• Presentation

Doing What They Say: Do Surveys Predict Behavior?

Citation

Olmsted, M., Schwerin, M. J., Kline, T. L., & Wilcove, G. L. (2006, May). Doing What They Say: Do Surveys Predict Behavior?. Presented at American Association for Public Opinion Research 61st Annual Conference, Montréal, Canada.

Abstract

Retention of highly trained, qualified and motivated personnel is a high priority for most organizations. This is especially true of the military, where the all-volunteer force structure creates a constant challenge for leadership as they seek to balance the number of personnel incoming (accessions) and outgoing (losses) to meet operational needs. Over the past few years, researchers have begun using surveys to collect information related to Sailor career intentions. The basic assumption of this research is that the career intentions will predict career choice behavior. However, little attention has been given to the issue of whether or not these career intentions actually predict behavior. Published research in the civilian literature has been mixed, indicating that many people leave an organization for a variety of reasons even after indicating that they plan to stay.

The present study used data from the Navy Quality of Life (NQOL) survey and personnel records to explore the relationship between intentions and behavior. The NQOL is a recurring survey that assesses the quality of life of Sailors. Data included in analyses are from 8,165 respondents to the 1999 survey and 5,114 respondents to the 2002 survey. The analyses explores the relationship between intentions and behavior by comparing major groups of Sailors such as enlisted vs. officers, married vs. non-married, parents vs. non-parents, and major paygrade groups.

Intentions to stay or leave the Navy were more or less accurate in predicting actual behavior depending on which dataset was used. Also the relationship between intentions and behavior was somewhat influenced by the characteristics of Sailors groups (i.e., paygrade, marital status, and parental status). The results of the analyses along with information on which Sailor groups provide more accurate responses on surveys will be presented. Implications for future civilian and military survey research along with applications to organizational policy will be discussed.