Martine van der Pluijm

141 How to support lower-educated parents? dyads, listening to their conversation, and exchanging backgrounds. During our observations, we did not join this conversation of dyads and made sure that the parent-child conversations were not influenced by others. The second activity was designed for use in classrooms where parents were familiar with parent- child activities. This activity focused on stimulating rich interaction between parent and child by taking turns, eliciting language, and having fun. Parents were encouraged to challenge their child to think and talk, instead of directing the child to give correct answers. The dyads with the younger children (aged 4-6) played hide and seek (What’s gone) with wooden fruit. Different kinds of fruit (e.g., apple, lemon, orange) were put on the table and covered. We showed how a piece of fruit was removed and explained that they were going to play the game together. They were going to guess which fruit the mother (or father) or the child had removed and continue turn-taking. Dyads with the older children (aged 7-8) played Rory Story. Rory’s Story cubes, a set of nine six-sided dice, each with a different image on them (e.g., glasses, the sun, a bicycle), are meant to inspire storytelling and creative play. Parent and child take turns to make a story based on the image. Again, we explained how child and parent were invited to play together by throwing dice in turns and thinking of associations based on the images that appeared. Coding of the interactions We used the first ten minutes of each of the two activities (average Activity 1: 12 minutes and 16 seconds, Activity 2: 13 minutes and 57 seconds) for our analyses. Videos were transcribed and coded by using transcriptions and video recordings. Videos were translated to Dutch (from Turkish, Arabic, Berber, and Urdu) by the students that were involved and checked by lecturers who master these languages. We developed a coding scheme based on three dimensions: the quality of the interaction, the quantity of the language, and the quality of the language: Quality of the interaction : We used a coding scheme based on the scales of Erikson, Sroufe, & Egeland (1985) and Landry et al. (2008) that measure child involvement and aspects of parental responsive behavior. We used four constructs (see Appendix A): 1) child involvement, 2) parental support of autonomy, 3) parental emotional responsive behavior, and 4) parental cognitive responsive behavior. We used a 5-point scale to measure frequencies of observed behavior (1= none, 2 = sometimes, 3 = several times, 4= most of the time, 5= continuously). Quantity of language : We counted the total number of words used by both children and parents (Boyce et al., 2010) and the total number of turn-taking (Jiménez et al., 2006). Quality of language : We used an adapted version (see Appendix B) of the coding scheme of De la Rie (2018) based on levels of abstraction of Blank et al. (1978), Van Kleeck et al. (1997), and communicative functions of speech by Joyner (2014). We compressed the coding scheme