5 THE CASE STUDY
5.6 D ATA ANALYSIS
The analysis of a case study is said to be the least developed and the most difficult aspect of doing case studies, however, the analysis strategy developed needs to treat the data fairly to arrive at compelling analytic conclusions and rule out alternative findings (Yin, 1994). I began with the analysis by keeping these in mind. Creswell (2007) notes that the process of data collection, data analysis, and report writing are not distinct steps in the process, rather they are interrelated and happen simultaneously throughout the research process, this was also experienced during this research. The analysis for this research took place in four major steps as shown below.
Stage 1: Analysis of interview transcripts
Stage 2: Comparison of interview 1 and interview 2 Stage 3: Analysis of Observation notes
Stage 4: Comparison of findings from interview and observations
Stage 1: Analysis of interview transcripts
The analysis of the interview transcripts began with the goal of wanting to be able to draw comparisons between the pre-interview and the post-interview.
After the initial steps of transcribing and reading through the transcriptions to familiarize myself with the transcripts, I started looking more closely at the data to plan my approach for analysis. A framework that best explains the steps taken in this stage of analysis is the interactive model for data analysis suggested by Miles and Huberman (1994) shown in FIGURE 10. The interactive model involves three steps that are performed simultaneously after the data collection, they are data reduction, data display, and drawing conclusions. I further explain below how I used these steps in my own data analysis.
FIGURE 10 Interactive model for data analysis (Miles & Huberman, 1994, p. 12)
• Reduction
The reduction phase involved going through the transcription multiple times, removing pieces of data from there, and transferring them into cells on an excel sheet. After this stage, I had all the data from the transcription transferred to an
excel sheet in the form of small pieces of data. An example of the reduction stage is shown in FIGURE 11 and FIGURE 12. At this point, I knew my data very thoroughly. As Miles and Huberman (1994) note, data reduction does not stop here, the reduction is part of the analysis process and the process continues until the end of the research.
FIGURE 11 Interview transcript
FIGURE 12 Reduced data
• Visual display of data
Tracy (2013) suggests the use of visual data displays in the form of a table, matrix, network, or a flowchart is not only a way of making sense of large amounts of qualitative data but it is also as a useful layer of analysing and thinking creatively about the data. Data displays organize data in a systematic format so that the researcher can make valid conclusions and perform necessary actions (Miles &
Huberman, 1994). In this research, I used concept maps as a way of displaying the interview data. Using concept maps to display data allows for visual identification of themes and patterns and in turn facilitate the process of comparison (Daley, 2004). Concept maps not only helped me with reducing and displaying data but also in the analysis and comparison process. Concept maps of the transcriptions were drawn using a software tool called Cmap tools. The
symbols used in the maps and their meanings are represented in TABLE 5. An example of a concept map is shown in FIGURE 13.
TABLE 5 Concept map symbols
Symbols Meaning
Represents the concept Links two concepts together
Proposition: A logical connection of 2 concepts
Cluster: A group of propositions
• Identification of clusters and assigning general codes to the clusters
After the concept maps were drawn, I identified clusters that represent the same concept. General codes were assigned to these clusters. The general codes are representative of what the cluster was about. The list of codes is represented in TABLE 6 and some examples of its respective clusters used for the comparative analysis are represented in TABLE 7. The reason for this approach to data analysis is that I wanted to interpret the data holistically without having to take apart the data too much by assigning too many codes.
FIGURE 13 Concept map and clusters
TABLE 6 General codes General codes
Descriptions of PBSL Materials Assessments
Group work Talk about past experience Goals
Difficulties Teacher’s role in PBSL Reference to Finnish curriculum
Scientific method self-guided learning Science learning through projects
Stage 2: Comparison of interview 1 and interview 2
The comparison of interview transcripts began by first creating a word document by name ‘comparative document’. A table was created like the one presented in TABLE 7. Clusters with the same general codes from both the interviews were placed next to each other for comparison. The similarities or differences in the way the teacher spoke about each of the codes was noted and a detailed description of the comparison was written down.
TABLE 7 Example of coding and comparisons
Codes Interview 1 Interview 2
Materials
Stage 3: Analysis of Observation notes
The observation notes were analysed using the critical incident analysis technique. This method is used for many different purposes. Flanagan (1954) used critical incidents (CI) to study activity requirements in different professions.
Thomson and Hall (2017) used this method to study schools. Tripp (2012) used CI to help teachers with reflection and learning. The critical incident analysis technique is flexible and can be adapted in many situations (Flanagan, 1954). This section explains what is meant by a CI, the purpose of using CI for this research, and how this technique was used in this research.
Powell et al. (2003) note that the events that are considered critical differs based on the research questions (RQ) pursued. This research seeks to understand how the teacher’s conceptions relate to the implementation and how the teacher developed her conceptions of PBSL as a result of practical experience. Dilemmas, as seen earlier, stop teachers from implementing PBSL the way it was conceptualized, but they also act as learning opportunities for the teacher. These dilemmas reveal important insights that help in answering the RQs. Therefore, any incident that caused the teacher to encounter a dilemma is considered a CI.
CI technique was best suited for analysing the observation notes for two reasons.
One, because of the type of data that was collected. The observation notes had recorded sequentially and in detail, the events that occurred in the class. Due to the availability of rich descriptions of the events, the data benefited most by identifying and analysing the CIs. Two, the implementation of PBSL was looked at from Windschitl’s (2002) framework of dilemmas and therefore, rich interpretations could be drawn by analysing the dilemmas encountered. The analysis of the observation notes followed four steps. They are: 1) identifying the CIs 2) describing the CIs 3) forming categories 4) drawing interpretations. These steps are described below in detail.
• Identifying the CIs
As mentioned above CIs in this research represents the dilemmas encountered by the teacher during the project implementation. Some of these CIs were already identified during the observations and others were identified through multiple read-throughs of the notes. The observation data had a total of 10 CIs from the WST project and 25 CIs from the Electricity project. Each CI was given a number based on the sequence of its occurrence.
• Describing the CI
After examining each CI closely, a detailed reflection of the critical incidents was produced. The reflection included answers to questions such as What happened during the CI? How did this CI come to be? How did this incident impact the project? What could have been done differently? Discussions with the teacher played an important role in forming my understanding of the CIs.
• Forming categories
CIs that were related to a similar aspect/topic were categorized together. TABLE 8 provides a list of all categories along with an example of a CI under each category for the reader to have a glimpse of what the CIs under the categories
looked like. An example of how one category named ‘use of teaching materials’
was formed is shown in TABLE 9.
TABLE 8 Categories
Category CI Description
Students readiness CI 3 Students forget to follow the steps of the investigation
Time CI 1 Setting up the class- teacher preparing for the project
Group work CI 16 Students experiment with the equipment in the circuit construction kit in their groups
Unexpected results CI 5 Paper clip sinks Connecting science topics
to students’ real-life
CI 10 Teacher and students discuss the role of WST when washing clothes
Self-guided learning CI 33 Student self-guided learning – testing if paper conducts electricity
Equipment in the school CI 12 Printer does not work
Assessment in projects CI 32 Teacher’s concern about student learning – are they learning anything?
Teacher’s content knowledge
CI 22 Teacher unable to answer students question on current
Use of teaching materials CI 24 Students find it difficult to follow instructions on the booklet
Student absence CI 9 Teacher spends time explaining the experiments again to students that were absent
Teacher learning CI 8 Teacher change in instruction
TABLE 9 Formation of categories
Week Number Description Category
Week 1 CI 4 Getting the students to write in the investigation form
Use of teaching materials Week 4 CI 20 Teacher ensured students note down
experiments in their notebooks Week 6 CI 24 Students find it difficult to follow
instructions on the booklet
Week 6 CI 25 Students filled booklets are being collected Week 6 CI 26 Students copy answers from each other Week 7 CI 28 Ensuring students fill the booklet
Week 7 CI 31 Answers arrived at do not match answers in the booklet.
• Drawing interpretations
In the last step, each category was looked at closely to interpret what the dilemmas involved tell about the project implementation. As mentioned before, Windschitl’s (2002) framework of dilemmas was used as a frame of reference for the dilemmas identified in the observations. Therefore, the CIs under each category were seen in the light of this framework to understand what kind of dilemmas are present in the categories. For example, TABLE 9 represents the CI under the category named ‘use of teaching materials’. This category reveals important insights into the use of teaching materials throughout the project.
Due to the limitation of space, a detailed description of the CIs is not provided in this thesis, however, these CIs will be referred to in the Findings section and a general description of them will be provided where necessary. How these critical incidents were used for analysis and interpretation is shown further with one example in section 6.3.
Stage 4: Comparing interview findings and observation notes
In the final step of the analysis, the findings from the interviews were compared with the findings from the observation. This was done by looking at how the general codes produced in the interview findings compared with the categories formed from the observation notes analysis. As an example, TABLE 7 and TABLE 9 can be compared with each other to provide important insights into the way materials were used in the projects. Further explanation about this comparison is provided in section 6.6.