CHAPTER 4: RESEARCH DESIGN
4.4 Ensuring reliability and validity
4.4.1 Multi-method approach
The multi-method or multiple methods approach was used under a role-playing mode in the first exploratory study (ES1) which aimed to investigate the perceived potential of GBL through the eyes of trainee teachers. Although the multiple data
92 collection approach was only applied once in ES1, the use of the multiple data
analysis approach extended to the explanatory stage of this research.
ES1 was divided into two sessions, in which all participants played the role of SME in the first session and game designer in the second. Five different methods were deployed to gather four kinds of data and to test their suitability and practicality for other studies (see Table 4.5).
Table 4.5: The types of instruments used and the nature of data collected in the first exploratory study
Instruments/methods Kinds of data Nature of data
Brainstorming Perceived potentials of GBL Qualitative
5W1H Worksheets Self-justifications of the perceived top ranked
potential of GBL
Six Thinking Hats Worksheets Self-evaluations: the evaluation of the self-
justifications
Single-page treatment form Game design treatments
Post-session feedback questionnaire survey
Responses and comments on the conduct of the study and the use of instruments
Quantitative & qualitative
Apart from the six closed questions about the conduct of game design activity in the second feedback form, all collected data were qualitative. Written text was the only form of qualitative data collected in this study. Although it would be common to express game or GBL ideas through graphical or auditory representations, written text was preferred, to match the choice of data analysis techniques. The collected data were sorted using NVivo (Bryman 2008, p. 569). Three analysis techniques were used to analyse the data: narrative text coding, pattern matching and logical model mapping. These techniques were employed in all qualitative data analyses throughout the doctoral research.
The narrative text coding technique was an adaptation of Creswell‘s (2008) qualitative process of data analysis. A narrative is a spoken or written account of connected events (The Oxford Dictionary of English 2005). Written accounts were coded to form themes and to describe either the themes or interesting issues
mentioned by participants.
The second analysis technique was pattern matching (The Oxford Dictionary of English 2004). Pattern is a form of qualitative finding (Patton 2002); which appears
93 in a regular and intelligible form or sequence discernible in certain actions or
situations, especially one on which the prediction of successive or future events may be based (The Oxford Dictionary of English 2009). In the research, the process of identifying, analysing and interpreting the presence of comparable patterns in perceptions was named as pattern matching. This technique has been used in psychological studies since the 1960s (Campbell 1966), but it was Yin (2009) who transferred the concept to case study research in the 1980s, when he claimed it as
‗the most desirable analytic strategy in case study research‘ (Hak & Dul 2009). The third technique used to analyse the qualitative data is called ‗logical model mapping‘. This technique was inspired by the idea of logical model used in computer database design; as Phelan (2003) explained, ‗a logical model is a way to draw your
mental roadmap from a problem specification to an entity-based storage system.‘ In the preliminary literature review of this doctoral research, key concepts and issues related to GBL were identified and connected to construct a working logical model—the GBL model version 1.1 that imitated the idea used in database design (see Figure 4.6). The GBL model juxtaposes the inputs needed in a linear game production process, which could be acquired from academia, the creative industry and a specific GBL context. Academics who study games for use in educational contexts could define learning objectives at the beginning of the production using a generic ADDIE instructional design approach. Instructional design is ‗the systematic
approach to the Analysis, Design, Development, Implementation, and Evaluation [ADDIE] of learning materials and activities‘ (McGriff 2000). After that, experts in the creative industry could create storylines and plots during the game design and development process, in which the core mechanics are set. This would be followed by determining the artistic direction of the game, which involves interface and graphic design. Once the game is produced, SMEs or teachers could develop the instructor‘s guide which depicts how the game could be used to provide or facilitate fun and engaging learning experience.
The GBL model was used in a logical model mapping process to identify the gaps of knowledge and skills in three domains of study: academia, the creative industry and GBL context. The term ‗mapping‘ was adopted from Mathematics, where it means ‗be associated with or link to‘ (The Oxford Dictionary of English 2005). The gaps
94 were identified by comparing the desired roles and responsibilities of subject matter experts (Felicia 2009) and game experts (Rollings & Adams 2003), with the findings drawn from the actual and perceived usual practice of participants.
The use of multiple methods in both data collection and analysis processes was indeed an attempt to enhance the trustworthiness of the overall research outcomes, particularly in the self-justification and the self-evaluation processes. The use of NVivo was another effort made to technically enhance the consistency of the analysis procedure (Bazeley 2003).
Figure 4.6: GBL model version 1.1, the logical model developed based on the preliminary literature review
In terms of external validity, like all other methods employed in exploratory studies, the multi-method approach was not designed to make any statistical generalisation due to the small number of participants involved in these studies (Lincoln 1985). However, analytic generalisations were made based on the conclusions of those studies. Such generalisations were relatively more intuitive, ideographic and
empirical (Lincoln & Guba 1994). The results of these exploratory studies were used to build hypothetical propositions which were tested analytically in both
95 As for internal validity, the development and the use of the logical data mapping aimed to cover most if not all components of the GBL concept.
In terms of ecological validity, trainee teachers and teachers should be able to apply the findings in their teaching because the research outcomes were context-oriented, which means based on the justification and evaluation by the trainees who completed the PGCE programme.