SURVEY AS DESCRIPTIVE QUANTITATIVE DATA

The use of Educational Design Research often entails complex teaching designs where it’s necessary to collect large amounts of data where a single method is rarely sufficient to answer the research question itself (Brinkmann & Tanggaard, 2010; Lehmann-Rommel, 2000). According to Greene (2007), it is the combination of different data sources that contributes to a more complete and comprehensive understanding of the educational design and the participants’ experience of it (Greene, 2007). The integration of the different data types thus provides an improved insight when data and analysis, through an iterative process, inform design elements in the next iteration (Bazeley, 2018). Based on these arguments, quantitative data are collected through questionnaires to supplement the qualitative data. Thus, where the qualitative interviews seek to obtain a rich description of how the students have worked with the learning game in their process, the quantitative data collection has a specific focus on describing the changes over time. The purpose of the quantitative data is thus to gain knowledge about how the students use and experience the game seen throughout the entire period of the project (Kvale & Brinkmann, 2014; Watt Boolsen, 2008).

Within humanities research, two types of quantitative statistical analysis exist ‒ descriptive and inferential ‒ with both aiming to provide answers to questions about social phenomena based on quantitative data collection. Descriptive statistics are about summarising and describing the data set, while inferential statistics through randomised studies aim to make predictions.

In educational research, it can be challenging to talk about randomised controlled studies as teaching situations often contain many types of variables (Agresti & Finlay, 2014; Brinkmann & Tanggaard, 2010; Lehmann-Rommel, 2000;). According to Alan Agresti and Barbara Finlay (2014), one should, therefore, be extremely cautious about using inferential statistics within the humanist paradigm: “Much social science research uses observational studies, which use available subjects to observe variables of interest. One should be cautious in attempting to conduct inferential analyses with data from such studies. Inferential statistical methods require probability samples that incorporate randomisation in some way” (Agresti & Finlay, 2014, p. 24). Therefore, since inferential statistical methods require probability samples that incorporate

Descriptive

Frequency Frequency

Continuous

randomisation in some way, the PhD project is based on a descriptive approach to the quantitative data (see Figure 59). Also, when it is possible to find the actual values of the parameters studied, i.e. the entire population is known, there is no need to use inferential statistical methods (Agresti & Finlay, 2014). The PhD project’s data collection contains the total population of individuals who are of interest in the study, as the significance of the context for the results is vital.

Figure 59 ‒ An overview of the choices made for the quantitative data collection.

The quantitative analysis is based on the third and last iteration. This demarcation is partly because there is no comparable basis between the three iterations as the testing of the game takes place during a different number of weeks (see illustration 58). Further, it is chosen that the results of the questionnaires from both iterations 1 and 2 are not included as separate presentations, as the number of participating students is considered to be so low that it is not meaningfull to conduct a quantitative analysis. The results from iterations 1 and 2 thus only contribute valuable knowledge in order to prepare (1) a question guide for the focus group interviews, and (2) new design perspectives in relation to the design workshops (Kvale & Brinkmann, 2014; Watt Boolsen, 2008).

The students received four identical questionnaires over three weeks in order to see any change within the students’ perception or their game behaviour during the intervention.

Also, the students received a final and evaluative questionnaire at the end of the intervention. The purpose of the final questionnaire is to get a general assessment of the students’ experience of working with Game-Based Learning (Kvale & Brinkmann, 2014; Watt Boolsen, 2008).

Physical questionnaires were used to ensure a high response rate, as it makes it possible to collect all the questionnaires at once (Goodman et al., 2015; Kvale & Brinkmann, 2014; Watt Boolsen, 2008). Also, all students answered the questionnaire at the same time, which can be a challenge with digital questionnaires. Since the quantitative data are supposed to help us see how the student behaviour changes over time, this choice was significant. The quantitative approach thus aims, through four physical and identical questionnaires, to continuously collect data of actions and events related to the application of the gaming principles over a period. At the end of the course, the students were asked to fill out the questionnaires one last time, but this time with a focus on evaluating the entire course. The purpose of the quantitative data collection is to clarify various indicators that are interesting to investigate further from a qualitative perspective. The physical questionnaires also allow for a high degree of anonymity, which is essential because of the insider perspective on which the PhD project rests.

In the same way as in the qualitative data collection, the students were not asked to give their names when filling out the quantitative questionnaires. The students used a self-selected alias in connection with the quantitative survey. In addition, all questionnaires were physically delivered and thus not sent out by mail (Kvale & Brinkmann, 2014;

Watt Boolsen, 2008).

The response rate for each round of quantitative data collection can be seen in the table below (Kvale & Brinkmann, 2014; Watt Boolsen, 2008).

Figure 60 – The response rate for each round of quantitative data collection.

Design of question

The quantitative questionnaire consists of a total of 31 questions (see Appendix C) based on the three main areas, i.e. Motivation & Autonomy, Exploration & Analysis, and Reflective Practice, defined in Section 1.4. The questionnaire is built around questions that investigate the extent to which Game-Based Learning has supported specific elements within the three main categories. The use of a large number of questions is chosen to create many perspectives from the quantitative data that serve to identify

1. collection 2. collection 3. collection 4. collection Final collection

87,3 90,9 83,6 83,6 % 89,1 %