The generalizability of Caregiver Strain Index in family caregivers of cancer patients.

The purpose of this study was to estimate the variance components associated with individual differences and various sources of measurement error of Caregiver Strain Index (CSI). A two-facet (item and occasion) crossed design generalizability study (G study) was conducted using a sample of 14 caregivers of cancer patients. Based on the findings of the G study, alternative decision studies were then designed to search for optimal generalizability of CSI. The measurement of CSI was not reliable with a single test. Averaging two or more measurements would make CSI scores more generalizable for either relative or absolute assessment of caregiver strain.