It is widely accepted that metamemory judgments such as people’s predictions of their future memory performance during learning (judgments of learning, JOLs) rely on an experience-based process (i.e., fluency) and a theory-based process (i.e., metamemory beliefs). Whereas research initially focused on examining the contribution of fluency to JOLs, recent research addresses the role of metamemory beliefs. Nevertheless, there are many open questions regarding the measurement and the nature of metamemory beliefs. For example, it is unknown to what extent demand characteristics afflict established measures of metamemory beliefs. The present research aimed to address this question by using an indirect measurement approach to assessing metamemory beliefs and their contribution to JOLs. In two preregistered experiments (N = 128 each), we examined whether indirect measures capture beliefs about how font size affects memory and explain the contribution of metamemory beliefs to the font-size effect on JOLs. In each experiment, participants studied words in small and large font sizes, provided JOLs, and took a free recall test. We assessed participants’ beliefs about font size indirectly through the Implicit Association Test (Experiment 1) and the Propositional Evaluation Paradigm (Experiment 2) and directly through global performance predictions. JOLs and both belief measures revealed that participants expected to remember words printed in large font sizes better than words printed in small font sizes. However, only direct beliefs explained the font-size effect on JOLs. Our results thus indicate that direct measures of metamemory beliefs are better suited to capture metamemory beliefs than indirect measures.