Assessing children’s speech using picture-naming: The influence of differing phonological variables on some speech outcomes

Abstract

Purpose: Sometimes, children’s speech is judged as typical on picture-naming tests and not on other speech tests, even other picture-naming tests. This study attempts to explain this observation by examining the impact of increasing the numbers of polysyllabic words in a picture-naming test on the outcomes of percentages of consonants and vowels correct and phonological processes.

Method: The participants were 283 randomly sampled, typically-developing children, aged 3–7-years. These Australian English speakers named pictures, yielding ∼166 selected words, varied for syllable number, stress and shape, which repeatedly sampled all consonants and vowels of Australian English.

Result: Most mismatches occurred in di- and polysyllabic words, with few in monosyllabic words: the usual words in picture-naming speech tests. Significant age effects existed for percentages of consonants and vowels correct and for 12 of the 16 phonological processes investigated. Many age effects were present at 7 years; later than often reported.

Conclusion: These findings imply that additional di- and polysyllabic words add value to routine speech testing. Further, children, especially school children, who “pass” speech tests dominated by monosyllabic words may benefit from further examination of their productions of di- and polysyllabic words before finalizing clinical judgements about their speech status.

Keywords:

• Assessment
• children
• phonology

Acknowledgements

The Channel 7 Children’s Research Foundation of South Australia (Inc) funded this study. The first two authors were employed at Flinders University when the data were collected. The authors wish to thank the anonymous reviewers for their very useful feedback and comments on earlier drafts.

Supplementary material available online

Supplementary Tables I–III

Notes

†Childcare centres cater for children under 4 years of age.

‡Hierarchically arranged data is nested, consisting of sets (the jurisdiction of South Australia) and sub-sets (schools/pre-schools, then the groups of five children) such that sub-sets are at lower rankings than the set.

†Words were penalised, irrespective of whether the constituent phonemes were accurate or not, to recognise that the integrity of other tiers, such as the foot or prosodic word were compromised.

†Please refer to Long and Fey (1996) for definitions and descriptions of each of the phonological processes except WSD, epenthesis and metathesis, which are defined and described within the text.

‡WSD was treated in a traditional phonological processes framework whilst cognisant that it disrupts the foot structure.

†The modified Bonferroni adjustment rather than the Bonferroni was used throughout because the latter is considered too conservative for a large number of comparisons (Jaccard & Wan, 1996). The comparisons are ranked according to the level of significance obtained from the multiple tests from smallest to largest. These levels of significance are then compared to a modified alpha level to determine significance. The alpha level for each comparison is determined by dividing it by the number of comparisons minus its ranking.