We investigated whether recent discoveries about the cognitive representation of numbers would predict performance in a graph comprehension task. Participants decided verbally whether statements of the form 'A > B' were correct descriptions of subsequently presented bar graphs. We observed longer decision times for horizontal compared to vertical bar graphs and for negative compared to positive number graphs. Comprehension was faster when the spatial layout of magnitude information matched an internal 'mental number line'. These results show that the design of graphs can benefit from research into number representations.

In the psychology and educational science programs at Ghent University, only 36.1 % of the new incoming students in 2011 and 2012 passed all exams. Despite availability of information, many students underestimate the scientific character of social science programs. Statistics courses are a major obstacle in this matter. Not all enrolling students master the basic mathematical skills needed to pass statistics courses. Therefore, we propose a test that measures these skills. Our aim is to examine the predictive validity of the test with regard to the statistics course and also as to overall academic success. The results indicate that a test of very basic mathematics skills helps identify at-risk students at and before the start of the academic year. The practical implications of these results are discussed. The test aids the efficient use of means for remedial interventions and supports future students in choosing a higher education program that suits their potential