Results

Structural Modeling

A preliminary structural model was proposed to account for the observed relationships between the reading and vision measures (see Figure 12). It was hypothesized that visual measures (Acuity, Mean CS and Peak CS) would determine the legibility of printed material, which would in turn predict its readability (i.e., Reading Time, Reading Distance and Composite Subjective Ratings, derived from Mean Reading Ease, Reading Difficulty, Clarity of Print and recoded Simplicity measure, due to its inverse relationship with the other subjective reading measures). The visual measures were interrelated (Far Acuity with Far Mean Contrast Sensitivity r(22) = -0.61, p < .05; Far Acuity with Near Peak Contrast Sensitivity r(22) = -0.41, p < .05; and Far Mean

Figure 12: Initial proposed theoretical model of the relationship between vision, legibility and reading performance.

Contrast Sensitivity with Near Peak Contrast Sensitivity r(22) = -0.66, p < .05) and all three variables predicted Legibility (r(20) = 0.89, p < .05). With all three visual measures predicting Legibility, only Far Acuity provided significant independent predictability for Legibility Size (Beta = 0.92, p < .05 for Far Acuity; Beta = -0.01, p > .05 for Near Peak Contrast Sensitivity; and Beta = 0.05, p > .05 for Far Mean Contrast Sensitivity). The correlation between Far Acuity and Legibility was significant (r(22) = 0.894, p < .05) accounting independently for 80% of the variance in Legibility Size. The additional information provided by Mean Contrast Sensitivity and Near Peak Sensitivity only provided an additional 0.1% predictability in the Legibility variance. Thus, once the Acuity/Legibility relationship was specified, Mean Sensitivity and Peak Sensitivity added little information. R-to-T transformations for dependent measures were conducted to determine if far acuity was a better predictor of legibility than the other visual measures. Far Acuity was the best predictor of Legibility, however, it was only marginally superior to Near Acuity (see Table 17).

Table 17: R-to-T Transformations to Examine Differences Between Far Acuity and Far Contrast Sensitivity Measures and Near Acuity as Predictors of Legibility
Measure T Value Measure T Value
CSF 3 c/deg 1.7334* Peak Frequency 6.6391*
CSF 6 c/deg 2.1559* Peak Sensitivity 3.1695*
CSF 12 c/deg 3.9430* Mean Sensitivity 2.1284*
CSF 24 c/deg 4.5326* Near Acuity 0.9468
* = Significant at p< .05

The different measures of readability were not related to one another; Reading Distance and Reading Time r(22) = 0.039, p > .05; Reading Distance and the Composite Subjective Ratings r(22) = 0.159, p > .05; and Reading Time and the Composite Subjective Rating r(22) = 0.139, p >.05). Only Reading Time was associated with Legibility (Reading Time Beta = 0.473, p < .05, Reading Distance Beta = -0.102, p > .05, Composite Subjective Rating Beta = -0.010, p > .05). The correlation between Legibility and Reading Time was significant (r(22) = 0.470, p < .05). Thus, this model was reduced to a very simple one in which acuity predicted Legibility Size, and thus in turn related to Reading Time (see Figure 13).

Figure 13: Reduced theoretical model of the relationship between vision, legibility and reading performance.

This page last updated June, 1998 by: Kevin Connolly - <gkconnol@acs.ucalgary.ca>