RT Article
T1 Forms of Attrition in a Longitudinal Study of Religion and Health in Older Adults and Implications for Sample Bias
JF Journal of religion and health
VO 55
IS 1
SP 50
OP 66
A1 Hayward, R. David
A1 Krause, Neal M. 1948-
LA English
PB Springer Science + Business Media B. V.
YR 2016
UL https://ixtheo.de/Record/1566095425
AB The use of longitudinal designs in the field of religion and health makes it important to understand how attrition bias may affect findings in this area. This study examines attrition in a 4-wave, 8-year study of older adults. Attrition resulted in a sample biased toward more educated and more religiously involved individuals. Conditional linear growth curve models found that trajectories of change for some variables differed among attrition categories. Ineligibles had worsening depression, declining control, and declining attendance. Mortality was associated with worsening religious coping styles. Refusers experienced worsening depression. Nevertheless, there was no evidence of bias in the key religion and health results.
K1 Religion
K1 Health
K1 Attrition
K1 Longitudinal Analysis
DO 10.1007/s10943-014-9949-5