Content area
Abstract
Health and Retirement Study data from 2006 was used to examine whether health has a moderating effect on the relationship between age and depression. The sample consisted of 9,988 respondents in the 2006 data sample. Ages ranged from 41 to 102 with an average age of 69 and a standard deviation (SD) of 11.0 years. Gender, race/ethnicity, marital status, number of years of education, work status, and perceived health status were each analyzed by Ordinary Least Squares to determine the true population relationship between each of these factors and the depression of the participants. The majority (54%) of the sample was female; 50% were married, and 25% were widowed. Only 26% were working full-time. The large majority were Caucasian, with 16.5% being African-American. The mean level of education in years was 12.4 with an SD of 3.3 years. Health, measured by ADL and IADL was allowed to freely vary, rather than restricting its variance by selecting only those with ADL or IADL limitations. The average level of depressive symptomatology, ranging from 0 to 8, was quite low with a mean of 1.6 depressive symptoms and an SD of 2.0. This indicates the average person experienced less than 2 symptoms. DSM-IV, for example, requires 5 or more for a diagnosis of depression. Self-rated health had the largest unique effect on health (B = -.25), indicating that as health declined, the number of depressive symptomatologies increased. ADL limitations exerted the next largest unique effect (B = .18), indicating an increase in depressive symptoms with increasing ADL limitations. IADL limitations operated in a similar manner. As hypothesized, age was significantly related to the level of depressive symptoms with the number of symptoms growing with increasing age. With the present data it seems that age does have a significant, independent effect on depression for those suffering from at least one limitation of ADL or IADL. Only one interaction, age by IADL, had a significant unique effect on depressive symptomatology (t(1) = 3.4; p<.01). The contribution of the set was trivial, explaining only .004% additional variance in depressive symptomatology.
