Stronger Families in Australia study: the impact of Communities for Children
- Previous: Appendix B: Outcome measures and their internal consistency and validity
- Next: Appendix D: The validity of the contrast site as a counterfactual
Appendix C: Using the Adapted PPVT-III in the Stronger Families in Australia project
Sheldon Rothman, Australian Council for Educational Research
The Adapted PPVT-III is a shortened version of the Peabody Picture Vocabulary Test, Third Edition (Dunn & Dunn 1997), which is a test of receptive vocabulary. It is often used as a screening test of verbal ability. This adaptation is based on work done in the United States for the Head Start Impact Study, with a number of changes for use in Australia for the Longitudinal Study of Australian Children (LSAC).
Adapted PP VT-III
The PPVT-III was adapted for use with 4 year olds in LSAC by altering the administration procedures, resulting in a reduced number of items administered during testing. LSAC also used separate versions with 6 year olds and 8 year olds. SFIA used only the version adapted for 4 year olds.
The Adapted PPVT-III used in LSAC and SFIA contained 40 items, although no child was required to respond to more than 30 items. The test was administered by asking all children to respond to a core set of 20 items. Children who made 15 or more errors were then administered a basal set of 10 items. Those who made six or fewer errors on the core set were administered a ceiling set of 10 items.
SFIA results were placed on the LSAC scale using the raw score to scale score conversion table, thus allowing comparison between the SFIA sample and the LSAC sample.
A full description of the procedures for the development of the Adapted PPVT-III can be found in Rothman (2003) and Rothman, Ainley and Hillman (2007).
Sample
For the 1,708 children with scores on the Adapted PPVT-III, ages ranged from 47 months to 61 months (mean=53.6); 99 per cent were aged 60 months or older. Forty-seven per cent of children were administered only the core set (including 5 per cent who did not complete the core set); 3 per cent were administered the core and basal sets; and 50 per cent were administered the core and ceiling sets (Table C1).
Note: For the Adapted PPVT-III, it was assumed that children who were not required to answer 10 basal items had answered these items correctly. Reliability reported here is the person separation reliability (Wright & Masters 1982).
The test was calibrated using results for 1,519 children who had completed the test as designed. The test had a person separation reliability of 0.82, which is higher than the reliability of 0.76 reported for the original LSAC sample. Raw scores were translated to scale scores using conversion tables used in LSAC. Children with raw scores of 0 could not be assigned scale scores.
Results
For the 1,708 children who completed at least some of the Adapted PPVT-III, the mean scale score was 56.4 and the median was 56.1. Scale scores ranged from 5.4 to 87.8.
Quality of the test
The statistics indicate that the items used for the Adapted PPVT-III test fit the Rasch model well. This is shown in Figure C1, the item fit map. The infit mean square ranged from 0.87 to 1.21 for items 11–30 (the core set) and items 31–40 (the ceiling set). Among the core and ceiling sets, discrimination ranged from 0.26 to 0.50, with most discrimination for most items around 0.40. On items in the basal set (items 1–10), the infit mean square ranged from 0.86 to 1.02, because only 54 children (3 per cent) were administered these items.
Figure C1: Item fit map for items on the Adapted PPVT-III
- Previous: Appendix B: Outcome measures and their internal consistency and validity
- Next: Appendix D: The validity of the contrast site as a counterfactual
[ top ]