From worker to retiree a validation study of a psychological retirement planning measure /

by Davis, Guyla D.

Yiwei Chen, Advisor In order to determine whether psychological retirement planning is a viable construct, a measure was developed and the psychometric properties of the measure were assessed. Using the approach that retirement is a role and work provides meaning that is potentially lost with retirement, a 23-item instrument was developed. Data collected from 369 full time workers were used to validate the psychological retirement planning measurement. Exploratory factor analysis was used to condense items into four cohesive factors (i.e., developing a structure, anticipation of loss, spousal/partner relationships, and identity) and confirmatory factor analysis was used to fit the factor structure to the data. Confirmatory factor analysis indicated a good fit of the scale to the data and alpha coefficient estimates of reliability for the newly developed scale were found to be good (overall scale ? = 0.92, subscale ? ranged from 0.79 to 0.91). In addition, convergent, divergent, and criterion-related validity was assessed in order to validate inferences about psychological planning from the scale. The new Psychological Retirement Planning measure was positively correlated with existing measures of psychological retirement planning measures, indicating evidence of convergent validity. The Psychological Retirement Planning measure was significantly related to measures of financial retirement planning, to a lesser degree than the relationship with psychological planning, providing evidence of divergent validity. In terms of criterion-related validity, psychological retirement planning did contribute to the explained variance of attitudes toward retirement, but results are mixed depending upon how attitudes are measured. iii While the construct still needs some further development and more supporting evidence is required, the foundation has been set to further explore this new construct. In the future, different criterion variables will be chosen to provide further validation and research needs to be conducted on different populations, such as lower SES samples as well as target individuals who are specifically low or high on worker identity salience. This dissertation is dedicated to my support team of family, teachers, and friends iv v