Abstract
Measuring visual performance typically requires an in-person visit to a clinic or laboratory. As tele-visits continue to grow in popularity, the need to obtain measurements remotely in the absence of a testing device is increasing. Can we utilize subjective self-report to get a reliable estimation of visual performance? To answer this question, we compared subjective estimation to objective measurement on letter recognition. Eight normally-sighted young adults participated in the study. The confusion matrices measured by Yu & Watson (2018) were used as the objective measurements for comparison. To obtain corresponding subjective estimations, we adopted similar experimental design as in Yu & Watson (2018). Single lowercase letters were presented at 10deg eccentricity left and right of the fixation. Three fonts, Courier, Times New Roman (TNR), and Century Gothic (CG), were tested. Unlike the 2018 study, our subjects did not perform a letter recognition task following each letter presentation. Instead, they viewed each letter ten times in a row and then provided self-report on their perception of the letter images. Specifically, they reported the possible identity of the letter and also letters with which they might possibly confuse it at any point, and assigned a percentage value for each reported letter (sum to 100%). A confusion matrix was constructed with the data collapsed across all subjects for each font. Consistent with the objective measurement (Courier 82%, TNR 75%, CG 71%; Yu & Watson, 2018), our data indicated highest recognition accuracy for Courier and lowest for CG (Courier 78%, TNR 75%, CG 72%). The consistency was further confirmed by remarkable correlation between the corresponding cells in the two confusion matrices (subjective and objective) for all three fonts (rs≥0.98, ps<0.0005). Our finding revealed the possibility of obtaining a reliable estimation on objective measurement in visual perception such as letter recognition performance from subjective self-report.