Editorial Board:
Editor:
George H. Conklin,
 North Carolina
 Central University

Board:
Bob Davis,
 North Carolina
 Agricultural and
 Technical State
 University

Richard Dixon,
 UNC-Wilmington

Ken Land,
 Duke University

Miles Simpson,
 North Carolina
 Central University

Ron Wimberley,
 N.C. State University

Robert Wortham,
 North Carolina
 Central University




Editorial Assistants

Rob Tolliver,
 North Carolina
 Central University

Shannon O'Connor,
 North Carolina
 Central University

John W.M. Russell,
 Technical
 Consultant

® 

Volume 5, Number 1

Spring 2007
 
 

Exploring Health Disparities Affected by Stressful Life Events.
 Are Older African Americans At Greater Risk?*

by

Jeffrey D. Brooks 
Fayetteville State University

Eva Kahana and Andre Nauta
Case Western Reserve University

Boaz Kahana 
Cleveland State University

    One of the early and possibly most widely known uses of the idea that minority elders are somehow more vulnerable than others was Robert Butler's use the phrase "multiple jeopardy (Butler, 1975, p. 30)."  Butler's use of multiple jeopardy includes attributing higher poverty rates and lower life expectancy in minorities to the influences of racism and ageism.  Now researchers apply Butler's idea to differences in mental and physical health.  It should also be noted that more recent studies refer to this as double jeopardy (Ferraro & Farmer, 1996), and debates are ongoing about validity of the double jeopardy hypothesis.  A goal of this study is to shed some light on this debate.

    Evidence of health disparities experienced by black and other minorities in early in life is well documented (Weitz, 2004).  Whereas infant mortality statistics are well established and widely published, information about the elderly is less common, and published research about minorities in later life is relatively rare.  Just a decade ago a special report by Gerontological Society of America (GSA) described minority elders as remaining statistically invisible (Gerontological Society of America, 1994).  Unfortunately, concern about the relative invisibility of minority elders remains valid today.  An additional incentive for this project is the ongoing efforts by the GSA and the Council on Social Work Education to enhance training of workers so they have gerontological skills.  Hence, with these points in mind and the support a National Institute of Health sponsored grant, we report results from our interviewers with retirees who belonged to an HMO.  The sample provides interesting insights for those interested in examining predictions of health disparities between older African Americans and Whites.  We are also fortunate that a rich heritage of stress research and theory is available to give direction to this investigation. 

    Our approach to exploring the double jeopardy hypothesis about health disparities was to investigate exposure to stressful life events and mental health status.  We should point out that our sample is not representative, but rather has characteristics of a unique quasiexperimental method that is well-suited for exploring race differences in exposure to stress and any mental health outcomes experienced as a result of the stressful life events.  Respondents are fairly homogeneous in that they have similar occupations and equal access to healthcare, and so we felt these respondents were suitable for examining the double jeopardy hypothesis.  More specifically, we utilize a stress paradigm because it provides existing methods, theoretical strength and offers practical implications for understand what factors influence mental health in later life.

    In this study we used general predictions derived from the stress paradigm to examine the double jeopardy hypothesis using psychological well being as the dependent variable.  In addition, we attempt to employ specific predictions about the impact of various types of stressful life events (positive, negative, egocentric, nonegocentric) on the measures of mental health and morale.  As such, an additional goal is to examine the fruitfulness of disaggregating the stressful events when examining their influences on well-being.  However, our primary goal is to offer additional results to the debate suggesting elderly African Americans do or do not suffer from the risks or vulnerability implied by the notion of double jeopardy. 

    We will be examining two aspects of double jeopardy.  First, we will look at increased risk for black elderly by considering the number of life events experienced, since increased risk implies that black elderly will experience more stressful events.  Second, we will look at hypothesized increased vulnerability of black elderly.  If racism increases the number of stressors, it is reasonable to assume that the cumulative impact will be greater for black elderly, and therefore the negative impact of events late in life will be larger for blacks.

 Disaggregating Dimensions of Life Events

    A strong precedent exists for examining differences of desirability or the positive nature of events and the undesirability or negative nature of events (Kessler & Mcloud, 1983, Tausig, 1982; Thoits, 1983,).  In addition, the recent work by Aldwin (1990) offers a promising second distinction into egocentric/nonegocentric types of events.  As both typologies are used in this study, a discussion of each follows.

Positive versus Negative Life Events

    Initial interest in the positive and negative nature of events concerned questions about whether the relationship between stressful events and psychological disturbances was due to change per se, or undesirable change (Thoits, 1983).  Studies using a judges ratings or the person's assessment of the desirability or undesirability of an event (Vinokur & Selzer, 1975) found that undesirable events have a significant relationship to a variety of measures of distress, while desirable events do not.  A review found a consistency of findings suggesting that undesirability is the crucial dimension of events implicated in the etiology of psychological disturbance and not the total change associated with events (Thoits, 1983, p. 59).  One additional point is that desirable events have been reported as being associated with reduced symptoms (Grant, Sweetwood, Yeager & Gerst, 1981).  While it would appear undesirable events best predict psychological outcomes, Thoits also points out that there is some evidence that total change (which includes positive and negative events) best predicts physical health outcomes.  It should be kept in mind that the studies of stress reviewed by Thoits did not include older subjects and were not concerned with age differences.  Chiriboga (1982) did study different age groups and employed a longitudinal design in an effort to study consistency in functioning which suggests some interesting points to consider about age and reports of positive and negative events.  The seniors (retirement age at the beginning of the study) showed the greatest increase in reports of life events of the four age groups.  He also reports that those in the retirement stage reported steady increases in positive events concerning the family. 

Egocentric and Nonegocentric Life Events

    The study of egocentric and nonegocentric life events was initially proposed to be useful in demonstrating one aspect of positive changes in later life development (Aldwin, 1990). Nauta and his co-authors (Nauta, Brooks, Johnson, Kahana & Kahana, 1994) tested the egocentric/non egocentric classification of stressful life events and their method of classification will be used here.  Aldwin was interested in disaggregating events because she posits increased concern for others reflects Erikson's (1959) notion of generativity. It is important to point out that while Aldwin accepts the notion that changes in sensitivity to stress occur with advancing age, but she does not theorize that this change should be seen as a deficit or deterioration.  It is not our intention to evaluate Aldwin’s proposition about generativity, rather we are interested in looking for race differences and any impact of the disaggregated groups of stressful events have on mental health. 

    Two studies reporting on the validity of stressful life event scales suggest that it is appropriate to consider egocentric and nonegocentric events as separate dimensions of stressful events.  Miller, Bentz, Aponte and Brogan (1974) report factor analysis for the social readjustment rating scale (Holmes & Rahe, 1967) as producing four factors: two for lifestyle changes such as work and recreation changes (similar to Aldwin's egocentric events), and two for interpersonal changes such as relationship changes and relationship terminations (similar to Aldwin's nonegocentric events).  The measure of stressful events used in this study was initially concerned with validating the change values for older people.  Results from using the Geriatric Scale of Recent Life Events (GSRLE) support both the notion of sensitivity to others and the greater reactivity to nonegocentric events (Kiyak, Liang, & Kahana, 1976).  Using the GSRLE revealed that sensitivity to others differed by age, where the elderly subjects assigned higher change scores for the interpersonal events of marriage, reconciliation, arguments with spouse, divorce, death of spouse and other family problems. 

    Given the literature on the double jeopardy hypothesis, we propose the following hypotheses:

  1)  Black elderly should experience a greater number of life events, particularly those that would be labeled negative and/or nonegocentric, given the combined influences of racism and ageism. 

  2)  Second, if the concept of "double jeopardy" holds true, one would expect that Black elderly will be more vulnerable to the events occurring to themselves and their family and friends than their white counterparts, given the cumulative effects of a lifetime of stress related to racism.  Thus, the impact of these events on psychological well-being should be greater for black elderly than for white elderly. 

    Data were collected from 397 community residing elderly persons who were enrolled in a Detroit-area health maintenance organization (HMO) as part of a study concerned with life stress and illness.  All respondents were aged 55 or older, living in their own homes or apartments, and retired from their lifetime occupations (generally from the auto industry) at the time of the interview.  Face-to-face interviews were conducted with respondents in their homes.  The sample had relatively equal proportions of males and females, as well as blacks and whites (no other racial groups were represented in the sample).

Life Events: Egocentric/nonegocentric and positive/negative events were measured using the GSRLE (Kiyak et al., 1976).  This scale contains a checklist of 33 events (see Table 1) adapted for older respondents from items utilized by Holmes and Rahe (1967).  Respondents were asked to indicate which events they had experienced in the year prior to the interview, and describe the type/amount of change they experienced as a consequence of the event.  However, change scores were not used for this study because of the potential confounding of change with the dependent measures, such that those with lower morale/poorer mental health might give higher change scores for the events they experienced.

    Assignment of life events to positive and negative categories was based on respondents' evaluations of whether the event changed their lives in a positive or negative way.  Table 1 shows the ratings based on the majority response of those experiencing each event.  In those instances where the respondents were relatively equally divided between the two choices, the event was judged to be ambiguous and was dropped from further consideration.  In addition, one event (Respondent divorced) was dropped due to the lack being reported often.  Thus, only 28 of the original 33 items were used in the analyses reported.
 
 

    Assignment of life events to egocentric or nonegocentric categories was done according to the scheme described by Aldwin (1990), and operationalized from the GSRLE by Nauta and his coauthors (Nauta et al., 1994).  Egocentric items are those which represent circumstances that impact directly upon the individual, while those which impact on family members are described as nonegocentric.  Table 1 also shows the various events ranked by their frequency in this sample.

    Two measures of psychological well-being were used in this study, one measuring psychiatric symptoms of impairment, herein referred to as mental health, and one measuring morale.  The mental health measure is an additive scale consisting of 15 items from Langner's Twenty-Two Item Scale (1962).  It was designed to provide a measure of psychological distress relevant to the aged (see Haug, Belgrave & Gratton, 1984).  Seven items relating to physical symptoms often unrelated to psychiatric distress among older adults were not asked of respondents.  Scale items included are psychophysiologic symptoms, symptoms of anxiety, and symptoms of depression.   Responses to the mental health scale were coded so that a higher score indicated better mental health (fewer symptoms).  The reliability coefficient (Cronbach's alpha) for the 15 items in the current study was .78.

    Nine of the 22 items from the Philadelphia Geriatric Center (PGC) Morale Scale (Lawton 1972; 1975) were asked of respondents to assess their psychological well-being.  The nine items included three items measuring each of the three dimensions of morale described by Lawton in his 1975 modification of the scale (Agitation, Attitude Toward Own Aging, and Lonely dissatisfaction). The morale scale was coded so that higher scores represented higher morale.  Cronbach's alpha reliability coefficient for the nine-item scale used with this sample was .76. 

    By using two different measures of psychological well-being--each tapping somewhat different content--the current study provides a broad range of outcomes to assess the differing effects of egocentric, nonegocentric, positive, negative, and combined stressors on the two racial groups represented in the data set.

    The first hypothesis regarding the number of events experienced is examined by comparing the mean number of each type of life event experienced for Blacks and whites using analysis of variance (Table 2).
 
 

    The results indicate that there is a significant difference in total events experienced; however, the difference is opposite that hypothesized.  Rather than experiencing more events than whites, blacks experienced fewer events.  Nearly all of this difference is due to black respondents showing fewer positive life events.  In fact, the number of negative life events is virtually equal for the two groups.  If the concept of "double jeopardy" is to apply here, it is in a negative sense of reducing opportunity for positive events, rather than increasing the likelihood of negative events.  Hence, these results contradict the double jeopardy hypothesis.

    The second hypothesis dealt with the differing impact of life events on the two races, with the expectation that black respondents would be more adversely affected than their white counterparts.  Table 3 presents results of correlations between the dependent measures (mental health and morale) and the types of life events, for the total sample and by race.  When one looks at the correlations for all respondents, it appears that not all life events have a significant impact on psychological well-being.  Specifically, both egocentric and positive events show no significant affect on either mental health or morale.  Nonegocentric events do not significantly impact morale, although the correlation approaches significance (p = .054).

    However, when one looks at the two racial groups separately, a number of interesting contrasts arise.  While all the correlations for whites are negative (although not all are significant), the majority of the coefficients for black respondents are positive (although most are not significant).  In contrast to whites, whose mental health appears to be adversely affected by positive events (r = -.1264, p = .043), black respondents are positively affected by positive events (r = .1419, p = .034). Since these groups are nearly equally represented in the total sample, the net effect appears insignificant when all respondents are considered.  Similar differences occur for other event type/well-being associations, although the example cited is the only one in which both coefficients are statistically significant and are opposite in impact.
 Examining if mental health differences in the variables are significant cannot be assessed simply by comparing correlation coefficients.  Others (Ferraro & Farmer, 1996, Nauta et al. 1994, Jaccard, Turrisi & Wan, 1990) suggest using multiple regression to look for interaction effects between variables. Hence, we examined the impact of the race/event type interaction on the measures of well-being using multiple regression, controlling for the independent affects of race and event type.  We tested all possible interaction term/dependent measure models, but only five are statistically significant.  All five are related to the impact of life events on mental health.  Next, we examined these five significant regression models with and without the interaction terms.  The impact of race is statistically significant and positive in all five models when the interaction term is excluded.  However, when the interaction of race and life event type is included, the coefficient for race is no longer significant.  A second noteworthy point is that adding the interaction term into the model increases the impact of life event type on mental health for all event types.  In three cases (egocentric, positive, and all events), the coefficient for event type goes from nonsignificance to being statistically significant.
 
 

    Still another insight from examining the effect of the interaction term is that for all event types, the coefficient for the interaction term is positive.  Using a simple formula (Jaccard et al., 1990), one can calculate the separate slopes for the predicted effect of life event on mental health for the two races.' The formula is: b, at X, = b, + b,X2' where b, is the coefficient for event type in the model, X2 is the value for race (1 = White, 2 = Black), and b, is the coefficient for the interaction term in the model.   Thus, for white respondents, nonegocentric events have a slope of -.50 [-.98 + (1)(.48)], while the slope for blacks is -.02 [-.98 + (2)(.48)]. Similar calculations can be made for egocentric events (whites = -.44; blacks = +.22), positive events (whites = -.31; blacks = +.37), negative events (whites - .68; blacks = -.11), and all events (whites = -.41; blacks = +.06). These results indicate a lesser negative impact (and sometimes a positive impact) of all types of life events for black respondents compared to white respondents.

    Taken together, these results do not support the hypotheses.  Black respondents did not experience more life events (either in total or of the negative type) than whites, nor was the impact of these events on mental health more adverse for blacks than whites.  These results do  not support the double jeopardy hypothesis.

    The results do not indicate a great deal of support for the double jeopardy hypothesis (the exception being fewer positive life events experienced by Black elderly in our sample).  In fact, the opposite appears to be true, in that elderly Blacks do better than whites in withstanding the potential negative impact of stressful life events on mental health.

    There are a number of potential explanations for this discrepancy.  One has to do with the people sampled.  It may be that the relatively equal social status of the black and white respondents, having all retired from the union paid careers with ample and equal health care benefits in the auto industry, masks differences that might be more evident if a more diverse sample were to have been used.  For example, Ulbrich, Warheit and Zimmerman (1989) note in their study that lower-SES blacks are more vulnerable to undesirable events than middle-SES blacks.

    It is also possible that the measure of stressful life events does not include enough of the type of events one might experience as the result of double jeopardy.  An alternative to using only a specific a list of events with a relatively recent time period does not fit with the way others defined the impact of race on health (see Ferraro & Farmer, 1996).  Nonetheless, if the results of this study accurately represent the impact of stressful life events on elderly of the two races, then one needs to explain why the double jeopardy approach prediction for minorities to be at greater risk was not supported here.  In fact, African Americans in our sample reported less experience for most events, and a reduced impact of those events on their mental well-being.

    One possibility for the results obtained here is that elderly blacks may have more social support than their white counterparts.  If true, the higher level of support may act as a buffer against the strain caused by life events experienced by black elderly.  The level of social support was not included in this study, but should be in future studies as it may provide some explanation for the findings.

    We are confident that the research reported here is useful to researchers interested in exploring the stress of later life and health disparities across races.  We do see the value in disaggregating stressful life events into subcategories.  We also call for others to explore samples of diverse persons with relatively equal social status for access to health care and incomes so that assessments of the double jeopardy hypotheses can continue.  Unequal access to healthcare may be of greater importance than previously realized.  In the United States it is alarmingly common for employers to discontinue these benefits for current work force and retiree benefits.  In other words, given the changing status of American automobile companies, we hope a replication and extension of this approach will be conducted soon. 

    In this project we tested the double jeopardy hypothesis that predicts health disparities for minorities in later life.  The research design allowed an investigation of race differences in mental health and stress for a group of comparably compensated retirees. We found virtually no support for the hypothesis that blacks would experience more stress or be more vulnerable to it.  We attribute this non-finding to the sample having similar occupations, good incomes and health benefits.  It appears that older African Americans do as well as white retirees in terms of encountering stress and coping with it.  We hasten to point out that more research should be conducted using this methodological approach. Urgency is needed because of job losses in the industry.  We fear the ongoing economic changes (job losses and benefit cuts) in this industry are limiting our chances to study health disparities.

    The study extends beyond examining the amount of stressful life events and the impact of stress on measures of mental health and morale.  We also explored subcategories of the stressful events and found doing so offered additional insights about stress and health.  Overall, results do not support the overarching double jeopardy hypothesis.   Blacks did not report more experiencing more stressful life events than whites.  More importantly, hypotheses about correlations between stressors and the measures of mental health were not as predicted.

    Our research suggests there is good reason to investigate the effects of different kinds of stressful events on well-being.  Research into race differences on stress and mental health may be helpful in understanding quality of life issues for individuals and also for issues of social justice and public health.  In short, investigations of health disparities using a stress paradigm should generate numerous practical findings.  However, it should be acknowledged that our results are modest at best and must remain tentative. We welcome others to replicate this research and encourage that other variables are included in the designs. It is our hope that the methods utilized in this report provide confidence that a longitudinal study with a much larger sample would be worthwhile.

Aldwin, C.M. 1990.  "The Elders Life Stress Inventory: Egocentric and Nonegocentric Stress."  In M.A.P. Stephens, J. H. Crowther, S. Hobfoll and D.L. Tennenbaum. (Eds.) Stress in Later-life Families, pp. 49-70.  New York: Hemisphere.

Butler, R. 1975.  Why Survive? Being Old in America.  New York: Harper and Row.

Chiriboga, D. A. 1982.  "Consistency in Adult Functioning: The Influence of Social Stress. Ageing and Society 2: 7-29.

Erikson, E. 1959.  Identity and the Life Cycle.  New York: International Universities Press.

Ferraro, K, &. Farmer, M. 1996.  "Double Jeopardy to Health Hypothesis for African Americans: Analysis and Critique." Journal of Health and Social Behavior 37: 27-43.

Grant, I., Sweetwood, H., Yeager, J., and Gerst, M. 1981.  "Quality of Life Event in Relation to Psychiatric Symptoms." Archives of General Psychiatry 38: 335-339.

Gerontological Society of America 1994.  Minority Elders: Five Goals Toward Building a Public  Policy Base. (Second Edition).  Special Issue- The Gerontologist: GSA:Washington, DC.

Haug, M., Belgrave, L., & Gratton, B. 1984.  "Mental Health and the Elderly: Factors in Stability and Change Over Time."  Journal of Health and Social Behavior:  25: 100-115.

Holmes, T.H. & Rahe, R.H. 1967.  "The Social Readjustment Rating Scale."  Journal of Psychosomatic Research 11: 213-218.

Jaccard, J., Turrisi, R. & Wan, C. 1990.  Interaction Effects in Multiple Regression. Newbury Park, CA: Sage Publications.

Kessler, R. & McLeod, J. 1984.  "Sex Differences in Vulnerability to Undesirable Life Events." American Sociological Review 49: 620-631.

Kiyak, A., Liang, J., & Kahana, E. 1976. "Methodological Inquiry into the Schedule of Recent Life Events." Paper presented at the American Psychological Association Meetings. Washington, DC.

Langner, T.  1962.  "A Twenty-two Item Screening Scale of Psychiatric Symptoms Indicating Impairment." Journal of Health and Social Behavior 3: 269-276.

Lawton, M. P. 1972.  "The Dimensions of Morale." In D. Kent, R. Kastenbaum and S. Sherwood. (Eds.) Research Planning and Action for the Elderly, pp. 144-165.  NewYork: Behavioral Publications.

------. 1975. "The Philadelphia Geriatric Center Morale Scale: A Revision." Journal of Gerontology 30: 85-89.

Miller, F.T., Bentz, W., Aponte, J., and  Brogan, D. 1974.  "Perceptions of Life Crisis Events: A Comparative Study of Rural and Urban Samples."  In Dohrenwend and  B.  Dohrenwend.(Eds.)  Stressful Life Events: Their Nature and Effects, pp. 259-273. New York: Wiley & Sons.

Nauta, A., Brooks, J. Johnson, R. Kahana B. and  Kahana, E. 1994.  "Egocentric and Nonegocentric Life Events: Effects on the Health and Subjective Well-Being of the Aged." Journal of Clinical Geropsychology 2: 3-22.

Tausig, M. 1982.  "Measuring Life Events." Journal of Health and Social Behavior 23: 52-64.

Thoits, P. 1983.  "Dimensions of Life Events that Influence Psychological Distress: An Evaluation and Syntheses of the Literature."  In H. Kaplan, (Ed.),  Psychological Stress: Trends in Theory and Research pp. 33-103. Academic Press: Orlando.

Ulbrich, P.M., Warheit, G., and Zimmerman, R. 1989.  "Race, Socioeconomic Status, and Psychological Distress: An Examination of Differential Vulnerability." Journal of Health and Social Behavior 30: 131-146.

Vinokur, A. and  Selzer, M. 1975.  "Desirable versus Undesirable Life Events: Their Relationship to Stress and Mental Distress." Journal of Personality and Social Psychology 32: 329-337.

Weitz, R. 2004. The Sociology of Health, Illness, and Health Care: A Critical Approach.  Belmont, CA.: Thompson. 

Acknowledgment:  Completion and revision of this project was supported by a Fayetteville State University Research Center for Health Disparities (P20 MD001089). 

Return to Sociation Today, Spring 2007