Internal consistency has been measured and found to be excellent on each of the StimQ forms, with Crohnbach’s alpha ranging from 0.88 to 0.93. Test-Retest reliability is also excellent, with an intraclass correlation coefficient of 0.93 for StimQ-T.
Convergent validity has also been measured for the StimQ-T. The correlation between the StimQ-T and the gold standard, the IT-HOME inventory was .55, with p<0.001.
Correlations between subscales of the StimQ and the HOME also support the validity of the StimQ. For example, the learning materials subscale on the StimQ was most strongly correlated with the play materials subscale on the HOME, with r=0.56.
Predictive and concurrent validity of StimQ have also been measured, by looking at its association with cognition and/or language. Studies have been performed with each of the three forms.
Higher StimQ-I scores at 8 months were associated with higher Bayley MDI scores at 24 months, with a correlation of 0.38 and p=0.05. The subscale with the largest correlation with MDI was verbal responsivity, with r=.37.
The simple correlation between Bayley MDI and StimQ-T was .45, with p=0.002. In multiple regression analysis, adjusting for child’s age and sex, and mother’s age, Hollingshead SES and language, each one point increase in StimQ-T was associated with a 1.2 point increase in adjusted mean MDI, with semipartial r equals 0.42 and p=.005.
When we looked at the relationship between StimQ-T subscales and Bayley MDI, the Learning Materials and Parental Involvement subscales were each significantly related to MDI score, with correlations of 0.51 and 0.42 respectively.
When children were analyzed by quartile of StimQ, 88% of children in the lowest quartile of StimQ-T had an MDI score below 85; in contrast, only 20% of children in the highest StimQ quartile had an MDI score below 85. This suggests that children with StimQ scores in the lowest quartile are likely to need developmental intervention; likewise, children scoring in the highest quartile are likely to have normal development.
StimQ-P was significantly related to concurrent expressive vocabulary, with a correlation of 0.36, and to receptive vocabulary, with a correlation of 0.30.
Subscale analysis showed that learning materials, reading and involvement were significantly related to expressive score, while learning materials and verbal responsivity were significantly related to receptive score.
In multiple regression analysis, adjusting for child’s age and sex, and for mother’s age, SES and language, each one point increase in StimQ-P was associated with an adjusted mean 0.45 point increase in receptive vocabulary score, with semipartial r=.36 and p=.03.
In multiple regression analysis, each one point increase in StimQ-P was associated with an adjusted mean 0.5 point increase in expressive vocabulary score, with semipartial r=.35 and p=.02.