CHAPTER FOUR RESEARCH METHODOLOGY AND DESIGN
4.11 Data Analysis
The collection and analysis of data was undertaken in a number of phases, the first phase consisting of the trend analysis of the quantitative results of the NSS and qualitative analysis of the verbatim comments from the NSS. This was followed by the analysis of the qualitative data gathered from the semi-structured interviews. The results of the analysis from both phases are then triangulated with the other identified sources including the critical review of the literature, the review of the documentation
and the direct observation during the interpretation and discussion of results. For this study the findings from the analysis of the data was used to build the case study and the different sources of data to validate the overall conclusions of the case study.
4.11.1 Use of Nvivo 10
Nvivo 10 is a software package that is used to assist researchers to manage, organise and analyse qualitative data. It has many features to assist with the analysis of large quantities of qualitative data although the software does not perform any of the analysis. The nature of this research project with a large volume of qualitative data from seven years of NSS verbatim comments, the semi-structured interviews and relevant literature it is necessary to organise the data effectively to allow the data to be analysed. Nvivo 10 was used within this research project in the following ways;
• Literature in the form of journal papers, policy documents, book chapters and additional reference material was stored in Nvivo 10 to allow for cross-referencing of identified themes across all sources of data.
• The verbatim comments from the NSS were imported into Nvivo.
• The digital voice recordings of the semi-structured interviews were imported into Nvivo along with the transcription of the interviews.
• The project for this research project was created with a suitable node structure to manage the information and to assist with cross-referencing.
• The saved information was coded following the steps outlined by Smith and Osborn (2008). The analysis initially followed pre-established nodes based on the categories of questions in the NSS survey and to reflect ‘positive’ or ‘negative’ views from the students. New nodes were generated as the data collection and analysis progressed.
• The identified themes and concepts were analysed further using the software to highlight any relationships. These relationships were explored further to establish links between the research data and to compare and contract with the literature. • The Modelling tools within the software were used to graphically represent the
4.11.2 Analysis of the published NSS results.
The analysis of the available data has been undertaken in several stages including a general trend analysis of the quantitative data published by IPSOS MORI relating to the University of Salford and the specific programmes offered by the School of the Built Environment. The data was gathered over a period of seven years to ensure a sufficient period of time to be confident of identifying if any trends are present and the nature of any identified trends.
The aim of the analysis of this data is to discover any patterns, concepts, themes and meanings. A qualitative computer software package, Nvivo 10, was used to organise data into manageable nodes that according to Richards (1999), helps to manage and synthesise themes from large amounts of qualitative data. A detailed analysis of the verbatim comments from the NSS for Built Environment programmes was undertaken using Nvivo 10 to identify the key themes relating to the areas covered by the survey. According to Leedy and Ormord (2001), content analysis is used to establish the presence of certain words or phrases within a wide range of texts while Krippendorff (2004) describes content analysis as a “research technique to make replicable and valid inferences from text to a context of their use”. The detailed verbatim comments were entered into Nvivo 10 and a process of coding the information was undertaken to reflect the nature of the comments into positive or negative comments and how they relate to the questions on the NSS in terms of the question category for example an comment could be coded as relating to ‘assessment’ and could also be coded as a ‘positive’ comment. Coding of the data can be undertaken using inductive and deductive coding (Krippendorf, 2004; Bernard, 2000; Marying, 2000) methods depending upon the source of the data. Typically, the literature is coded deductively and the primary data coded inductively. This information was then analysed separately in terms of the comment in terms of the positive perception and how that relates to the overall student experience and separately in terms of how assessment impacted on the student perception of their experience. Further analysis was then undertaken to explore the relationships between the two factors and to explore if any further links could be identified. This has been undertaken to explore the reasons behind the positive or negative perceptions of the student experience as reported by the quantitative data. The data was further integrated
using tools within the software to establish the word frequency within the verbatim comments in any given area or within the entirety of the verbatim comments. The basis of this analysis was to identify any common themes that could be further explored within the semi-structured interviews (Jackson and Trochim, 2002; Silverman, 2001).
4.11.3 Semi-Structured Interview Data Analysis
The semi-structured interviews were recorded using a digital media recorder, with the consent of each participant, with an average duration of 45 minutes. According to Saunders et al., (2012) there is a need "to create a full record of the interview, including contextual data” as soon after its occurrence as possible to control bias and to produce reliable data for analysis. The interview data was initially analysed using content analysis to assist with the organization of the data into general themes. A key aspect of the analysis of the data collected using the semi-structured interview is the ability to search for meaning through the direct interpretation of what is being observed by the researcher as well as what is experienced and reported by the participants. When using case study as the research method, Yin (2003) stresses the importance of checking the data for patterns that may explain or identify causal connection in the database. The process of data analysis begins with the open coding of the data, which is the organisation and categorization of data in search of patterns, themes and meaning that emerges from the data. Dey (1993) and Yin (2009) describe the process of generating categories and reorganizing data as the beginning of the process of engaging with the data and the commencement of the analysis. Assigning the data into categories assists the researcher in making an initial identification of any emerging patterns. This is followed by a comparison of any identified patterns and any contrasts between patterns in order to reflect on any emerging complex threads in the data in order to make sense of them.
The data collected as a result of the semi-structured interviews with students was initially analysed using content analysis in Nvivo 10 to organise the data into general themes. Open coding of the data was used to categorise the data into the identified themes and to include any emerging themes not yet identified. Coding is the process of recording the responses a particular respondent gave to a question in terms of the category
established by the researcher using a tree node and free nodes. Axial coding was then used to look for any relationships between the identified categories of data. This process is to “explore and explain a phenomenon (a subject of your research project) by identifying what is happening and why, the environmental factors that affect this (such as economic, technological, political, legal, social and cultural ones) how it is being managed within the context being examined, and the outcome of action that has been taken” (Saunders et al., 2009). Axial coding of the data was undertaken to identify any emerging relationships between the themes. The researcher can then attempt to verify the outcomes against the actual data in order for a process of testing these relationships. The outcome of the data analysis from all identified sources will be triangulated to produce evidence that can be used to draft a final conceptual framework for the increase in reported levels of student satisfaction within the School of Built Environment.
The final phase of the process is to ensure credibility of the findings. There are three ways in which to do this; the first is by validation which is generally used for studies that take on a more deductive approach, reliability is generally used for studies that take on a more inductive approach, however, where mixed data collection techniques have been employed, triangulation can be a valuable way of ensuring validity and that the data are telling you what you think they are telling you (Saunders et al., 2009). Denzin (1978) defines triangulation as ‘the combination of methodologies in the study of the same phenomena’. This is a method used by qualitative researchers to check and establish validity in their studies by analysing a research question from multiple perspectives (Guion et al., 2011). The data collected from the literature, the analysis of the NSS results plus the semi-structured interviews will be triangulated by mapping across both the qualitative and quantitative findings. This in turn will assist with refinement of the framework.