The Testing Effect and Judgments of Learning in Memory for Number Facts

Document Type


Degree Name

Doctor of Philosophy (Ph.D)


Psychology and Special Education

Date of Award

Spring 2016


Previous research using verbal materials has shown that intermediate testing can enhance the long-term memory of individuals (e.g., Roediger & Karpicke, 2006a). Nikopoulos (2014) investigated testing as a tool in the improvement of long-term memory for math facts and found that individuals who were repeatedly tested recalled more math facts than those who re-studied. The current study builds on Nikopoulos (2014) and included two experiments to examine whether participants’ assessment of their own learning differed when they repeatedly restudied information or were repeatedly tested over the information. The purpose of the first experiment was to analyze differences in individuals’ sensitivity to retention intervals when individuals were asked how many math facts they thought they would forget versus how many they thought they would remember. Extant testing effect studies have only focused on predictions framed in terms of remembering (Agarwal, P.J., Karpicke, J.D., Kang, S.H., Roediger, H.L., & McDermott, K.B., 2008; Karpicke & Roediger, 2008). The goal of Experiment 2 was to investigate individuals’ monitoring accuracy of their learning at an item level. The focus of much of the testing effect research has been global predictions (e.g., Agarwal et al., 2008; Roediger & Karpicke, 2006b).Therefore, item-by-item judgments of learning (JOLs) were the focus of Experiment 2. Although previous research (Nikopoulos, 2014) found a testing effect, neither of these experiments replicated this effect, but the numerical trends in both experiments indicated individuals in the Test Condition recalled more math facts than those in the Study Condtion. Of primary concern in Experiment 1 were participants’ Global JOLs; those who were repeatedly tested predicted lower recall of math facts when predictions were framed in terms of forgetting compared to remembering.When participants’ item-by-item JOLs were measured in Experiment 2, participants assigned higher confidence ratings to math facts in the last learning block and to tie math facts. The latter finding may indicate that individuals adjust their confidence levels based on what they think will be easier to remember.


Lacy E. Krueger

Subject Categories

Education | Educational Psychology