In educational psychology, finding the best predictors of achievement is one of the most important research problems. This research quest is socially important considering that objective measures of achievement, like the grade point average (GPA), provide tangible insights for policy makers, program evaluations, career placement, etc. Historically speaking, the study of individual differences in intelligence originates from researchers trying to predict scholastic performances (Binet & Simon, 1916). Ever since, the field of educational psychology has been pursuing this endeavor, mostly by contrasting the influence of cognitive and non-cognitive factors (Richardson, Abraham, & Bond, 2012). Researchers have often shown that past assessments of achievement (e.g., high school GPA; r = .40) and general measures of intelligence (r = .20) are strong predictors of present/future achievement. We advance that these two main correlates of academic achievement can be re-conceptualized with the proposed framework of Ackerman and his colleagues (Ackerman, 1994; Goff & Ackerman, 1992), which differentiates individuals in terms of their typical intellectual engagement and their maximal/potential performance. On the one hand, in education, assessments of typical achievement would describe the stable pattern of achievement depicted by the general trend of a student’s achievement. Past achievement (e.g. high school GPA) can thus be conceived as an indicator of typical achievement, as it denotes the capacity to learn different materials taught by different professors across different learning environments. On the other hand, assessments of maximal/potential achievement should be understood in terms of general cognitive ability that captures the capacity to treat and manipulate complex information. Working memory has been targeted by educational researchers as one of the most important cognitive ability for both short-term and long-term learning (Fenesi, Sana, Kim, & Shore, 2015). Empirical studies have demonstrated that working memory and academic achievement are related with medium-to-large effect sizes (r ≈ .30; e.g., Alloway & Alloway, 2010; Swanson & Alloway, 2012). However, most if not all studies examining this relation were conducted on samples of elementary and high school students, which can give us an inaccurate estimate of effect size for university students. In this presentation, we will report the results of three studies (n = 271, 257, 262) in which we examined the unique contribution of working memory (maximal performance) and admission GPA (typical achievement) to predict the semester GPA of university students. At the first step of our multiple regression, working memory significantly predicted semester GPA in two of our three samples (βs = .169, ts = 2.492). At the second step, admission GPA was also a significant predictor in all the samples (∆r2s =.093, ∆Fs(1,260) = 29.589) and working memory remained a significant predictor of semester GPA (βs = .116, ts = 1.837) for two of the three samples. Overall, the predictive effect size of working memory was found to be lower than what is usually observed with elementary and high school students. This result could be due to the selection procedures, which reduce the variance in intelligence from one level of education to the other (Furnham, Chamorro-Premuzic, & McDougall, 2003). Cognitive factors might hold less predictive properties at university as a result of the different demands of the collegial environment on the students. Implications for educational research will be discussed as well as future directions in investigating the role of cognitive factors in the achievement of university students.
