An empirical investigation of the relationship between multidimensional degree of structure and the validity of the employment interview

Abstract

The first purpose of this investigation was to present a conceptual framework for systematically classifying interview studies according to their degree of structure. This framework was based on the premise that interview structure serves to reduce procedural variability across applicants and was comprised of five different dimensions of interview structure. Such a framework helped to clarify the defining features of interview structure and thus provided a basis from which to empirically examine the relationship between the degree of structure and validity of the employment interview. The second purpose of this investigation was to delineate the potential effects of structure on the various stages of the interview process. A conceptual integration of the structural framework with the process model of the interview developed by Dipboye (Dipboye, 1982; Dipboye, 1989; Dipboye & Macan, 1988) suggested a number of important effects that can result from standardization of each the five dimensions of structure. These effects were then used to make specific predictions regarding the impact of increased standardization along each structural dimension on interview validity. The final purpose of this investigation was to empirically examine the relationship between degree of structure and the validity of the employment interview. Results indicated that standardization of interview questions was an important determinant of interview validity. Increased standardization of questions was consistently associated with higher overall validity. Results also suggested a curvilinear relationship between standardization of response evaluation and interview validity. Increased standardization was initially associated with higher overall validity, but subsequently was associated with lower overall validity. The optimum level of response evaluation appeared to be evaluation along multiple specified criteria rather than a global evaluation or evaluation of each individual response. In addition, results suggested that mechanical combination of ratings was superior to subjective combination of ratings and that use of the same interviewer or interview panel across the entire applicant pool was associated with higher overall validity than use of different interviewers.