due to equipment failure. These missing values were replaced by the average scores of the rest of the
participants.
For the purpose of follow-up analyses, children with a relatively low Peg Tapping score (the
lowest 50%) were assigned to the low group (N = 11, M = .81, SD = .05) and children with a relatively
high score (the highest 50%) were assigned to the high group (N = 10, M = .98, SD = .03). One
outlying Peg Tapping score (Z = -2.55) was winsorized and for the low Peg Tapping group one
outlying value for the total fixation duration at the text-irrelevant parts of the illustration for the first
simultaneous reading (Z = -2.52) was winsorized. This winsorized eye-tracker variable was used in the
separate analyses for the low and the high Peg Tapping groups.
Results
Data Inspection
The assumption of normality was met for children’s total fixation duration at the text-relevant
and text-irrelevant parts of the illustration, for both the simultaneous and non-simultaneous readings.
Children’s scores on the Peg Tapping Task and the Meaning Recognition Pretest and Posttest were
normally distributed as well.
For both the low and the high Peg Tapping group, the eye-movement variables – including the
winsorized variable of total duration of fixations at the text-irrelevant parts in the first simultaneous
reading – and the winsorized Peg Tapping variable had a normal distribution.
Eye-gaze Patterns
In order to answer the first research question – whether children’s looking behavior at the
illustrations differed between the simultaneous and non-simultaneous presentation of oral text and
illustrations – a 2 x 2 x 2 repeated measures analysis of covariance (ANCOVA) was conducted. Total
fixation duration within the illustrations was entered as the dependent variable, condition
(simultaneous and non-simultaneous), relevancy (parts highlighted by the text and text-irrelevant
parts), and session (first and second reading sessions) as the within-subject factors, and Peg Tapping as
the covariate. The results revealed a significant main effect of relevancy (F(1, 19) = 5.67, p = .028, ηp²
= .23) and significant interaction effects between relevancy and condition (F(1, 19) = 15.92, p = .001,
η
p² = .46), between relevancy and Peg Tapping (F(1, 19) = 5.71, p = .027, ηp² = .23), and between
relevancy, condition, and Peg Tapping (F(1, 19) = 14.68, p = .001, η
p
² = .44).
The main effect of relevancy implied that children looked longer at the parts of the
illustrations that were highlighted by the text (M = 6.12, SE = .46) than at the text-irrelevant parts (M =
6.09, SE = .28). According to the interaction effect between relevancy and condition this was only true
for the simultaneous readings (F(1, 19) = 12.94, p = .002, ηp² = .41). When text and illustrations were
presented simultaneously, children looked longer at the parts highlighted by the text (M = 6.55, SE =
.77) than at the text-irrelevant parts (M = 5.89, SE = .35). In contrast, there was no difference in
children’s looking times for the text-relevant parts (M = 5.69, SE = .49) and the text-irrelevant parts
(M = 6.30, SE = .43) when the illustrations were only shown after the text (F(1, 19) = .10, p = .754).