Evaluating a pre-screening questionnaire's predictive validity at a large financial institution
Master Thesis
2008
Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
University of Cape Town
Department
Faculty
License
Series
Abstract
This study investigated whether a pre-screening questionnaire was able to predict job performance at a large financial institution. The pre-screening questionnaire comprised a biographical, integrity and interest subscale as predictors of job performance. A sample of 2145 job applicants completed the pre-screening questionnaire and these scores were used to establish the internal consistency of the pre-screening questionnaire. For the validity testing, a smaller sample of 449 job applicants was drawn to compare their test scores of the pre-screening questionnaire to their performance scores obtained, as the criterion. Each subscale was evaluated in terms of its predictive validity. The loglinear analysis revealed that the biographical subscale was a valid predictor of job performance. Despite being reliable i.e. integrity subscale with Kuder-Richardson coefficient of (r =.63) and interest subscale with an overall Cronbach Alpha of (r =.98), both subscales were found to be poor predictors of job performance. The findings suggest that the pre-screening questionnaire demonstrates predictive validity and utility in the selection process to some extent. Recommendations are listed to improve the overall reliability and validity of the pre-screening questionnaire.
Description
Includes bibliographical references (leaves 61-66).
Keywords
Reference:
Adams, R. 2008. Evaluating a pre-screening questionnaire's predictive validity at a large financial institution. University of Cape Town.