CONSTRUCT VALIDITY ISSUES IN ORGANIZATIONAL BEHAVIOR RESEARCH
Eugene F. Stone-Romero State University of New York at Albany
In the past several decades researchers in organizational behavior (including the related fields of industrial and organizational psychology, vocational psychology, occupational psychology, human resources management, and vocational psychology) have devoted a great deal of attention to the development and testing of theories. Progress in such effort hinges on the use of research methods that produce valid findings. Unless studies use appropriate measures, manipulations, statistical tests, subjects, and so forth, inferences that stem from them are of little or no value in theory building, theory testing, or the use of theory based interventions to change individuals, groups, or organizations.
The overall validity of research is a function of internal, external, statistical conclusion, and construct validity (e.g., Cook & Campbell, 1979). As has been noted by Cook and Campbell ( 1979), research has (a) internal validity to the degree that it provides a sound basis for inferring that one variable (e.g., job design) causes another (e.g., job satisfaction); (b) external validity to the extent that its findings can be generalized to and across different types of persons, settings, time periods, and operational definitions of cause and effect constructs; (c) statistical conclusion validity to the degree that it allows for valid statistical inferences about the degree of association (covariation) between measured or manipulated variables; and (d) construct validity (CV) to the extent that measures or manipulations used in research (i.e., operational definitions of constructs) provide complete and unbiased representations of the underlying constructs. The focus of this chapter is on this latter type of validity. However, this limitation in focus is not meant to imply that the other three facets of validity are of lesser