73 4 of items, responded to by students on a 4-point Likert-scale, are: “The teacher clearly indicates in the lessons what I amgoing to learn” and “The teacher usually makes sure we work hard during lessons” (items were formulated in Dutch). Teachers in the experimental group completed two additional, digital questionnaires during the intervention period about the use of the tool (at M1 and M2). In these digital questionnaires and in the questionnaire at M3, teachers in the experimental group reported whether they had obtained insight into where they could improve their lessons based on the feedback from the tool. In addition, teachers reported whether they had undertaken improvementoriented actions in response to the feedback during follow-up lessons (e.g., “At the end of the lesson, summarize what students had learned”, in response to item 14 of the Impact! tool) and/or outside the lesson (e.g., “I discussed the feedback with the students”, or “I discussed the feedback with my supervisor”). 4.3.6 Data analysis Descriptive statistics were calculated for the frequency of use of the Impact! tool by teachers in the experimental group. A minimum of four and a maximum of 17 measurements (tool use frequencies at M0 and M3) per teacher were included; only those measurements in which at least one-third of the students gave feedback about the quality of teachers’ instruction were used in the analysis (calculated based on the number of tool results compared with the total number of students in the class). Students who did not attend the lesson were not notified to fill in the questionnaire. Bar charts showing percentages of teachers indicating particular aspects of teaching they could improve their lesson were used to test hypothesis one of this study. The development of teachers’ professional reflection (hypothesis two) was analysed by conducting an analysis of covariance (ANCOVA), using SPSS Statistics version 24. For the variable intensity of reflection (at M3 for all participating teachers), the assumptions for an ANCOVA were checked. Independent measurements were guaranteed by the sampling procedure (randomized control). Based on looking at the Q–Q plots, the variable seemed to have a normal distribution. Levene’s test showed equality of variances. It is assumed that covariates were measured without errors. Participation in the intervention was included as a factor in the ANCOVA. The means at M3 for intensity of reflection in both the experimental group and the control group were as the dependent variable. The means at M0 for intensity of reflection were included as a covariate. Bar charts were used to examine the number of improvement-oriented actions undertaken by teachers in the experimental
RkJQdWJsaXNoZXIy MjY0ODMw