Two columns present F ration and p values for the simplest case, without covariates. The last two give the results with maternal IQ covaried. Because concurrent maternal education added little, given the other variables, results of the second covariance analysis are not shown. With one exception, the preschool X school-age interactions were not significant and are also not shown in Table 5.
A significant linear trend was found for the Verbal IQ but not Performance IQ. Similarly, when the 2 X 2 models were tested, a significant effect for preschool treatment was found for the Verbal IQ but not for Performance IQ.
The linear term was significant for the WJ Reading, Knowledge, and Written Language scores. Significant positive main effects for preschool treatment were found for the Reading and Knowledge scores. None of the four academic subjects showed independent main effects of the school-age treatment.
There were significant main effects for maternal IQ for every intellectual and academic outcome tested. Despite the strength of maternal IQ as a predictor of child cognitive and academic performance, however, the covariance analyses indicated that the effects of treatment were essentially independent of maternal IQ. The mother’s educational level generally added little to the prediction of child cognitive and academic outcomes. The one exception is that, with both covariates in the model, there was a significant preschool X school-age interaction for the Verbal IQ score, F (1,85) = 4.70,p < .04. Examination of the adjusted means suggests that, only for children untreated in preschool but given the school-age program, adjusting for levels of maternal IQ and the environmental mother’s education does strengthen the apparent effect of treatment on children’s performance.
Analysis of variance followed by Tukey’s studentized contrasts showed that, for all three WISC-R and all four Woodcock-Johnson scores, the LPS group scored significantly higher than all groups of high-risk subjects on IQ and academic test scores.
Grade retention and use of special services. –Grade retention and the use of special services during the first 7 years in school for the Abecedarian and LPS groups are summarized in Table 6. After 7 years in school, there was a strong trend for children with preschool treatment to be retained in grade less often, but a log linear model analysis indicated that the association between earlier treatment and grade retention was not statistically significant at the .50 level.
Similarly, a strong trend for children having preschool treatment to avoid placement in special education (code 2) did not attain statistical significance. The log linear analysis did show a significant preschool X school-age interaction, such that, as predicted, CC children were more likely than the CE group to be assigned to special education, but EE children were more likely than EC children to be so assigned.
When the data for the use of related services (code 1) or special education (code 2) were combined, it developed that, across the four Abecedarian groups, from 64% to 76% had received one or both forms of help. Combining retention and special education placement to create a single (negative) index of school progress shows that children with preschool intervention were more likely to avoid these pitfalls.
The