	.47
5. I would think about my experience with racism when I did not mean to	2.72	1.30	.81	−.18	.38	−19	.66
Factory 2: Anticipatory Race-Related Stress
6. When I am around White people, I expect them to say or do something racist	3.41	1.93	.20	−.48	.31	−.06	.32
7. I believe that most Black people will experience some form of racism in the future	6.12	1.33	.19	−.63	.08	.07	.41
8. I know if I go where there are mostly White people, there is a good chance I will experience racism	4.77	1.84	.13	−.58	.28	−.06	.39
9. I believe there is a good chance that I will experience racism in the future	5.70	1.49	.22	−.68	.07	.10	.48
Factor 3: Anticipatory Bodily Alarm Response
10. I can feel my hands start to shake whenever I think I am about to experience racism	4.24	2.01	.38	−.24	.59	−.24	.41
11. I get chest pains whenever I think I am about to experience racism	2.74	1.65	.24	−.10	.69	−17	.48
12. My hands (or other body parts) sweat . . .	2.15	1.46	.32	.05	.73	−17	.56
13. I get a lump (or dryness) in my throat . . .	2.57	1.73	.30	−.10	.76	−.21	.57
Factor 4: Secondary Appraisal
14. Black people have always had to deal with these kinds of events/situations, so my experience with racism was something I could manage	5.39	1.66	−.24	−.25	−.12	.41	.24
15. At the time the event/situation occurred, I felt prepared to deal with it	4.51	2.01	−.24	−.11	−.18	.74	.55
16. At the time the event/situation occurred, I was able to think of ways to deal with it	4.93	1.71	−.26	−.03	−.22	.86	.75
17. I felt I had what I needed to deal with the event/situation	4.81	1.78	−.31	−.05	−.19	.75	.57

Note: PARS = Prolonged Activation and Anticipatory Race-Related Stress Scale. Data appearing in boldface represent the highest factor loading for each item.

The four-factor oblique solution accounted for 56% of the total variance. Factor 1 (eigenvalue = 6.57) accounted for 28.57% of the common variance, Factor 2 (eigenvalue = 2.54) accounted for 11.05% of the common variance, Factor 3 (eigenvalue = 2.2) accounted for 9.73% of the common variance, and Factor 4 (eigenvalue = 1.6) accounted for 6.99% of the common variance. Rotation is performed to obtain a simple structure of the data and to aid in the interpretability of the solution (Reise, Waller, & Comrey, 2000). It has been recommended that researchers retain items that produce a strong loading to a single factor with a minimal loading to other factors (Matsunaga, 2010). Although there is no single criterion for retaining items to factors, a meta-analysis of published exploratory factor analysis studies conducted by Hensen and Roberts (2006) identified .30 to .50 as a general range, with .40 being the most commonly used cutoff. Consequently, our criteria for retaining items to factors were as follows: (a) items with factor loadings of .40 or higher were retained, (b) items that cross-loaded (i.e., loaded .40 on more than one factor) were dropped, and (c) items that loaded .15 higher on their primary factor than on their next highest factor loading were retained.

Based on results from the principal axis factoring, six items were eliminated. The remaining 17 items represented the most robust indicators of the prolonged activation and anticipatory stress response to race-related events experienced by African Americans. A content review of the items representing each factor resulted in the assignment of descriptive labels. Factor 1 (five items) was labeled Perseverative Cognition Scale (PCS; e.g., “In the days/weeks after my experience with racism, I thought about it”: 1 = not at all, 2 = once weekly, 3 = 2-3 times a week, 4 = 3 or more times a week, 5 = once a day, 6 = 2-3 times a day, 7 = more than 3 times a day). Items on the PCS represent the frequency, intensity, and controllability. Factor 2 (four items) was labeled Anticipatory Race-Related Stress Scale (ARS; e.g., “I believe there is a good chance I will experience racism in the future”). Factor 3 (four items) was labeled Anticipatory Bodily Alarm Response Scale (ABARS; e.g., “I can feel my hands start to shake whenever I think I am about to experience racism”). Factor 4 (four items) was labeled Secondary Appraisal Scale (SAS; e.g., “At the time the situation occurred, I felt prepared to deal with it”).

Internal consistency and subscale intercorrelations

Cronbach’s alpha coefficients with 95% confidence intervals (CIs) were calculated for scores on the four factors of the PARS. The alpha coefficients were .83 (95% CI = .79-.86) for PCS, .70 (95% CI = .64-.75) for ARS, .83 (95% CI = .80-.86) for ABARS, and .78 (95% CI = 74-.82) for SAS.

Pearson product-moment correlation coefficients were calculated to evaluate the subscale score intercorrelations for the PARS. These findings indicated that PCS was correlated .29 with ARS-PSY, .35 with ABARS, and −.33 with the SAS (ps < .01). ARS correlated .24 with ABARS (p < .01) and .05 with SAS (p > .05). The ABARS correlated −.19 with the SAS (p < .01). On the basis of these low to moderate intercorrelations, the PARS subscale scores can best be described as measuring related, yet distinct constructs associated with prolonged activation and the anticipatory stress response to race-related stressors in African Americans.

Study 2: ConfirmatoryFactor Analysis of the PARS

Method

Participants

The participants for Study 2 consisted of 227 African American college students from a large public university (48), a Historically Black University (142), and adults from the community at large (37). All the data collection sites were located in the Mid-Atlantic region of the United States. Among the 227 participants in the current study, 79 (34.8%) were male and 145 (63.9%) were female. Participants age ranged from 18 to 62 years, with a mean age of 24.98 years (SD = 11.59).

Measures

The PARS is a 17-item self-report measure of prolonged activation and anticipatory race-related stress responses in African Americans. The PARS is composed of four subscales measuring related factors of prolonged activation and the anticipatory stress response to race-related stressors. The first subscale, PCS, consists of five items and measures the degree to which an individual cognitively perseverates about a specific encounter with racism. The Anticipatory Race-Related Stress Scale–Psychological consists of four items and measures the degree to which an individual cognitively anticipates future encounters with race-related stressors. The Anticipatory Race-Related Stress Scale–Physiological consists of four items and assesses the degree to which respondents have a physiological reaction to an anticipated encounter with race-related stressors. Respondents use a 7-point Likert-type scale (1 = Strongly disagree to 7 = Strongly agree) to indicate their response to the race-related stressor. Higher scores indicate a greater expectation that an individual will experience race-related stress in the future. Cronbach’s alphas are reported in the “Results” section of this study. The SAS consists of four items and assesses the degree to which respondents feel that they have the resources to cope with the race-related event. Respondents are asked to indicate the degree to which they agree or disagree with the statements (1 = Strongly disagree and 7 = Strongly agree) about whether they felt they had the resources to cope with the event described. Higher scores indicate more confidence in one’s coping resources.

The Experiences of Discrimination Scale (EOD; Krieger, Smith, Naishadham, Hartman, & Barbeau, 2005) is a 19-item self-report measure of discrimination. The EOD measures experiences with discrimination in terms of frequency, response to unfair treatment, worry, as well as global questions about individual and group experiences with discrimination. Only the Frequency, Worry, and Global scales were used in the current study. The Frequency Scale consists of nine items and asks respondents to indicate how often (0 to 4 or more times) they have experienced discrimination across a variety of domains (school, work, housing, etc.). To score the Frequency Scale, item frequency responses are tallied across domains to create a total frequency score. Higher scores indicated greater frequency of encounters with discrimination. The Worry Scale consists of four items and requires respondents to indicate, on a 3-point Likert-type scale (1 = most of the time and 3 = rarely or never), how much they worried about experiencing racial discrimination during childhood (two items) or in the past year (two items). To score this measure, the items are first reverse-coded and then summed across the two scenarios to derive a total Worry Scale score. The Global Scale consists of two items and asks respondents to indicate, using a 4-point Likert-type scale (1 = never and 4 = often), how much they (one item), or members of their racial group (one item), experienced discrimination. The Global Scale, although separate single-item indicators, are scored by summing both items to create a total global scale score. The initial development study provided some preliminary evidence for the validity and reliability of the measure. For example, the EOD was found to be positively correlated with other measures of racial discrimination (Krieger et al.). The EOD produced Cronbach’s alphas ranging from .74 to .86 and test-retest reliability coefficients were .69 and higher. Cronbach’s alpha coefficients for the current study were .69 for the Frequency Scale and .81 for the Worry Scale. The Global Scale consisted of only two items, so Cronbach’s alpha was not calculated for this measure.

The Multidimensional Ethnic Identity Measure–Revised (MEIM-R; Phinney, 1992) is a 14-item self-report measure of ethnic identity. The MEIM measures two components of ethnic identity; these are as follows: ethnic identity search (five items) and affirmation, belonging, and commitment (seven items). The MEIM was developed primarily for use with college students, but has been used with adults as well (Ponterotto, Gretchen, Utsey, Stracuzzi, & Saya, 2002). The measure asks respondents to rate items on a 5-point Likert-type scale (1 = Strongly disagree and 5 = Strongly agree). Higher scores indicate a stronger ethnic identity. A Cronbach’s alpha of .90 was reported for college students in the initial development study of the MEIM (Phinney, 1992). Cronbach’s alpha coefficient for MEIM in the current study was .67 for the Ethnic Identity Scale and .89 for the Affirmation, Belonging, and Commitment scale.

A demographic data questionnaire was administered to participants, who were asked to indicate their age, gender, ethnicity, and academic status.

Procedures

Procedures identical to those of Study 1 were used for this study. However, several additional questionnaires were added to this study to assess the PARS’ convergent validity. Prior to administration the measures were counterbalanced to reduce the potential for response set.

Results

Internal consistency and subscale intercorrelations

Cronbach’s alpha correlation coefficients with 95% CIs were calculated for the four subscales of the PARS. The coefficient alphas for the four subscales were .77 (95% CI = .72-.82) for the PCS, .70 (95% CI = .64-.76) for the Anticipatory Race-Related Stress Scale–Psychological (ARS), .85 (95% CI = .82-.88) for the ABARS, and .80 (95% CI = .76-.84) for the SAS. With regard to the PARS subscale intercorrelations, PCS correlated .29 with ARS, .41 with ABARS, and −.35 with SAS. The ARS correlated .26 with the ABARS and −.01 with the SAS. The ABARS correlated −.18 with the SAS. The subscale intercorrelations would suggest that the PARS’ subscales are measuring similar, yet distinct, constructs related to the prolonged activation and anticipatory stress response to race-related stressors among African Americans.

Confirmatory factor analysis

A CFA using the maximum-likelihood estimation method was conducted on the covariance matrix of the 17-item PARS. Prior to conducting the CFA, the data were screened for multivariate normality, outliers, and missing data. Findings indicated that the data failed to meet the assumption of multivariate normality and had approximately 4% missing data. No outliers were found to exist in the data set. The issue of a multivariate nonnormal distribution was addressed via the Bollen-Stine bootstrap procedure. Bootstrapping is a resampling procedure that enables researchers to randomly draw multiple subsamples from the parent sample for an empirical examination of the variability of parameter estimates and fit indexes (Byrne, 2001). A Bollen-Stine bootstrap procedure was performed on 1,000 samples using the ML estimator. Missing data were addressed using the listwise deletion procedure. According to Yuan and Bentler (2000), when there are fewer than 20% missing data listwise deletion is just as effective as the (preferred) full information maximum likelihood method.

Following established procedures for conducting SEM with AMOS 16 (Arbuckle, 2006), several fit indices were calculated for the hypothesized four-factor oblique model (see Table 2) and several competing models. Comparisons were made between a null model, a unidimensional model, an alternative nested four-factor orthogonal model, and the hypothesized four-factor oblique model. To determine the degree of fit among the competing models, both absolute and incremental fit indexes were assessed (see Table 2): Chi-square value (χ²), Tucker-Lewis index (TLI), comparative fit index (CFI), incremental fit index (IFI), and root mean square error of approximation (RMSEA), were calculated. A model is said to have a good fit to the data when there is a nonsignificant χ²; the TLI, CFI, and IFI values are greater than .95; and the RMSEA is less than .05 (Bentler, 1990; Hu & Bentler, 1999). Findings from the current CFA indicated that the chi-square values for all of the competing models were significant. This would suggest that none of the models produced an acceptable fit to the data. However, the chi-square test is sensitive to sample size and a statistically significant result does not necessarily reflect a poor fit to the data (Bollen, 1989). The AMOS program provides additional measures of fit that are less sensitive to sample size and considered to be better estimates of a model’s fit to the data (Bentler, 1990). Based on the TLI, CFI, IFI, and RMSEA values, the hypothesized four-factor model produced a good fit to the data and was superior to the competing CFA models.

Table 2.

Confirmatory Factor Analysis Model Fit Summary.

Model	χ²	df	Bollen-Stine Bootstrap	TLI	CFI	IFI	RMSEA (90% CI)
Independence	1599.30**	112	p = .001	.00	.00	.00	.22 [.21, .23]
Unidimensional	1081.08**	132	p = .001	.33	.35	.35	.18 [.17, .18]
Four-factor oblique	168.78**	112	p = .058	.95	.96	.96	.05 [.03, .06]
Nested four-factor orthogonal	311.94**	119	p = .001	.83	.85	.85	.09 [.08, .10]

Note: N = 227. TLI = Tucker Lewis index; CFI = comparative fit index; IFI = incremental fit index; RMSEA = root mean square error of approximation; 90% CI = upper and lower bounds of the RMSEA 90% confidence interval.

p < .01.

Because the data failed to meet the requisite assumption of normality required for conducting a CFA, each model was reevaluated in AMOS 16 using the bootstrap method to obtain a Bollen-Stine corrected chi-square statistic. With nonnormal data, the default maximum likelihood–based p value is likely to be inflated and therefore, not optimal for comparisons against the central chi-square distributions (Hu, Bentler, & Kano, 1992). For nonnormal data, Bollen-Stine bootstrap provides a corrected p value for the chi-square statistic to assess the overall fit of a model. This method produces a much more exact statistic than the maximum likelihood–based p value (Bollen, 1992). Using 1,000 bootstrap samples, a Bollen-Stine corrected chi-square statistic was obtained for each of the PARS models. The findings indicated that the hypothesized four-factor oblique model fit the data worse than expected in 57 of the 1,000 samples, or 57/1,000 = .057, which is the obtained p value of overall model fit. Using a conventional significance level of .05, this model fits the data well. There was no change in p value for the alternative PARS models, as they remained statistically significant, Bollen-Stine corrected p value 0/1,000 = .001. These findings are consistent with the initial CFA results in that the four-factor oblique model produced a superior fit to the data when compared with the competing alternative models.

Convergent validity

It was hypothesized that higher scores on PCS, ARS, and ABARS of the PARS would be positively and statistically significantly correlated with higher scores on the EOD Frequency, Worry, Global scales, and the MEIM-R’s ethnic identity search and affirmation, belonging, and commitment scales. The PARS’ SAS was hypothesized to be inversely correlated with the EOD Frequency, Worry, and Global scales. No relationship was hypothesized to exist between the SAS and the MEIM-R scales. As hypothesized, the PCS, ARS, and ABARS all positively correlated with the EOD Frequency, Worry, and Global measures. The SAS was inversely correlated with the EOD Global and Worry scales, but not the Frequency scale. Only the ARS scale of the PARS correlated in the anticipated direction with the ethnic identity search and the affiliation, belonging, and commitment scales of the MEIM-R. The PCS, ABARS, and SAS were not correlated with the MEIM-R. Means, standard deviations, and Pearson correlation coefficients for the study’s variables are reported in Table 3.

Table 3.

Means, Standard Deviations, and Pearson Product-Moment Correlations for Measures in the Study.

Scale	M	SD	1	2	3	4	5	6	7	8	9
1. PCS	13.41	5.05	—	.29**	.41**	−.35**	.24**	.36**	.30**	.10	.01
2. ARS	13.14	4.51	—	—	.26**	−.01	.27**	.47**	.44**	.27**	.12*
3. ABARS	10.61	2.32	—	—	—	−.18**	.14*	.26**	.15*	.03	.01
4. SAS	5.64	4.31	—	—	—	—	.05	.12*	−.14*	<.01	<.01
5. Frequency	8.98	2.05	—	—	—	—	—	.33**	.27**	.03	.01
6. Worry	6.18	1.03	—	—	—	—	—	—	.37**	.30**	.15*
7. Global	3.55	.66	—	—	—	—	—	—	—	.07	.05
8. EIS	9.29	.2.62	—	—	—	—	—	—	—		−.62**
9. ABC	10.40	3.44	—	—	—	—	—	—	—	—	—

Note: PAS = Perseverative Cognition Scale; ARS = Anticipatory Race-Related Stress Scale; ABARS = Anticipatory Bodily Alarm Response Scale; SAS = Secondary Appraisal Scale; Frequency = Experience of Discrimination Frequency Scale; Worry = Experience of Discrimination Worry Scale; Global = Experience of Discrimination Global Scale; EIS = Ethnic Identity Search scale of the Multigroup Ethnic Identity Measure; ABC = Affirmation, Belonging, and Commitment scale of the Multigroup Ethnic Identity Measure.

p < .05. **p < .01.

Discussion

The aim of the current study was to develop, validate, and evaluate the preliminary psychometric properties of a measure of the process of the prolonged activation of race-related stress. Based on a series of statistical and psychometric procedures, across three separate studies, initial validation evidence was established for scores on the 17-item PARS. Specifically, the PARS produced a distinct underlying four-factor structure; the factors were Perseverative Cognition, Anticipatory Race-Related Stress, Anticipatory Bodily Alarm Response, and Secondary Appraisal. Furthermore, the instrument produced adequate indices of internal consistency, and mediocre-to-fair test-retest reliability estimates. Some initial evidence of the convergent validity of scores on the PARS was also established by its correlation, in the anticipated direction, with related constructs. Overall, the preliminary evidence establishing the validity and reliability of scores on the PARS as a measure of the prolonged activation and anticipatory race-related stress is promising. It is anticipated that the instrument will contribute to conceptual and methodological advancements in the assessment of race-related stress.

Findings related to the associations between scores on the PARS and scores on the EOD and MEIM-R were primarily for validation purposes, but yield important findings beyond their initial intent. For example, that scores on the PARS’ PCS demonstrated moderate-to-good positive correlations with scores on the EOD’s Frequency, Worry, and Global measures not only demonstrate that the measures are assessing similar constructs but also that perseverative cognitions about a race-related stressor increase (or decrease) as the frequency of events increase (or decrease, accordingly). Similarly, as individuals worry more (or less) about future experiences of perceived discrimination perseverative cognitions increase (or decrease, accordingly). Other related findings include moderate correlations between scores on the ARS and ABARS and scores on the EOD’s Frequency, Worry, and Global measures. Intuitively, and beyond the validation goals of the study, these findings suggest that anticipatory race-related stress and anticipatory bodily alarm response to impending race-related stressors increase (or decrease) as the frequency of perceived discrimination or worry about future experiences with perceived discrimination increase (or decrease). The SAS was inversely correlated with the Global measure. Although somewhat modest, this correlation was in the anticipated direction. Taken together, these findings suggest that chronic exposure to perceived discrimination and worry about future experiences with perceived discrimination are linked to the prolonged activation of race-related stressors in the current sample of African Americans. Additional research will be needed to flesh out the full implications of these findings and to determine whether these findings can be replicated with other African American samples from similar and distinct demographic backgrounds.

As it concerns the convergent validity of scores on the PARS, the ARS was positively and moderately positively correlated with the ethnic identity search and affiliation, belonging, and commitment scales of the MEIM-R. There were no other correlations between the PARS subscales and the MEIM-R scales. These findings are particularly intriguing in that the relationship between the anticipatory race-related stress response of the prolonged activation of race-related stress process and the ethnic identity of African Americans in the current sample seem to suggest that the more one identifies with (and is committed to) their ethnic group the more likely they are to anticipate experiencing race-related stressors. Moreover, that the perseverative cognition and anticipatory bodily alarm response scores of the PARS were not correlated with the ethnic identity scores of participants, although unexpected, might suggest that the former are autonomic responses to stress. There was no expectation that secondary appraisal would correlate with ethnic identity. Overall, the findings regarding the relationship between the prolonged activation of race-related stress and ethnic identity seem to indicate that one’s ethnic identity is only related to the degree to which one expects to be treated badly because of their race-group membership.

The rationale for the development of the PARS was to improve on existing measures of race-related stress by developing a more complete conceptual framework for assessing this construct. The theoretical and conceptual frameworks undergirding most existing measures of perceived racism and race-related stress focus mainly on the primary appraisal of a given stressor. More recently, however, research has increasingly pointed to other factors of the stress process as significantly impacting an individual’s capacity to recover from environmental stressors. Specifically, the concept of prolonged activation has received increased attention in the stress and coping literature as researchers continue to explore the link between stress and health (Brosschot et al., 2006). The PARS sought to address the conceptual limitations of existing measures by including items that assess factors associated with the prolonged activation of stress, namely Perseverative Cognition, Secondary Appraisal, Anticipatory Race-Related Stress, and the Bodily Alarm Response to impending race-related stressors. Second, methodological limitations of existing measures of racism- and race-related stress include an overreliance on the checklist approach to identifying relevant stressors. The checklist approach, where respondents select items from a preestablished list of race-related stressors, is problematic because some race-related events may not appear on the list. In such instances, significant race-related stressors will be underreported. Additionally, respondents might endorse an item that only approximates their actual experience, since the exact event does not appear on the list, resulting in the overreporting of race-related stressors. The current measure addresses this issue by allowing respondents to indicate and describe a specific race-related event they experienced, and refer to this experience when responding to the PARS items.

The PARS is not intended as a measure of the cumulative effects of multiple race-related events across time, as the majority of existing measures do, but is intended to assess the process of the prolonged activation stress response to race-related stressors in the context of a specific race-related event. Applying the conceptual framework delineated earlier in this article, the process of the prolonged activation of race-related stress includes the following: perseverative cognitions about a race-related event, an evaluation of one’s resources to be summoned in managing the race-related stressor, anticipation of future race-related events, and the potential for bodily alarm response related to impending race-related stressors. There are several advantages to this approach. First, by focusing on a specific event that the respondent identifies the PARS is a more exact assessment of race-related stress. When race-related stressors are combined for a cumulative assessment more significant events are collapsed with less significant events for an average of all events. Second, by focusing on the process of the prolonged activation of race-related stress as opposed to the average of the magnitude of multiple perceived events, the PARS is well suited to assess the complexity of the link between chronic exposure to race-related stress and health outcomes. Last, better understanding of the process of the prolonged activation of race-related stress provides more information for the development of effective interventions at specific sites of the conceptual framework. For example, if it is determined that secondary appraisal is the critical point in the conceptual link where the appropriate intervention can reduce the prolonged activation of race-related stress, and thereby reduce any associated negative health outcomes, then clinicians can focus their interventions accordingly. Additional research will be necessary to more completely understand the clinical implications of the PARS.

Another important component of the PARS is its inclusion of network events in assessing the prolonged activation of race-related stressors. Network events are stressors experienced by a family member or someone close to the respondent (Kessler & McLeod, 1984). A number of stress and coping researchers have examined this phenomenon in the context of stress and coping. For example, Utsey and Ponterotto (1996) included the assessment of network events in their initial version of the Index of Race-Related Stress. Harrell (2000) later described the phenomenon of network events in the context of vicarious racism experiences, which she characterized as those “Experiences of prejudice and discrimination that happen to members of one’s family and close friends . . .” (p. 45). A study by Alvarez, Juang, and Liang (2006) examined the degree to which vicarious racism impacted the perceptions of Asian Americans about their experiences with racism. Although racism and race-related stress scholars have long been concerned with the impact of network events, the PARS provides a method for assessing the degree to which they result in comparable harm.

Limitations and Recommendations for Future Research

Despite some promising preliminary evidence of the PARS’ viability as a valid and reliable measure of the prolonged activation and anticipatory race-related stress response in African American adults, there are several limitations to be noted. First, both studies had samples that were overwhelmingly female. Second, the studies relied exclusively on self-report data from participants. Self-report measures are prone to increased measurement error due to retrospective recall bias, problems with accuracy of reporting, and a potential for respondents to be confused by the directions. Last, given a nonrandom sample of African American college students from one urban city in the Mid-Atlantic region of the United States, the findings likely do not generalize beyond the parameters of the current sample.

The primary purpose of the current study was to develop and establish some initial psychometric data for the PARS. Additional studies will be required to further establish the validity and reliability of the measure with nonstudent and clinical populations. It will be important for future studies to examine how well the PARS four-factor model fits the data using community and clinical samples of African American adults. Furthermore, the underrepresentation of males in all three samples of the current studies raises the possibility that future studies may require oversampling for male participants. Additionally, researchers will need to examine other indices of reliability, such as parallel forms and split half reliability. Likewise, additional forms of validity, such as divergent validity, criterion validity, and discriminant validity, should be assessed. All these efforts would contribute to a thorough psychometric evaluation of the PARS and would be a welcomed contribution by the original developers of the instrument as well as researchers in the area of race-related stress.

Footnotes

Appendix

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Alvarez

A. N.

Juang

Liang

C. T. H.

(2006). Asian Americans and racism: When bad things happen to “model minorities.” Cultural Diversity & Ethnic Minority Psychology, 34, 684-718.

Arbuckle

J. L.

(2006). AMOS 7.0 user’s guide. Spring House, PA: AMOS Development Corporation.

Bentler

P. M.

(1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238-246.

Bollen

K. A.

(1989). Structural equations with latent variables. Oxford, England: Wiley.

Bollen

K. A.

(1992). Bootstrapping goodness-of-fit measures in structural equation models. Sociological Methods & Research, 21, 205-229.

Borders

Liang

C. T. H.

(2011). Rumination partially mediates the associations between perceived ethnic discrimination, emotional distress, and aggression. Cultural Diversity & Ethnic Minority Psychology, 17, 125-133.

Bowen-Reid

T. L.

Harrell

J. P.

(2002). Racist experiences and health outcomes: An examination of spirituality as a buffer. Journal of Black Psychology, 28, 18-36.

Brondolo

Gallo

L. C.

Myers

H. F.

(2009). Race, racism and health: Disparities, mechanism, and interventions. Journal of Behavioral Medicine, 32, 1-8.

Brosschot

J. F.

Gerin

Thayer

J. F.

(2006). The perseverative cognition hypothesis: A review of worry, prolonged stress-related physiological activation and health. Journal of Psychosomatic Research, 60, 113-124.

10.

Brosschot

J. F.

Pieper

Thayer

J. F.

(2005). Expanding stress theory: Prolonged activation and perseverative cognition. Psychoneuroendocrinology, 30, 1043-1049.

11.

Brosschot

J. F.

Van Dijk

Thayer

J. F.

(2002). Prolonged autonomic activation, perseverative negative cognition, and daily stressors. International Congress Series, 1241, 329-336.

12.

Brown

T. A.

(2006). Confirmatory factor analysis for applied research. New York, NY: Guilford Press.

13.

Byrne

B. M.

(2001). Structural equation modeling with AMOS: Basic concepts, applications, and programming. Mahwah, NJ: Lawrence Erlbaum.

14.

Cattell

R. B.

(1966). The Scree test for the number of factors. Multivariate Behavioral Research, 1, 245-276.

15.

Clark

Anderson

N. B.

Clark

V. R.

Williams

D. R.

(1999). Racism as a stressor for African Americans: A biopsychosocial model. American Psychologist, 54, 805-816.

16.

Comas-Diaz

Jacobsen

F. M.

(2001). Ethnocultural allodynia. Journal of Psychotherapy Practice and Research, 10, 246-252.

17.

Davis

S. K.

Liu

Gibbons

G. H.

(2003). Disparties in trends of hospitalization for potentially preventable chronic conditions among African Americans during the 1990s: Implications and benchmarks. American Journal of Public Health, 93, 447-455.

18.

Franks

Muennig

Lubetkn

Jia

(2006). The burden of disease associated with being African American in the United States and the contribution of socio-economic status. Social Science & Medicine, 62, 2469-2478.

19.

Gee

G. C.

(2002). A multilevel analysis of the relationship between institutional and individual racial discrimination and health status. American Journal of Public Health, 92, 615-623.

20.

Geronimus

A. T.

Hicken

Keene

Bound

(2006). “Weathering” and age patterns of allostatic load scores among Blacks and Whites in the United States. American Journal of Public Health, 96, 826-833.

21.

Harrell

S. P.

(2000). A multidimensional conceptualization of racism-related stress: Implications for the well-being of people of color. American Journal of Orthopsychiatry, 70, 42-57.

22.

Hatzenbuehler

M. L.

Nolen-Hoeksema

Dovidio

(2009). How does stigma “get under the skin”? The mediating role of emotion regulation. Psychological Science, 20, 1282-1289.

23.

Hensen

R. K.

Roberts

J. K.

(2006). Use of exploratory factor analysis used in published research: Common errors and some comment on improved practice. Educational and Psychological Measurement, 66, 393-416.

24.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55.

25.

Bentler

P. M.

Kano

(1992). Can statistics in covariance analysis be trusted? Psychological Bulletin, 112, 351-362.

26.

Karlsen

Nazroo

J. Y.

(2002). Relation between racial discrimination, social class, and health among ethnic minority groups. American Journal of Public Health, 92, 624-631.

27.

Kessler

R. C.

McLeod

J. D.

(1984). Sex differences in vulnerability to undesirable life events. American Sociological Review, 49, 620-631.

28.

Kressin

N. R.

Raymond

K. L.

Manze

(2008). Perceptions of race/ethnicity-based discrimination: A review of measures and evaluation of their usefulness for the health care setting. Journal of Health Care for the Poor and Underserved, 19, 697-730.

29.

Krieger

(1999). Embodying inequality: A review of concepts, measures, and methods for studying health consequences of discrimination. International Journal of Health Services, 29, 295-352.

30.

Krieger

Smith

Naishadham

Hartman

Barbeau

E. M.

(2005). Experiences of discrimination: Validity and reliability of a self-report measure for population health research on racism and health. Social Science & Medicine, 61, 1576-1596.

31.

Kwate

N. A.

Valdimarsdottir

H. B.

Guevarra

J. S.

Bovbjerg

D. H.

(2003). Experiences of racist events are associated with negative health consequences for African American women. Journal of the National Medical Association, 95, 450-460.

32.

Landrine

Klonoff

E. A.

Corral

Fernandez

Roesch

(2006). Conceptualizing and measuring ethnic discrimination in health research. Journal of Behavioral Medicine, 29, 79-94.

33.

Lazarus

R. S.

Folkman

(1984). Stress, appraisal, and coping. New York, NY: Spring Publishers.

34.

Matsunaga

(2010). How to factor-analyze your data right: Do’s and don’ts, and how-to’s. International Journal of Psychological Research, 3, 97-110.

35.

McCord

Freeman

H. P.

(1990). Excess mortality in Harlem. New England Journal of Medicine, 322, 173-177.

36.

Merritt

M. M.

Bennett

G. G.

Williams

R. B.

Edwards

C. L.

Sollers

J. J.

(2006). Perceived racism and cardiovascular reactivity and recovery to personally relevant stress. Health Psychology, 25, 364-369.

37.

Meyer

I. H.

(2003). Prejudice as stress: Conceptual and measurement problems. American Journal of Public Health, 93, 262-265.

38.

Monat

Averill

J. R.

Lazarus

R. S.

(1972). Anticipatory stress and coping reactions under various conditions of uncertainty. Journal of Personality and Social Psychology, 24, 237-253.

39.

Outlaw

F. H.

(1993). Stress and coping: The influence of racism on the cognitive appraisal processing of African-Americans. Issues in Mental Health Nursing, 14, 399-409.

40.

Paradies

(2006). A systematic review of empirical research on self-reported racism and health. International Journal of Epidemiology, 35, 888-901.

41.

Phinney

(1992). The Multigroup Ethnic Identity Measure: A new scale for use with adolescents and young adults from diverse groups. Journal of Adolescent Research, 7, 156-176.

42.

Pieterse

A. L.

Carter

R. T.

(2010). An exploratory investigation of the relationship between racism, racial identity, perceptions of health, and health locus of control among Black Americans. Journal of Health Care for the Poor and Underserved, 21, 334-348.

43.

Ponterotto

J. G.

Gretchen

Utsey

S. O.

Stracuzzi

Saya

(2003). The Multigroup Ethnic Identity Measure (MEIM): Psychometric review and further validity testing. Educational and Psychological Measurement, 63, 502-515.

44.

Reise

S. P.

Waller

N. G.

Comrey

A. L.

(2000). Factor analysis and scale revision. Psychological Assessment, 12, 287-297.

45.

Sharpley

C. F.

(2002). Heart rate reactivity and variability as psychophysiological links between stress, anxiety, depression and cardiovascular disease: Implications for health psychology interventions. Australian Psychologist, 37, 56-62.

46.

Shavers

V. L.

Shavers

B. S.

(2006). Racism and health inequity among Americans. Journal of the National Medical Association, 98, 386-396.

47.

Smith

T. W.

(2002). Measuring racial and ethnic discrimination (GSS Methodological Report No 96). Chicago, IL: National Opinion Research Center, University of Chicago.

48.

Utsey

S. O.

(1998). Assessing the stressful effects of racism: A review of instrumentation. Journal of Black Psychology, 24, 269-288.

49.

Utsey

S. O.

Giesbrecht

Hook

Stanard

P. M.

(2008). Cultural, sociofamilial, and psychological resources that inhibit psychological distress in African Americans exposed to stressful life events and race-related events. Journal of Counseling Psychology, 55, 49-62.

50.

Utsey

S. O.

Hook

(2007). Heart rate variability as a physiological moderator of the relationship between race-related stress and psychological distress in African Americans. Cultural Diversity & Ethnic Minority Psychology, 13, 250-253.

51.

Utsey

S. O.

Ponterotto

J. G.

(1996). Development and validation of the Index of Race-Related Stress (IRRS). Journal of Counseling Psychology, 43, 490-501.

52.

Williams

D. R.

(1999). Race, socioeconomic status, and health: The added effects of racism and discrimination. Annals of the New York Academy of Sciences, 896, 173-188.

53.

Wyatt

S. B.

Williams

D. R.

Calvin

Henderson

F. C.

Walker

E. R.

Winters

(2003). Racism and cardiovascular disease in African Americans. American Journal of the Medical Sciences, 325, 315-331.

54.

Yuan

K.-H.

Bentler

P. M.

(2000). Three likelihood-based methods for mean and covariance structure analysis with nonnormal missing data. Sociological Methodology, 30, 165-200.