Martine van der Pluijm

165 General discussion to join the activities (see Chapter 3). This is in line with the theoretical model of Hoover- Dempsey et al. (2005) that predicts that parents are more motivated to be involved in their children’s development when their child invites them explicitly. The authors found that parental involvement was high, irrespective of parents’ background. However, several teachers who participated in AHL reported difficulties with this whole classroom approach and with providing tailored support to lower-educated parents’ needs. The high number of parents that participated resulted in full classrooms. We observed up to 49 people in a kindergarten classroom (one teacher, 24 children, 24 parents). This was exhausting for teachers and often difficult for them to differentiate their support to the specific needs of lower-educated parents (see Chapter 4). This problem occurred less in preschool classrooms with smaller groups (max. 14 children). These groups have two teachers who could coordinate their attention to child-parent dyads effectively. Teachers of classrooms beyond preschool generally work alone and have more children in their groups. By observing parents and children during the weekly parent-child activities, teachers gained a better understandingof theneeds of lower-educatedparents. Some teachers reported that there was too little time to be involved in activities and modeling language strategies. They would have liked to give more individual support to these lower-educated parents. This situation was even more complicated when children participated without a parent. In this situation, teachers tended to prioritize supporting the child without the parent. During our research, students or a parent-educator helped teachers to support these children. This assistance allowed teachers to focus on providing support tailored to the needs of individual parents. No structural solutions have been found for situations with kindergarten, grade 1, and grade 2 teachers with large numbers of children in the classroom, and no assistants. This situation might undermine teachers to continue applying a tailored approach towards lower-educated parents during parent-child activities in classrooms. Construing activities for fruitful parent-child interactions Our systematic observations of parent-child interactions (Chapter 5) show that interactive behavior improved from pretest to posttest. Three aspects of quality of interaction (i.e., child involvement, parental support of autonomy, and quality of emotional responsive behavior), improved in both the talk and play activities (i.e., Activity 1: a family activity, Activity 2: playing with fruit/cubes). This is an encouraging finding, considering that child initiative and parental responsiveness are important for language development (Hoff, 2006; 2013; Mol & Neuman, 2014). In addition, we found that in classrooms with a better quality of intervention delivery by teachers, cognitive support (e.g., scaffolding), all aspects of quantity (number of words used by parent, child, and number of turn-taking), and two aspects of quality of language (use of