7.4 Results of Experiment 2
The results are divided into four sections. The first section considers the Baseline measures
(baseline vocabulary knowledge test for naming and comprehension). The second section
considers the Production measures (naming task). The third section considers the
Understanding measures and is divided into two subsections (the first subsection discusses
the Direct measures of Understanding. For the present experiment the operational definition
of direct measures of understanding refers to the measures which ask about the word
knowledge explicitly (multiple choice, short questions task - (a) categorisation questions and
(b) world knowledge questions, and definition task) while the second subsection discusses
the Indirect measures OfUnderstanding. The operational definition of indirect measures of
understanding refers to the measures which ask about the word’s knowledge implicitly
(association task, contrast task, story generation task). The fourth section explores general
trends and consider comparisons between measures.
All of the research questions were explored for each post-test measure. Therefore each
section of the results considers:
1. Is there a differential impact of the type of exposure to the new lexical items that the
children receive ? (Critical differences should occur when the intervention change)
2. Does children’s performance improve with increased exposure to the lexical items ?
3. Does the children’s prior knowledge of the lexical item(s) influence acquisition?
4. Is the acquisition process influenced by the semantic domain of the lexical items ?
5. To what extent does the child’s prior lexical knowledge influence acquisition ?
Preliminary analysis of children’s performance of the four words across tasks revealed no
significant differences. Therefore, the results from all the words were added up using the 0-1
coding. Relevant statistics will be presented for each one of the questions. In general, non-
parametric statistics were applied. Because non-parametric tests are less sensitive to
significant differences, I report as significant anything which is <.05 and as a trend anything
between >.05 and <.07. All the other differences are described as non-significant. In the cases
where the variances among groups were equal (according to Levene test) parametric
statistics were used.