The Sessions

For the first and descriptive part of the study two activities were designed. The first was a group discussion about a topic belonging to the grade 4 science curriculum, which was chosen together with the four teachers involved. This activity will be referred to as the science topic discussion activity (hereafter STA). The second activity was a problem solving activity (PSA henceforth) based on the RTPM. The PSA was analysed in two ways. In the first part of the study the discussion of the groups when performing this activity (PSA) was analysed. For the second part of the study, the RTPM results obtained in each class were analyzed. These results aimed at measuring the reasoning skills in each examined group. The next sections present each activity in more details.

4.3.2.1 Science topic group discussion (STA)

As stated before, the STA was chosen as a highly representative group work activity in the L1 and CLIL science class. The topic was chosen in collaboration with the teachers. It had to be a topic that all students had previously worked on and which the teachers felt they would be comfortable with. No materials, apart from the group discussion prompt, were necessary as the idea of the prompt was to make them discuss a well‐known topic rather than assessing what they remembered about this topic. The questions aimed at promoting students’ reference to facts, reasons and opinions.

The topic chosen was living things. Two language versions of the same prompt, in English for the CLIL class (see Image 4.1) and in Spanish for the L1 class (see Image 4.2)., were designed and administered in the same way. Both classes were given 45 minutes to answer as many questions as they could. However, it was observed that the prompt was too long for the time provided so students were allowed not to finish if they ran out of time.

Figure 4.1: Extract from prompt on Living things for CLIL students.

Figure 4.2 Extract from prompt on living things for L1 students.

Both prompts included seven questions aimed to promote group discussions therefore the students were asked to reason their answers to the questions, independently whether these were correct or incorrect. Most of the questions (see Appendices 3 and 4) asked students to imagine different situations or requested

below). In this way, the prompt required no specific knowledge from the topic for students to remember but rather to use acquired knowledge to reason their answers (see Appendices 3 and 4 for the complete prompts in Spanish and English).

4.3.2.2 RTPM‐based problem solving activity (PSA)

In the science classroom, the teaching and learning process often heavily relies on specific contents and topics at hand, mainly reflected in scientific concepts. In the design of this study, this part of the learning process was reflected in the STA which was seen as a representative science class activity closely connected to a specific scientific topic. However, other elements of learning such us deductive reasoning are also a part of learning that are especially present in the science‐related subjects.

In order to take into account those elements, it was considered important to design a general problem‐solving activity (PSA) deprived of any specific topic‐related elements as it is also a valuable instrument in the group co‐construction of knowledge.

Therefore, the following two aims were stated for the PSA: (i) to design a different type of activity to trigger reasoning skills in the CLIL and L1 Science classrooms, and (ii) to obtain a more complete picture of the process of the co‐construction of knowledge in group work. Therefore, the PSA consisting of a group discussion around the RPTM was used in Part 1 of the study in order to compare its effect on students’ communicative and reasoning skills to those triggered by the topic‐related science prompt. During both sessions, students were told that the main purpose of the two activities, STA and PSA, was the discussion and that they had to aim to reach an agreement at the end. In this way teachers made sure that both STA and PSA generated group discussion, although of a dissimilar type, which was analyzed with the multi‐layered model which will be further presented in Chapter 5.

The RTPM‐based PSA in groups was used in a twofold way. As described in section 3.3.3.1, RTPM is a multiple‐choice intelligence test of abstract reasoning. It comprises 60 logical‐perceptual problems where students are asked to identify a missing item that completes a pattern presented in the form of a 2x2, 3x3 or 4x4

progressive matrix. These problems are divided into five sets of 12 items labelled from A to E, which become increasingly difficult. Below is an example of one of these items:

Figure 4.3: Example problem E7 in the RTPM

The first time students’ discussion while doing the RTPM in small groups was analysed in Part 1 of the study where it was examined as a non‐content related group reasoning discussion activity. In this case, the results of the test were not taken into account. Only the discussion around the RTPM was analysed using the multi‐layered analytical model with the aim to examine students’ use of speech functions, XXX and XXX in a small group discussion in a Science class within a content un‐related problem solving activity.

In Part 2 of the study, the RTPM was used to measure the results of the students’

reasoning skills. Therefore, the PSA results are considered not in terms of the

results in terms of effectiveness.

To sum up, in Part 1 the PSA is analyzed at the discourse, knowledge and interactional level. However, in Part 2, each examined group’s scores on the reasoning skills test are measured and compared in terms of the context variable (CLIL vs L1). However, Part 2 also considers another variable: the TT intervention program already presented in section 3.X. In this respect, Part 2 resembles previous studies on the TT program performed in the L1 context in which the RTPM of non‐

verbal reasoning was used to to measure the reasoning skills of two groups of students, native speakers of English (Britain; Mercer et al., 1999) and Spanish (Mexico; Rojas‐Drummond et al., 2003). The studies measured the results by setting up an experimental design with a control group and an experimental group which followed the Talk Reasoning and Computers program (Mercer et al., 1999), later called Thinking Together program (Dawes et al., 2004) designed to improve the quality of classroom talk.

Similarly, in the present study, one class from the CLIL school (CLILA) and another class from the L1 school (L1A) were chosen as the experimental groups. As already explained in section 3.3.3, the Thinking Together program was adapted to the CLIL and L1 contexts in the CAM and the CLILA and L1A teachers were trained in the program. The participating teachers were selected by the school, taking into account availability to follow the training in the TT program and its posterior implementation. The training took place in early February 2015 and the TT program was implemented between mid‐February and mid‐May 2015.

The following section will give a detailed account of the data collection process that took place between September 2014 and June 2015. For a better understanding, a chronological account of each of the stages will be provided.

4.4 Data collection process