Experiment 2
The results of Experiment 1 favor a verbal account of
PAL deficits in dyslexia. However, the extent of these deficits
and their relation to learning over time is difficult to
discern due to low performance levels, particularly in the
verbal-output conditions. Experiment 2 was designed to
capture a wider range of performance by increasing performance
levels across PAL conditions. To do so, we decreased
the number of items per mapping condition from six to
four in visual–verbal and verbal–verbal PAL and from six
to five in visual–visual and verbal–visual PAL. Given the
differences in task difficulty between the PAL conditions
in Experiment 1, we chose to use a different number of
items in the verbal-output and visual-output tasks to better
equate performance across the four conditions. An
additional aim of Experiment 2 was to assess the nature
of the errors made by the children, and to ask whether
the errors made by children with dyslexia differ from those
made by typically developing children. We hypothesized
that if PAL deficits are driven by verbal demands, as Experiment
1 suggests, the errors made by children with dyslexia
should be phonological in nature (i.e.,
phonologically distorted word, correct pairing). If however,
PAL deficits are caused by a specific difficulty with associative
learning, we would expect a higher proportion of associative
learning errors (i.e., correct word form, incorrect
pairing). To examine this question, we tracked the errors
made by participants in visual–verbal and verbal–verbal
PAL. Finally, given the frequent co-occurrence of dyslexia
and specific language impairment (SLI), it is possible that
children with dyslexia perform worse than controls due
to concomitant oral language deficits. To evaluate this
possibility, we measured expressive vocabulary in both
groups. Additionally, we measured digit span as a proxy
of verbal short term memory, as group differences in this
factor could potentially limit learning of the phonological
forms in visual–verbal and verbal–verbal PAL.