Page 137 - Second language development of newly arrived migrant kindergarteners - Frederike Groothoff

Pedagogical practices: focus on teacher behavior 137 6.3.4 Analyses The CLASS scores The score on each of the ten CLASS dimensions (See Appendix 3) was based on scores on different behavioral indicators for that dimension. These scores were based on detailed descriptions in the CLASS protocol with which indicators were scored “low,” “medium,” or “high.” The dimensions of the CLASS then received a score using a 7-point Likert scale, with 1 and 2 meaning a low score, 3–5 meaning a medium score (where 3 means a low- medium score and a 5 a high-medium score), and 6 and 7 a high score. A low score indicated that behavioral indicators were not or seldom seen during the observation. Furthermore, it could mean that there were many missed opportunities during the observational cycle, meaning that a teacher could have used an opportunity to show indicators of that dimension, but did not. A medium score meant that some indicated behavior had been observed, but not in a consequent manner. A high score meant that the behavior had been observed often and consequently during the observation cycle. One dimension, Negative Climate, had a reversed score, meaning a score of 1 or 2 was a good score and thus a score of 6 or 7 meant there was a very negative climate in the classroom. For the sake of analysis, the scores for Negative Climate were reversed. Statistical analysis The results will be presented for the two different school types (DL2-schools and Mainstream schools). With Linear Mixed models, using SPSS 22, the effect of school type was tested, taking into account differences between teachers within groups. The analysis thus contains a comparison between the scores of the group of 11 DL2-school teachers and a group of 6 Mainstream school teachers. Rater reliability All observations with the CLASS in the present study were completed by one observer; therefore no inter-rater reliability can be calculated. Nevertheless, this observer was trained by a certified trainer of the CLASS. After the training, a test was taken and the observer met the high reliability requirements (80% agreement) with a deviation of one scale point being allowed, following the protocol of the CLASS. 6.4 Results 6.4.1 Quantitative analysis of the CLASS Looking at Figure 6.1, a comparable pattern in both school types can be seen for the scores on the 10 dimensions of the CLASS (for a description of these dimensions see Appendix 3). In both school types the teachers had the highest quality on Negative Climate (reversed coded, mean = 6.55), meaning that there was very little negativity in the classrooms in this