Examining the Predictive Validity of an Instrument: Technology Attitudes and Learning

Abstract

ABSTRACT This article reports the results of a study examining the predictive validity of a computer attitude instrument. The researchers attempted to determine the extent to which this instrument predicts student learning. Data from two universities were collected using this instrument over a nine-year period and were sorted into three sets with a random n of 400 in each. Three procedures were performed. First, one set of data was used to develop a base model of prediction. Second, this model was used to calculate the predicted learning achievement scores for the other two sets of data. Finally, in those two sets of data, the means of the predicted and observed learning achievement scores were compared using inferential statistics. The predictive validity of the instrument was confirmed, as no significant differences were found between the mean predicted and observed learning outcome scores.