Academic journal article Social Behavior and Personality: an international journal

The Stability of Symptom Measures: Implications for Stress Research

Academic journal article Social Behavior and Personality: an international journal

The Stability of Symptom Measures: Implications for Stress Research

Article excerpt

Forty-nine female undergraduates completed life events and stress ratings on a daily basis over a 4week period. They also completed achecklist of physical and psychological symptoms at the end of each week. Data from this study as well as others reported in the literature showed that physical and psychological symptom scores are highly stable over time. Although life events and stress ratings predicted future symptoms, the correlations diminished after initial symptom levels were controlled. The correlation between stress measured at time 1 and symptoms measured at time 2 is largely a function of their concurrent correlation at time I as well as the stability of the symptom measure. Hence, the value of computing lagged correlations between stress and symptoms has to be reexamined. Possible reasons for the stability of symptom measures are discussed.

Over the past decade, there has been an increased interest in the use of longitudinal survey designs in the study of stress and health, and self-report measures are typically used in these studies (e.g., Banks & Gannon,1988; Compas, Wagner, Slavin, & Vannatta, 1986; Cutrona, Russell, & Rose, 1986; Monroe, Bromet, Connell, & Steiner, 1986; Siegel & Brown, 1988; Xiaojia, Lorenz, Conger, Elder, & Simons, 1994). The most crucial question in this area of research is whether illness, physical or psychological, is in part a consequence of psychosocial stress. Although longitudinal surveys do not really address the question of causality, they are an improvement over cross-sectional designs and are frequently the most rigorous designs possible when practicality is taken into consideration in field settings. Tpically, a measure of stress at time 1 is correlated with a measure of health status at time 2 to see if stress can predict future health status. The time interval selected for such an analysis should ideally depend on theoretical or empirical justifications as to when stress effects are most potent, although such justifications are rarely provided in studies.

That a lagged correlation is used to interpret the possible effect of stress on health is in part based on the assumption that if health status changes over time, the degree of change should be reflected in the measurement. However, this assumption may be erroneous when only self-report measures are used. This article will show, through a presentation of our own results and a review of the literature, that self-report measures of health status are remarkably stable over long periods of time. Because health measures are stable, how stress-health correlations should be interpreted becomes an important question.

In a two-wave two-variable design, the strength of a lagged correlation between two variables depends, in part, on two factors: (a) the strength of the concurrent correlation between the variables and (b) the stability of the variables. If both variables have a perfect test-retest correlation over a specific time interval, then the correlation between any pairs of cross-lagged measures will be the same as the correlation between any pairs of concurrent measures. Of course no single measure has yet been found to have a perfect test-retest reliability. However, if the two variables under study are highly reliable, then the cross-lagged correlation may have little interpretative value beyond what is obtained from a concurrent correlation. This may seem to be an obvious fact but the potential implications of this methodological limitation have been commonly overlooked by researchers. A theoretical framework concerning the longitudinal relationship between stress and symptoms should take into account the stability of both variables. (We will not deal with the stability of stress measures such as life event scales in this article because the stability of symptom measures is a more salient issue in the current status of the theory regarding stress and symptoms.)

In the study of the relationship between stress and symptoms, stress is often taken to be the independent variable and symptoms the dependent, although the causal relationship can be the other way round (Billings & Moos, 1982; Depue & Monroe, 1986; Hammen, 1991). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.