The distribution of children’s responses for both post tests is presented in Table 5.12. Results
from both target words are presented together, since children’s performance on the definition
task did not differ by target item.

Table 5.12 Children’s responses in the definition task for both post tests

From the responses given, “semantic properties” (use of a basic and superordinate level
word) and “functionalproperties” were the most frequent types of responses in the definition
task for both post tests. Statistical comparison among the number of properties mentioned
revealed that all the children provided less “perceptual” than “functional” (immediate:
Wilcoxon: Z=5.07, p<.0000; delayed: Wilcoxon: Z=6.1, p<.0000) and “semantic” properties
(immediate: Wilcoxon: Z=8.2, p<.0000; delayed: Wilcoxon: Z=7.9, p<.0000) for both post
tests. Furthermore, the children provided significantly more “semantic” than “functional”
properties for both post tests (immediate: Z=5.05, p<.0000; delayed: Z=3.3, p<.005).

Whether the same pattern was evident within each age group and linguistic condition was
also investigated. The distribution of responses separately by age and linguistic condition is
presented in Appendices 5.18 and 5.19 The same pattern was evident within each age group.
The older the children are, the more “semantic” properties they use in their definitions (see
Appendix 5.18). The differences were significant for the immediate (Kruskal-Wallis, I-Way
ANOVA: X²= 41.3, df=2, p<.0000) and the delayed post test (Kruskal-Wallis, I-Way
ANOVA: X²= 22.6, df=2, p<.0000).

Furthermore, the same findings were evident within each linguistic condition. Children in the
Definition condition provided more “semantic” properties in their definitions than the
children in the other conditions (see Appendix 5.19). Statistical analysis revealed that the
¹Use of basic and superordinate level words were added making one category called
semantic properties

141

<<   <   138   139   140   141   142   143   144   >   >>