Prejudice as Stress: Conceptual and Measurement Problems

Abstract

In the field of social sciences, there has been a renewed interest in studying prejudice and discrimination as stressors and assessing their impact on various health outcomes. This raises a need for theoretically based and psychometrically sound measures of prejudice.

As researchers approach this task, there are several conceptual issues that need to be addressed. The author describes 3 such issues related to (1) individual versus structural measures of the impact of prejudice, (2) objective versus subjective assessments of stress, and (3) measures of major events versus everyday discrimination.

How researchers approach the problem of measurement depends on the specific study aims, but they must consider these conceptual issues and understand the advantages and limitations of various approaches to the study of prejudice as stress.

MOST OF THE SOCIAL psychological work in the area of prejudice has been concerned with the perspective of the perpetrator (e.g., the authoritarian personality described by Adorno et al.).1 Allport’s description, nearly 50 years ago, of the effect of prejudice on members of minority groups was one notable exception.2 Recently, sociology and social psychology have seen a renewed interest in the study of prejudice from “the target’s perspective.”3 The stress model is often featured in these new works.4 In 1999, Clark et al.5 published an article titled “Racism as a Stressor for African Americans: A Biopsychosocial Model”; this article, which is already having an influence on the study of prejudice, explicitly placed racism within the stress conceptual framework. Clearly, conceptualizing prejudice and discrimination as stress fits well with the sociological notion of stress, which describes this concept as embedded in social structures.6 Thus, in the case of social groups, the stress model has been applied to stress related to disadvantaged class, sex, sexual orientation, and other positions in society.7–10

Regardless of the renewed efforts to study prejudice directly as stress, prejudice enters any stressful life event measure indirectly. Race and ethnicity are examples of stratifying social statuses that can be linked to potentially stressful experiences.11,12 For instance, as a result of the excess impact of discrimination, an African American individual is more likely to have experienced job loss than a White individual. Because job loss is a routinely studied stressor, life event measures may capture this differential exposure. However, there seems to be a consensus that it is insufficient to understand prejudice only as a mediator that leads to excesses in certain life events and that researchers should assess prejudice and discrimination as unique events.5

The renewed interest in prejudice as stress raises the need for theoretically based and psychometrically sound measures of prejudice. As we approach this task, there are several conceptual issues that need to be addressed. I describe 3 of these issues without attempting to resolve them, recognizing that their resolution depends, in part, on the purpose of particular investigations. I present these conceptual issues as polar problems related to (1) individual versus structural measures of the impact of prejudice, (2) objective versus subjective assessments of stress due to prejudice, and (3) major events versus daily hassles as measures of prejudice.

INDIVIDUAL VERSUS STRUCTURAL MEASURES

The first challenge is to sort out the extent to which prejudice and discrimination properly fit the stress model. Although concerned with social sources of stress, stress research has focused on assessing stress at the individual level of analysis. However, stress as an individual-level variable is limited in terms of capturing the impact of prejudice and discrimination. For example, Adams described institutional (as opposed to individual) racism as involving stressors that thwart prosperity, esteem and honor, and power and influence.13

Such institutional barriers—or, as Link and Phelan labeled them, structural discrimination barriers—are often impossible to detect at the individual level.10 This is especially true in the case of illegal forms of discrimination, such as employment practices discriminating against African Americans, in which the perpetrator is sure to have made efforts to disguise the discriminatory nature of the act. Thus, an African American respondent may honestly report to a researcher that he or she was not denied a promotion because of discrimination, not realizing that the institution for which he or she worked in fact engaged in a purposeful but hidden policy of excluding Blacks from promotions.

Structural barriers can also become invisible in individual-level research even when they are not concealed. If prejudice and discrimination are legal and widely practiced, they are likely to affect many or all members of a minority group; thus, there would be little or no variability to study. For example, gay men and lesbians are uniformly and legally excluded from marriage, but research that focuses only on these individuals would fail to reflect this practice as an instance of discrimination. Such a lack of variance is especially problematic in that many studies of prejudice and discrimination assess individual measures of stress and do so in samples that include only members of the minority group of interest (e.g., studies of hypertension among African Americans), thus detecting within-group variability in exposure to prejudice and discrimination but failing to detect the potentially stronger manifestation of structural prejudice.14

The impact of institutional stressors may best be documented via assessments of differences in population parameters (including economics and health) at the group rather than the individual level.13 However, other measurement approaches can be developed to capture structural variability and integrate such variability with individual-level observations tested by the stress model. For example, Darity, an economist, developed measures for comparing personal financial standing based on self-reports with group-level financial achievement based on population observations. This allowed him to estimate prejudice-related discrepancies between expected and actual achievements among Black and White respondents (W. A. Darity Jr., unpublished data, 2002).

OBJECTIVE VERSUS SUBJECTIVE ASSESSMENTS

Even if we were interested only in individual-level measures of prejudice as stress and left the structural level of racism to epidemiologists, economists, and other scientists, we would still face serious conceptual and measurement challenges. To understand this problem, it is important to note a distinction in the conceptualization of stress that is significant in discussions of stress due to prejudice. I refer to 2 general approaches that underlie stress discourse, one viewing stress as an objective phenomenon and the other viewing it as a subjective phenomenon. The objective view defines stressful life events as real and observable phenomena that are experienced as stressors because of the adaptational demands they impose on most individuals under similar circumstances.15 The subjective view defines stress as an experience that is contingent on the relationship between the individual and his or her environment. This relationship depends on properties of the external event but also, significantly, on appraisal processes applied by the individual.16,17

The distinction just described has important conceptual and methodological implications in the context of stress due to prejudice. Most, if not all, measures of discrimination events that have been developed to date rely on subjective perceptions. A typical item asks the respondent whether or not he or she has been discriminated against (e.g., in relation to employment); sometimes respondents are asked to recall such events over their lifetime, and sometimes they are asked to limit themselves to more recent periods. In certain instances, respondents are also asked to identify the reason for this discrimination (e.g., “Was this because of your race/ethnicity, gender, religion, social status, sexual orientation, or something else?”).18 However, individual reports of discrimination depend on perception, which produces discrepancies in findings. For example, as described earlier, discrimination can be hidden and thus undetected by its victims.

More relevant to the discussion here, however, is that even when individuals have opportunities to observe discrimination events, many factors affect the perception and reporting of these events. This is important because of the potential for confounding of the measurement of the stressor (as the independent variable) and the measurement of outcomes (especially mental health outcomes). Many individual psychological and demographic characteristics may affect perceptions and reporting of prejudice as a form of stress.

For example, Contrada and colleagues19 suggested that although minority group members are motivated by self-protection to detect discrimination, they are also motivated to ignore evidence of discrimination through a wish to avoid false alarms that can disrupt social relations and undermine life satisfaction. Similarly, some evidence suggests that, in ambiguous situations, people tend to maximize perceptions of personal control and minimize recognition of discrimination. Such observations indicate that healthier individuals may use strategies that lead them to underestimate prejudice and discrimination events. This may lead to bias that would attenuate the detected impact of perceived discrimination on health.

There are many other potential biases in perceptions and reports of prejudice and discrimination events that have been the focus of exciting new cognitive studies.19 Some interesting findings are as follows: (1) people who actively cope with prejudice are more likely to notice, recall, and report prejudice events; (2) minority group members have strong motivations to ignore prejudice-related events in some instances but to be hypervigilant of them in other instances; and (3) inclinations to report prejudice events may vary depending on concordance between respondents and interviewers in terms of minority status. Knowledge is incomplete regarding the correlates of variation in these biases, but they are certain to affect associations between stress and health outcomes. These motivational factors can lead to inaccurate reports of events of discrimination and prejudice and present serious challenges to researchers who are interested in an objective account of what actually occurred.

It is important to note, however, that an interest in the objective phenomenon is not uniformly accepted by researchers in the area of prejudice. Many researchers focusing on racism as stress see strong value in recording the minority person’s subjective perspective, that is, his or her perception of prejudice and discrimination. They view this as an important political choice, noting the presence of bias in previous studies of minority populations. They suggest that research focus on individual perceptions of prejudice and discrimination so as to empower the respondent’s perspective. Indeed, in part because of these reasons, Clark and colleagues5 called for studies of racism to employ Lazarus and Folkman’s model of stress and focus on perceived racism.16

However, the subjective stress model should not be adopted without careful deliberation. From a methodological perspective, relying only on subjective perceptions of stress is problematic because serious confounding can occur between an individual’s health and his or her perception of stressors.20 This is particularly the case in research on the association between life events and mental disorders. Such studies, which attempt to estimate prejudice as it relates to disease outcomes, require a methodology that conceptualizes stress as an objective phenomenon independent of an individual’s own views and feelings. From an ethical perspective, relying only on subjective perceptions may have the benefit of empowering the minority respondent’s voice, but it also may imply that prejudice and racism are merely problems related to perception, thus indirectly and unintentionally undermining the notion that racism and other forms of prejudice are social rather than individual stressors.

Still, an objective approach seems difficult to reconcile with the approach that views perceived racism as more important, and it may be at odds with the ideology expressed in the literature on perceived racism. For example, if stress research serves an expressive purpose in regard to minority concerns, the practice of rating events objectively and independently of individual perceptions raises ethical concerns, because it may involve devaluation of the perspectives of minority research respondents.

I recently confronted such an ethical dilemma in studying prejudice as stress at the intersection of race/ethnicity, gender, and sexual orientation. In this project, I used subjective (perceived) discrimination scales, but I also used an objective probed narrative method to study stressful events related to prejudice. This method, developed by Dohrenwend and colleagues, involves detailed probing of each event reported by a respondent. After the narrative has been recorded by the interviewer, it is rated by independent raters, according to specified criteria, on stress dimensions such as event valence, centrality, and magnitude.15

A case demonstrating the subjective versus objective dilemma involved a young Latino respondent who was an illegal immigrant. In a life event list, he reported moving to a new apartment and not being able to obtain a telephone as a stressful event. Because he made his living by cleaning homes, he depended on a telephone to receive referrals and calls from customers. Not being able to obtain a telephone could have had a significant impact on his income, making the event more detrimental than it might initially appear. In responding to a subjective perceived discrimination questionnaire, the respondent reported his experience as a discrimination event related to his ethnic/racial identity and immigrant status.

If our research team had stopped there, that is all we would have known. But in probing the event further, we discovered that the respondent was denied a telephone because he was unable to produce the documents and cash deposit routinely required by the telephone company for people with no established credit record in the United States. In discussing his narrative after completion of the rating procedure, the research team rated this event as not caused by discrimination, because we determined that the respondent was treated in a manner identical to that of any other person similarly lacking documents and money, and thus he was not singled out because of his race/ethnicity.

We were confident that a review conducted by a court or human rights commission would yield a similar result. Yet, we felt uneasy; it can certainly be argued that the respondent, despite being treated in an equitable manner, confronted a serious discriminatory social barrier related to his immigration status, low income, and, perhaps, ethnicity. In rejecting the respondent’s perspective and recording the event as not involving prejudice, were we complicit with an oppressive social structure whose effects we purport to study? On the other hand, is our research better served by recording the respondent’s subjective account? This could lead to biasing of the association between stress and health outcomes, possibly diminishing or masking a true effect.20

MAJOR LIFE EVENTS VERSUS DAILY HASSLES

The third challenge to measurement of prejudice as stress is related to the significance, or magnitude, of minor discrimination events. In traditional life events studies, researchers distinguish between major events and daily hassles. Daily hassles are ubiquitous; most people perceive hassles as an unavoidable part of life and are expected to recover relatively quickly from such experiences. Associations between daily hassles and mental health outcomes are likely to be overestimated, because mood states probably affect perceptions and reports of daily hassles as well as outcome measures. This raises questions about the utility of hassles scales as independent variables in the study of stress and mental health, and many researchers do not use them.20

Yet, minor discrimination events are pervasive and have an impact on many aspects of daily life. Williams and colleagues referred to “everyday discrimination” in describing occurrences such as African American men being followed in stores for suspicion of shoplifting or not being able to hail a cab.12 If we considered such occurrences as minor hassles, we would err: as a result of their meaning in a social context, they are more significant than traditionally defined daily hassles that are not related to prejudice. A seemingly minor everyday discrimination occurrence, such as not being able to hail a cab, can evoke among minority individuals painful memories related to personal and communal histories of prejudice.

In Race Matters, Cornel West described failing to get a cab in New York because taxi drivers would not stop for him; he stated that “[u]gly racial memories of the past flashed through [his] mind” as he recalled encounters with racism in his own life and the lives of others. He acknowledged that his experience paled against more serious acts of racism but nevertheless recalled that “the memories cut like a merciless knife at my soul as I waited on that godforsaken corner.”21^21(ppx–xi)

It is possible that such minor discriminatory events have greater effects on health outcomes than their seeming magnitude would suggest; for example, their effects on cardiovascular health, via activation of the sympathetic nervous system, have been examined.22 If we are concerned only with major events, we may miss an important aspect of the experience of prejudice and discrimination, and thus we may underestimate the burden of stress resulting from prejudice.

CONCLUSIONS

I have identified 3 issues that raise questions regarding measurement of prejudice and discrimination as stress, but I offer no resolution. Indeed, an attempt to find a resolution may be unwise, because the solutions to the methodological problems raised here depend on the specific research questions pursued and on investigators’ conceptual thinking about prejudice as stress. For example, a researcher who attempts to chronicle minority people’s experience in society may be interested in chronicling their world as they experience it, and therefore he or she may prefer a subjective measurement of prejudice stress. A researcher who aims to understand the impact of prejudice as a cause of disease and quantify its role in differential health outcomes may find it more important to measure prejudice as an objective stressor, reducing bias related to appraisal and recall as well as threats resulting from the confounding of individual characteristics and disease outcomes.

Regardless of their aims, researchers must understand various approaches to measuring prejudice as stress, and they must consider the advantages and limitations of their preferred approach. Integration of various approaches, collaboration among researchers from various disciplines who bring different perspectives to understanding prejudice as a risk factor for disease, and use of contextual and multiple levels of analysis may be the most promising routes in the study of prejudice and health.23,24