EXPRESS: What an Indirect Measurement Approach Reveals About the Susceptibility of Directly Assessed Metamemory Beliefs to Demand Characteristics and Ad Hoc Belief Construction

It is widely accepted that predictions of one's future memory (judgments of learning, JOLs) rely on metamemory beliefs. However, many open questions remain regarding the nature and assessment of metamemory beliefs, including the extent to which strategic influences such as demand characteristics or ad hoc belief construction afflict established measures. The present study addresses this question using an indirect measurement approach. In three experiments, we assessed metamemory beliefs about font size (Experiments 1 and 2) and relatedness (Experiment 3) with two indirect measures (Experiments 1 and 3: Implicit Association Test, Experiment 2: Propositional Evaluation Paradigm) and an established direct measure (global differentiated performance predictions). Results revealed that directly assessed metamemory beliefs captured the beliefs people hold and appeared to be largely unaffected by demand characteristics. At the same time, findings suggested that directly assessed beliefs were susceptible to ad hoc belief construction. In three experiments, directly and indirectly assessed metamemory beliefs were unrelated, and only directly assessed beliefs predicted the effects of font size and relatedness on JOLs. These findings suggest that established belief measures are suitable for assessing metamemory beliefs and their contribution to cue effects on JOLs but may not always provide an uncontaminated measure of pre-existing beliefs.