Confounding as a problem in relating life events to health status in elderly individuals

Explored two types of confounding relevant to research relating major life events to physical illness in elderly individuals: (a) contamination of life events lists by health-related and subjective items; and (b) failure to control for illness existing prior to the life event assessment period. Community-residing elderly individuals (M = 70.4 years) completed two measures of health status in each of two waves of testing. During the second wave, participants also completed a life events list. Independent judges categorized the life events as health-related or not and further categorized non-health-related events as subjective or objective. Results indicated that life events-illness correlations were influenced by the presence of confounded items and that when illness present prior to the life event assessment period was controlled the relationship between confounded life events and illness decreased. These results underscore the importance of assessing possible sources of confounding when conducting life event illness research with elderly individuals.