Estimate Maintenance is Somewhat but not Fully Explained by Motor Cost of Updating

Poster Session F - Tuesday, April 16, 2024, 8:00 – 10:00 am EDT, Sheraton Hall ABC

Julia Schirmeister¹ (jul.schirmeister@gmail.com), Britt Anderson¹; ¹Waterloo University

Previous research has found that when people are asked to keep consistent updates of their estimates of probabilities, they update only intermittently between runs of several trials. This has been used to argue that people fix one working model estimate of probabilities and only update their mental models if they detect change-points. However, this update-withholding behavior could alternatively be explained by a motor-cost confound. Participants have been given their previous response as default; this makes updating more effortful than letting the default remain untouched. In the current study, participants in the “manual” condition were given no default response; they were asked to respond anew every trial. Participants drew red or blue dots out of an urn in a computerized task. For each dot, they were asked to estimate the hidden proportion of dots in the urn. Manual participants did change their responses more often than did the participants in the “automatic” group who were given their old response as default. Some of those new adjustments did appear to reflect true new minor updating events. This supports the hypothesis that a motor threshold was indeed removed for manual participants. Nonetheless, update-withholding behavior was not entirely suppressed. Manual participants making the fewest adjustments showed undeniably intentional spontaneous response maintenance. This study demonstrates that motor cost is not sufficient to explain step-hold behavior found in previous studies. Long runs of maintaining the same estimate can be unprompted and meticulous.

Topic Area: PERCEPTION & ACTION: Other