Figure 1.
Frequency histogram of participant ages, in bins of width 2.5 years.
Vertical dashed lines separate the three age-subgroups examined. Dark bars indicate participants with no visual problems; lighter bars show participants who may have had a visual problem (see Methods for details). Note that in order to preserve anonymity, only year of birth was recorded. For brevity, “age” in this paper means “year of birth subtracted from year of testing”.
Table 1.
Stereo thresholds on the Frisby test for different groups.
Table 2.
Stereo thresholds on the FD2 test (in seconds of arc).
Table 3.
Viewing distances and resulting disparities (arcsec) for the Frisby stereo test, reproduced from the test documentation.
Table 4.
Contingency table showing relationship between visual problems and stereo classification on the Frisby test.
Table 5.
Contingency table showing relationship between visual problems and stereo classification on the FD2 test.
Table 6.
Number of participants in the 3 classes of stereoacuity, as assessed on the Frisby (rows) and FD2 tests (columns) for 363 participants for whom both tests were available.
Figure 2.
Frequency histogram of stereo thresholds estimated with the Frisby test.
NEG = stereonegative; unable to perform test at largest available disparity. Main plot: all participants (dark bars = “no visual problem”; light bars = “visual problem”, as in Figure 1). Subplots: age subgroups on the same horizontal axis. Solid vertical line shows median for “no visual problem” subpopulation (dark bars); dashed vertical line shows median for whole group.
Figure 3.
Frequency histogram of stereo thresholds estimated with the FD2 test.
Other details as for Figure 2.
Table 7.
Quantifying how reliably poor stereovision (threshold >20arcsec) is associated with visual problems, either any visual problem or a specifically binocular problem.
Figure 4.
Scatterplot of stereothresholds measured on the revised FD2 against those measured on the Frisby, for all 365 participants.
Participants who were stereonegative are plotted at a notional threshold of 700 arcsec. Since possible scores on both tests are quantized, we have jittered the data-points so they do not coincide: for purposes of plotting, each threshold was multiplied by a number between 0.9 and 1.1. The dashed lines indicate the floors and ceilings, i.e. the best and worst thresholds possible on each test. The central rectangle denotes the range of thresholds which were possible scores on both tests. For comparison, the blue cross indicates the 95% limits of agreement reported by Adams et al. [27] for the two tests, i.e. a factor of 1.74 for the near Frisby and 4,8 for the original FD2.
Figure 5.
Distribution of our three classes of stereovision in different age groups, as assessed on the two tests.