Validity and reliability of the Philadelphia Geriatric Center Morale Scale among Turkish elderly people

Abstract

Purpose: We translated the original 17-item Philadelphia Geriatric Morale Scale (PGCMS) into Turkish and examined its validity and reliability to determine whether it may used as a tool to measure quality of life (QOL) in a Turkish elderly sample people.

Methods: The sample included 398 participants aged 65 years living in institutions. Participants who were cognitively impaired (Abbreviated Mental Test score less than 7) or who could not answer questions for other reasons were excluded. Preliminary analysis was conducted to investigate multicollinearity, univariate and multivariate outliers, normality, item difficulty and discriminatory power of individual items. Confirmatory factor analysis (CFA) was used to test the structure of the PGCMS. By means of convergent-divergent validity, correlations between PGCMS and Turkish SF-36, correlations between PGCMS and social support scores, and correlations between PGCMS and hopelessness scores were investigated. Reliability was based on internal consistency investigated by Kuder-Richardson-20 (KD-20) and item-total correlation.

Results: By means of multicollinearity, we deleted two items. Neither univariate nor multivariate outliers were found. No items showed skewness and kurtosis value greater than recommended. A model containing 15 of the PGCMS items was found to fit Turkish data perfectly. We identified three underlying factors including agitation, attitude toward own aging, and lonely dissatisfaction similar to original PGCMS's three-factor solution. There were strong correlations between PGCMS's subscales. The correlations with the physical and mental domain in SF-36, correlations between PGCMS and social support, and correlation between PGCMS and hopelessness supported construct validity. We found satisfactory evidence of internal consistency (KD-20 was 0.92 for total PGCMS and it ranged from 0.76 to 0.85 for subscales in the PGCMS) with item-total correlations ranging from 0.39 to 0.73.

Conclusions: The PGCMS is a valid and reliable quality of life measure in elderly Turkish people.