Statistics concepts and applications

Validity and Reliability

Questions of a study's reliability and validity involve two, separate yet interrelated issues. A study is reliable if it produces consistent results. A study is valid if it measures what it purports to measure. According to the article "Does your coworker know what you're doing? Convergence of self- and peer-reports of counterproductive work behavior" from the International Journal of Stress Management, serious questions of reliability and validity often arise with work-related behavioral surveys, as workers may underreport negative behaviors for fear of reprisals. There are also more generalized problems with self-reportage, namely the tendency to view one's self in excessively favorable or unfavorable terms, or simply to be blind to routine personal behaviors obvious to others.

To ensure that there would be reliable results in a study on the potentially dicey subject of counterproductive work-related behaviors, the study's authors instated a kind of 'peer honesty' control, namely they paired participants, gave the focal or incumbent employee a self-reported survey but also gave a similar survey to a coworker familiar with the incumbent's work situation and behavior. The coworker was physically close to the incumbent at the workplace, and their work-related behaviors were characterized by common supervision, and task interdependency. Presumably, the reliability of the responses between a monitored study and an unmonitored study could be validated by consistent reportage from the peer and the incumbent. This method was also used to control for the study's overall validity: the study would be a more valid measure of counterproductive work actions and their relationship to work stressors if an outside source validated the incumbent's responses.