Naslov
Logistic regression analysis of personality and cognitive variables as predictors of employee turnover

Autori
Miloš, Marino

Vrsta, podvrsta i kategorija rada
Ocjenski radovi, magistarski rad

Mjesto
Cranfield, UK

Datum
31.03

Godina
2009

Stranica
40

Mentor
Kwiatkovski, Richard

Ključne riječi
turnover ; prediction

Sažetak
Methodologically founded employee turnover prediction models could be useful for early anticipation of voluntarily employee turnover. Such models may provide managers with time sufficient to intervene and prevent possible departures from the organization. That way negative consequences caused with turnover (financial costs, loss of social and/or intellectual capital, decreased motivation, job performance and organizational productivity) could be minimized. Using Cattell’s 16PF Questionnaire and tests of cognitive abilities (logical, verbal and numerical reasoning) this study tries to identify whether dispositional variables may give rise to the probability of voluntary employee turnover. Using data on turnover from leading Croatian bank (two years time span) a sample of 70 cases was created that consisted of employees that left the bank after the collection of psychometric data. Control sample of 210 cases was randomly created representing rest of 3546 employees that stayed at the bank. A logistic regression analysis was conducted in order to propose a model that will estimate the likelihood of separating from the firm. Exploratory nature of the study imposed a stepwise strategy in choosing a subset of variables. Method yielded the following significant dimensions that are included in the final model: logical reasoning, warmth, reasoning, sensitivity, independence and self-control. Correlating with previous research on Big Five inventories results partly confirmed importance of consciousness (self-control in 16PF) and openness to experience (independence in 16PF). Further, the model achieved a test sample correct classification rate of 82.7% which suggests that a simple logistic model is able to identify possible departures, and can become a useful tool for managers. Although model is adding visible value in valuation of dispositional variables in explaining turnover and improves utility of actual selection practices, it should be used with caution due to possible ethical issues and dilemmas (discrimination and/or impression management).

Izvorni jezik
Engleski

Znanstvena područja
Ekonomija, Psihologija, Interdisciplinarne društvene znanosti, Interdisciplinarne humanističke znanosti

POVEZANOST RADA