Glenberg, A. M., & Epstein, W. (1985). Calibration of comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition, 11(4), 702–718

Early experiment demonstrating poor comprehension monitoring in reading: university students’ judgments of their Reading comprehension are quite poorly calibrated. i.e. People often struggle to remember details of prose text because they never processed them in the first place

This figure depicts the calibration curves from the first experiment. “Delayed” here refers to the Judgment of learning—their estimate of comprehension—rather than to the testing. Students read 15 passages and were tested after all readings were complete in both groups. The “immediate” group judged their comprehension of the passage immediately after reading, and the “delayed” group judged all passages en masse after reading the last one.

In the second experiment, the authors guessed that perhaps the students weren’t sure how difficult the inferences would be. They added a “familiarization” condition” which showed students examples of inference questions and answers for the first three texts. This didn’t significantly improve calibration.

Finally, in a third experiment, the authors wonder if students would produce better calibrations if they tried performing some inferences, then got another chance to estimate their comprehension. Here, “initial calib” refers to their initial assessment, “calib. of performance” to their self-assessment of their performance on the initial inference question, and “recalibration” to their revised estimate of comprehension.

Correlation of the initial calibration is -0.04; correlations of the latter calibrations were 0.33 and 0.38 respectively.

This “updating on experience” finding aligns with The mnemonic medium may push readers to read more slowly and attentively.