Psychologists are emphasizing the importance of predictive conclusions. Machine learning methods, such as supervised neural networks, have been used in psychological studies as they naturally fit prediction tasks. However, we are concerned about whether neural networks fitted with random datasets (i.e., datasets where there is no relationship…

The current study evaluated the relationship between various operationalizations of the Advanced Placement[R] (AP) exam and course information with first-year grade point average (FYGPA) in college to better understand the role of AP in college admission decisions. In particular, the incremental validity of the different AP variables, above…

The overclaiming technique is a novel assessment procedure that uses signal detection analysis to generate indices of knowledge accuracy (OC-accuracy) and self-enhancement (OC-bias). The technique has previously shown robustness over varied knowledge domains as well as low reactivity across administration contexts. Here we compared the OC-accuracy…

There is much debate on the merits and pitfalls of standardized tests for college admission, with questions regarding the format (multiple-choice vs. constructed response), cognitive complexity, and content of these assessments (achievement vs. aptitude) at the forefront of the discussion. This study addressed these questions by investigating the…

A high correlation between predicted and criterion values in cross-validation may not always be a sufficient indicator of goodness of prediction. An alternative indicator of goodness of prediction is reviewed in terms of what kind of prediction is sought. (Author)

A new approach to partialling components is used. Like conventional partialling, this approach orthogonalizes variables by partitioning the scores or observations. Unlike conventional partialling, it yields a common component and two unique components. (Author/GDC)

Major objective in this study was the structuring of a predictive model that would assess combinations of variables that most effectively and parsimoniously measure and forecast college success. (Author)

Changes, like single scores, are demonstrated to be valid and reliable and potentially predictable by independent variables. Change and growth measures determined from individual examinees' test scores indicated excellent predictive value. (Author/PN)

The validity of predictors of academic performance in six postgraduate fields were compared. The fields included three liberal arts areas and three professional areas. The predictors included information about students's backgrounds, self-conceptions, values, nonacademic achievements, and curricular patterns as well as admissions test scores and…

The contribution of family environment to variation in achievement scores of children was found to be relatively small compared to the contribution of social status. Parental aspirations and their educational orientations assessed by home interviews and the implications of this study for previous environmental research are discussed. (Author/CM)

To determine if the observed correlation between two variables can be "explained" by a third variable, a significance test on the partial correlation coefficient is often used. This can be misleading when the third variable is measured with error. This article shows how the problem can be partially overcome. (Author/BW)

Paper discusses the nature of the suppressor variable, constraints upon the degree of gain in predictability by the addition of a suppressor, and the distinct aims of prediction and of construct measurement. (Authors/MB)

Study to investigate whether any increase in prediction of law school performance could be effected by a more thorough consideration of the undergraduate record. (Author)