Martine van der Pluijm

145 How to support lower-educated parents? Table 5.6 shows significant changes in cognitive support (quality of the interaction), all three aspects of the quantity of the interaction, and decontextualized language, and other language (quality of language). Inspection of the means shows that parents in the high delivery group increased more in these aspects compared to those in the low delivery group. Parents in the higher delivery group increased their cognitive support ( M pretest: 2.15, M posttest: 2.78), while the low delivery group showed a decrease ( M pretest: 3.06, M posttest: 2.37). Children increased their number of words in the higher delivery group ( M pretest: 68.73, M posttest: 128.73), whereas the low delivery group showed a decrease ( M pretest: 157.00, M posttest: 68.25). The amount of decontextualized language in the high delivery group increased ( M pretest: 7.67, M posttest: 16.20) but decreased for the low delivery group ( M pretest: 18.75, M posttest: 10.25). The partial eta squared effect sizes range from .22 (amount of decontextualized language), .38 (number of words) to .48 (cognitive support), which can be defined as medium and strong effect sizes (Cohen, 1988). The medians of the Mann-Whitney tests at pretest and posttest show that parents in the high delivery group improved more ( Mdn = 11.73) compared to those in the low delivery group ( Mdn = 3.50 ) in the number of words used by parents, in the amount of turn-taking ( Mdn = 11.73 and Mdn = 3.50, respectively) and in other language ( Mdn = 11.63 and Mdn = 3.88, respectively). This growth of r .60 and .61 represents a large effect size (Field, 2009). The high delivery group improved on these dimensions during Activity 2, whereas the low delivery group did not. DISCUSSION The two studies investigated whether the AHL program contributes to lower-educated parents supporting oral language development of young children. The results of our first study show that the intervention contributed to the high participation of all parents in the classroom. At the posttest, almost all parents (98%) reported that they had participated frequently in weekly parent-child activities. The results show that the parents already had positive perceptions of SFPs directed at language development support from the start of the program. In addition, their self-efficacy beliefs, and self-reported HLE were also quite high from the start. In general, these perceptions did not change significantly. However, parents with the lowest education levels had higher ratings of the HLE at the posttest compared to the pretest. Additionally, parents’ open answers at the posttest indicate a positive development of the SFPs to support oral language development. Several parents with the lowest education levels recalled activities and teachers’ suggestions at school and at home to support oral language development. No differences in effects were found when we compared the quality of delivery in classrooms. The results of our second study show a significant development of child involvement, parental autonomy, and emotional support (i.e., three aspects of the quality of interaction) in both