Validity, Concurrent

Assessment

Validity is context sensitive, particularly in the domain of social sciences. As new contexts are defined, new measures are in constant demand to advance theoretical assumptions or conclusions about predicting outcomes. The new measures must be tested, or validated, for accuracy. A commonly accepted validation process is to compare simultaneous results generated from new measures with results generated from previously validated measures—concurrent validity. The question becomes whether or not the two measures simultaneously predict the same outcomes within various contexts.

Consider the following example involving a communication scholar who wants to improve predicting the trustworthiness that employees have toward their managers. Over decades of replication, scholars have accepted the measure of interpersonal similarity as the standard by which to predict the trustworthiness employees have toward managers. That is, the more similarity employees believe to share with managers, the more employees tend to trust the managers. However, the scholar is interested in improving predictability of trustworthiness employees have toward managers and has created a new instrument called the task attraction measure. The new measure claims to predict the same outcome as the standard interpersonal similarity measure, except to generate even more predictive power. The scholar collects data using both the standard interpersonal similarity measure and the new task attraction measure. Data are collected at the same time, using both measures to generate evidence toward predicting employee trust toward managers. The scholar is now ready to test concurrent validity.

[Page 1818]The objective is to generate evidence that the new task attraction measure is indeed predicting the same expected outcome as the interpersonal similarity measure, employee trustworthiness toward managers. To generate the evidence, the scholar assesses the concurrent validity of the new measure in relation to the standard interpersonal similarity measure. The process often includes regression analysis to first determine the magnitude of prediction that both measures generate. Given both measures (i.e., the new and the standard) are indeed predictive of employee trustworthiness toward their managers, a typical next step is to test for concurrent validity by determining the correlation the two measures share with each other. The interpersonal similarity measure is already accepted as valid, so given the new task attraction measure shares a high statistically significant correlation with the standard interpersonal similarity measure, the scholar establishes concurrent validity. That is, the scholar has now generated evidence in support of arguing for the validation of a new measure to predict employee trustworthiness toward managers.

...