Students may prefer multiple choice tests to short answer or essay questions because they can guess when they do not know the correct answer. In theory, accuracy for those "guesses" should be at chance. However, implicit influences outside conscious awareness may bias students toward correct answers, even when they believe they are guessing.
This phenomenon — recognition without awareness — was explored by Craik, Rose, and Gopie (2015) in a new paper in Journal of Experimental Psychology: Learning, Memory, and Cognition (PDF, 116KB).
In four experiments, participants encountered words in an encoding phase, and in a subsequent test phase were asked to select which of four words they encountered during the encoding phase. Participants also provided confidence ratings (from 0 = pure guess to 2 = fairly certain) for each choice in the test phase.
Across all experiments, participants performed significantly above chance even when they reported that they were guessing.
Such recognition without awareness (correct responses for "guess" trials) occurred regardless of whether the encoding phase task was incidental (e.g., responding to word color during the encoding phase without knowing there would be a subsequent memory test for the words themselves) or intentional (e.g., participants were informed there would be a memory test for the colored words, or word-associate pairs were explicitly studied during the encoding phase).
Moreover, in contrast to previous studies that used visual patterns as stimuli, recognition without awareness was observed under both divided attention (e.g., concurrent auditory detection task) and full attention conditions.
Finally, the magnitude of recognition without awareness was not modulated by similarity between the encoding and test context.
These results demonstrate the robustness of recognition without awareness for word stimuli, but fail to support the notion that recognition without awareness is more likely under divided attention conditions or the hypothesis that the effect is related to differences between study and test context that reduce rates of explicit recognition.
Across experiments, the authors found that when the task was more difficult (measured by low hit rates), there was a large discrepancy between the proportion of trials assigned non-guess confidence ratings and the proportion of correct responses on those trials, suggesting that participants thought they were correct more than they were.
In contrast, when the task was easier (measured by high hit rates), participants made more non-guess ratings overall, but the proportion of non-guess ratings was lower than warranted by their actual performance. Furthermore, rates of recognition without awareness increased as the task got easier and the proportion of "guess" responses decreased.
The authors interpret these results to suggest that the magnitude of recognition without awareness is driven by the general difficulty of recognition within an experiment, which influences the subjective criterion for stating that a choice was a guess: when the task is easy, participants adopt a more conservative criterion for claiming confidence in their decision, and so some trials are rated as "guesses" despite being relatively easy.
This leads to a higher hit rate on guess trials in easy tasks compared to difficult tasks, where most trials classified as "guesses" are legitimately difficult and participants are unlikely to respond correctly.
Citation:
Craik, F. I. M., Rose, N. S., & Gopie, N. (2015). Recognition without awareness: Encoding and retrieval factors. Journal of Experimental Psychology: Learning, Memory, and Cognition, 41(5), 1271–1281. http://dx.doi.org/10.1037/xlm0000137
Note: This article is in the Basic / Experimental Psychology topic area. View more articles in the Basic / Experimental Psychology topic area.