Learning Research summary 2-minute read
People Reading on Screens Sometimes Misjudged What They Learned
Matched-text reviews found a small average paper edge. Some people reading on screens also estimated their own test results less accurately.
The short version
- Matched-text reviews found a small average comprehension edge for paper.
- Some people reading on screens estimated their own test results less accurately.
- Comparing predicted and earned scores separates test performance from self-judgment.
Score guesses exposed a second gap
Some people reading on screens judged their own learning less well, alongside a small average test edge for paper.
Researchers call that skill calibration. It is the match between a predicted test score and the score a reader earns. A person can expect a strong score while missing more than expected.
Reviews led by Pablo Delgado and Virginia Clinton compared tests after people read the same words on paper or screens. Paper led by a small amount on average.
Delgado's review also estimated what a similar new study might find. The range included no difference and even a small screen edge. The average paper result did not promise a paper win in each new test.
Self-paced reading made the gap visible
Rakefet Ackerman and Morris Goldsmith let people set their own study time. Those reading on screens scored lower, and their score guesses were less accurate.
In the same paper, a study with fixed time for reading found no clear score gap. The self-paced result did not appear in every setup the researchers tried.
Clinton's review also found more accurate score guesses after paper reading. It found no clear average difference in reading time itself.
These measures split reading into two results. One is what a person can answer after reading. The other is how well that person knows what will be answered.
One meta-analysis reviewed narrow screen-versus-paper comparisons
Delgado's review matched the content across paper and screens. It excluded most digital tools apart from scrolling. That made the reading forms easier to compare.
It also kept the claim narrow. The review did not test links, video, or alerts mixed into the text. Its result cannot stand in for every kind of reading on a phone.
A score and a forecast reveal different failures
A test score shows comprehension performance. Calibration appears only when that result is compared with the score predicted before the test. Neither number alone can show whether self-judgment was accurate.
The paired result separates two failures. A low score means test performance was weak. A poor prediction means the reader also misjudged that result.
One writer's proposal follows from this distinction: ask for a score forecast before showing the result. The studies did not test that feedback method, so it is an application rather than a finding.
The same low score can then point to two needs. One person needs help with the text. Another may also need a better way to judge whether the text was mastered.