Validity, Predictive

Assessment

That validity is context-sensitive means the various times at which the individual measurements are taken entail that each of the events remains contextually relevant. That is, the independent measures conceptually advance theoretical assumptions or conclusions by generating evidence to predict one another, given the same outcome criteria. A commonly accepted validation process of testing predictive validity, then, is to compare asynchronous results generated from the measure in question with results generated from an alternative, often previously validated, measure. The question becomes whether or not the two (or more) measures predict one another under the same outcome conditions within similar contexts at different times.

Consider, for example, a group of intercultural communication researchers who are interested in how relationship building takes place while living in a foreign nation. The researchers believe that language is important to building relationships, and thus become interested in generating accurate [Page 1834]predictions about individuals acquiring foreign language skills while living in a foreign nation. The researchers theorize that individuals who already perceive themselves as competent to communicate in their native language are more willing to use, and therefore learn, the foreign language.

First, the researchers measure the self-perceived communication competence that individuals report before departing to the foreign nation. A year later, the researchers test how well the same individuals score on the foreign language acquisition examination. Given the results from the self-perceived communication measure share a statistically significant relationship (typically correlation analysis) with the foreign language acquisition examination, predictive validity is established. In the context of intercultural communication and language acquisition, researchers determine the results as evidence to argue that self-perceived communication competence in a native language is predictive of foreign language acquisition while living in a foreign nation.

The objective is to generate evidence that the initial self-perceived communication competence is indeed predictive of the expected language acquisition outcome. To generate the evidence, the researchers assess the predictive validity of the initial measure in relation to the results of the language acquisition examination, 1 year later. The amount of time between measures is arbitrary, and in the field of communication, the necessary amount of time depends on the context being investigated. Often, the process includes correlation analysis between the initial measure and the later measure. However, regression analysis also helps to determine the magnitude of predictive power that the first measure (self-perceived communication competence) holds for the second measure (language acquisition examination). Given either a high correlation or a high degree of predictive power, or both, the researchers establish the predictive validity of the self-communication competence measure. That is, the researchers have now generated evidence in support of arguing for the initial measure to predict the second, later measure.

...