215 4.4 Further exploring the integrated approach to text structure instruction From the teaching materials developed and tested in the design study, we compiled the lesson series FLINT (Functioneel Lezen in Natuur- en Techniekonderwijs ‘Functional Reading in Science Education’). We used this lesson series in a quasi-experiment with a pre- and posttest in which the thirteen participating classes (grade 4) were divided over two conditions (Shadish et al., 2002). In doing so, we examined the effect of the integrated approach to text structure instruction in science education on students' text comprehension, summarization skills, writing skills, and subject knowledge. On none of the outcome measures examined did the text structure intervention differ significantly from the control condition. This could indicate that both approaches are equally effective and that text structure instruction has no added value over the approach in the control condition. To appreciate these results, however, it is important to consider several factors in the research design. First, the intervention was relatively short while introducing four different text structures. Short text structure interventions tend to have a smaller effect on text comprehension than longer interventions, and effective text structure interventions during which three or more structures are offered tend to be longer than our intervention (Pyle et al., 2017). A second concern is the implementation quality of the lesson series in the text structure condition. In line with results from both national and international research (Beerwinkle et al., 2018, Bogaerds-Hazenberg, 2023; Reutzel et al., 2016), we saw that teachers did not always properly grasped the essence of the lessons and, in particular, of text structure instruction. Third, the control condition contained several elements that, similar to those in the text structure instruction, have been shown to be effective for text comprehension: the combination of reading and writing tasks (Graham et al., 2018; Hebert et al., 2013), lots of student interaction about the text (Applebee et al., 2003; Nystrand et al., 2003; Pulles et al., 2022), and answering inference questions that elicited higher-order thinking processes (Hall, 2016). Moreover, teachers in the control condition appeared to have spent more time on reading lessons than teachers in the text structure condition. Finally, the outcome measures and the way these were operationalized might also have played a role, as the testing tasks differed from the tasks students performed during the intervention. Despite the fact that we found no differences in effects between the two conditions, this study still provides much information useful for follow-up research. For example, we would like to investigate the effects of longer interventions, in which more time can be spend on gradually increasing students’ knowledge of text structure, and the application of that knowledge in science-oriented reading and writing tasks. In doing so, it is essential to pay more attention to the professionalization of teachers, and to collect more data regarding the implementation of the intervention. This could also include collecting process data to gain insight into factors that appear to be contingent for students to benefit from text structure instruction. For example, if students have low vocabulary knowledge, little knowledge of 216
RkJQdWJsaXNoZXIy MTk4NDMw