Legibility size: A two-way repeated measures mixed-model ANOVA (2 Age groups X 8 Fonts) was carried out on the legibility threshold data. Legibility threshold varied significantly across fonts (F(7,16) = 32.64, p <.05), and while elderly observers required larger letters for correct identification than did young participants (F(1,22) = 13.88, p <.05), the age difference did not interact with font type (F(7,16) = 2.03, p >.05).
Visual inspection of the mean threshold sizes for each font suggested that they fell into three general legibility groupings (see Figure 4). Pairwise ANOVA comparisons were conducted to determine if these groupings were verifiable by the pattern of pairwise differences between them. Multiple-comparison error rate was controlled using Bonferroni adjustment based on the number of significant pairwise differences (22 tests). Consistent with the legibility levels among the fonts tested, the least legible font (i.e., Garamond Condensed) differed significantly from all other fonts (see Table 6), while the legibility of the next two fonts, Univers Condensed and Century Old Style, were significantly lower than that of the three most legible fonts (The Sans, Sabon, and Times New Roman). No other pairwise comparisons were significant.
Figure 4: Mean overall legibility size of print as a function
of font type.
Table 6: Significant T Values for Pairwise Comparisons of Legibility Size for all Participants (p<.002)
| Font |
Garamond Condensed |
Univers Condensed |
Century Old Style |
| Univers Condensed | 5.71 |
||
| Century Old Style | 6.65 |
||
| Bell Gothic | 7.85 |
||
| Times New Roman Sans | 8.26 |
||
| Times New Roman | 10.11 |
4.65 |
3.68 |
| Sabon | 12.99 |
5.47 |
5.00 |
| The Sans | 10.37 |
5.62 |
4.24 |
Visual predictors of legibility: The correlations of far visual measures and mean overall font legibility for all participants and each age group separately are found in Table 7. Acuity was a strong predictor of mean overall font legibility. To a lesser extent, mean contrast sensitivity was also predictive of Mean Legibility Threshold. None of the other composite contrast sensitivity measures were related to Mean Legibility Threshold. When the age groups were considered separately, it was found that spatial vision measures were generally more predictive of legibility among young observers than elderly ones. For young observers, legibility was related to acuity. High cut-off, frequency of peak sensitivity and peak sensitivity, approached, but did not reach significance in its predictability of mean overall font legibility for young observers. For elderly observers, acuity was a good predictor of Mean Legibility Threshold size. None of the composite measures of contrast sensitivity, however, were related to legibility size.
Table 7: Person Correlations Between Far Visual Measures and Average Legibility Overall, for Young and Elderly Participants
| Overall |
Young |
Elderly | |
| Acuity | 0.894* |
0.589* |
0.837* |
| Frequency of Peak Sensitivity | 0.076 |
-0.456 |
0.183 |
| Peak Sensitivity | -0.374 |
0.410 |
-0.082 |
| Mean Sensitivity | -0.518* |
-0.003 |
-0.229 |
| High-Cutoff | -0.251 |
-0.560 |
0.106 |
Reading time: A two-way repeated measures mixed-model analysis of variance (2 Age groups X 8 Fonts) was applied to the reading time data. There were no reading-time differences among fonts (F(7,16) = 1.23, p >.05), nor was there a significant age effect (F(1,22) = 1.18, p >.05). The interaction between age group and font, however, was significant (F(7,16) = 2.33, p <.05). Tests of simple main effects revealed that Reading Time did not differ among fonts for young readers (F(7,5) = 1.23, p >.025). Although there was a significant font effect in the elderly group, the effect was not significant (F(7,5) = 1.23, p >.025) with Geisser-Greenhouse correction. There were no significant differences in Reading Time between young and older readers for any font (see Figure 5).
Figure 5: Age differences in comfortable reading time as a function of font type (presented in order of increasing time for elderly participants).
Visual predictors of reading time: The correlations of near visual measures and mean overall font reading time for all participants and each age group separately are found in Table 8. Neither acuity nor composite contrast sensitivity measures correlated with the mean reading time overall across age groups, nor within either age group separately.
Table 8: Person Correlations Between Near Visual Measures and Average Reading Time Overall, for Young and Elderly Participants (p<.05)
| Overall |
Young |
Elderly | |
| Acuity | 0.084 n.s. |
0.024 n.s. |
-0.271 n.s. |
| Frequency of Peak Sensitivity | 0.078 n.s. |
0.113 n.s. |
0.162 n.s. |
| Peak Sensitivity | -0.014 n.s. |
0.115 n.s. |
0.243 n.s. |
| Mean Sensitivity | 0.064 n.s. |
0.302 n.s. |
0.405 n.s. |
| High-Cutoff | 0.048 n.s. |
-0.170 n.s. |
0.313 n.s. |