CHAPTER THREE RESEARCH METHODOLOGY
3.8 Validity and reliability
According to Golafshani (2003: 598), validity refers to the extent to which the findings represent what is happening in the real situation being measured. Borsboom, Mellenbergh and Van Heerden (2004: 1066) add that validity is concerned with whether the data and findings give a true picture of what is being studied and accurately assess what it intends to assess.
To ensure construct validity, the structured questionnaire was designed to capture all the three variables being analysed: staff interpersonal communication, staff interaction
and staff cohesion. The size and characteristics of subject population, the sample selection process, the time given for completion of questionnaires and the data collection adequately ensured that this entire research process and the findings obtained would meet internal and external validity required for a scientific research method. Furthermore, the questionnaire instrument was checked and verified by a registered statistician (ensuring that the questions link to the research problem, research questions and objectives) before sending to research participants. Validity was further verified in the data analysis through significance statistical proof in factor analysis. The Principal Component Factor Analysis with Varimax rotation (Bro &
Smilde, 2014: 2815), the results of the Kaiser-Meyer-Olkin (KMO) test for sample adequacy, the Bartlett test of Sphericity and the t-statistic significance test all indicated that the sample and method of analysis was appropriate.
In scientific studies, reliability relates to the degree to which research tools produce a stable and consistent result (Roberts, 2006: 43). Roberts adds that research findings can only be deemed reliable if another researcher can duplicate the tests and derive the same results. To test the credibility of the research findings in this study, Cronbach’s alpha was applied to test if the questions in the Likert scale survey instrument were reliable and correctly grouped. Cronbach’s alpha analysed the participants’ conscientiousness and openness, which are unobservable latent variables. Cronbach’s alpha further measured if the survey instrument was designed accurately to measure the variable of interest (Peterson, 2013: 384).
Reliability and validity were further tested through a pre-test questionnaire pilot study aimed at identifying any shortcomings to be certain that the questionnaire was properly understood as intended. The pre-test questionnaire was distributed via email to 10 faculty staff members and their feedback requested. The majority of the participants in the pilot study noted the arrangement of questions and the length of questions as problems with the survey questionnaire. Feedback from the pilot study also pointed out sections of the questionnaire that the respondents were hesitant to answer, or where they had made mistakes; and three overlapping questions were rephrased. Thus, this feedback was used to improve the survey questions and their logical flow so that the questions would be clearly understood.
3.9 Summary
The main theme of this chapter is an outline of how variables to be measured were empirically tested. This chapter began by outlining the purpose of this study: to
investigate the use of electronic communication and its impact on communication among faculty staff within the Faculty of Business and Management Sciences at UoT X. An interpretive phenomenological quantitative approach to research was applied in a case study with the aim of understanding participants’ experience from their point of view. The Faculty of Business and Management Sciences was selected as it is the largest faculty within UoT X, comprising the largest staff complement and student numbers; and it operates on five different, geographically dispersed campuses. It is assumed that, owing to staff being distant, email communications would be much relied upon to disseminate information to staff across the five campuses.
From the Faculty of Business and Management Sciences, a purposive sampling method was used to select a sample of 438 participants to which questionnaires were administered. The research participants were given two weeks (14 working days) to complete the quantitative self-administered questionnaire distributed to staff within the Faculty of Business and Management Sciences at UoT X and 46 completed questionnaires were returned via scanned email or hand delivery.
The 25-question survey questionnaire was designed to determine the perception of staff members regarding the use of email and the effect it has on staff interaction, social cohesion and interpersonal communication. To ensure that the study adheres to acceptable ethical standards, the researcher obtained clearance and approval from the Executive Director Human Capital, ethical clearance from the Faculty of Business and Management Sciences Research and Ethics Committee, and an informed participant consent form was attached to each questionnaire to explain to participants the purpose of the research study.
Data analysis was conducted by means of statistical techniques to investigate the variables (staff interaction, social cohesion and interpersonal communication), as well as their effect, relationships and patterns of involvement within the faculty staff.
Descriptive statistics was used to summarise the quantitative data, while inferential statistics was used to interpret the research findings. Further statistical analysis using the Principal Component Factor Analysis was conducted to analyse and understand participant’s perceptions of the effect of email communication on their interpersonal relationships, staff cohesion and staff interaction.
The chapter concludes by discussing the issues of validity and reliability to ensure that the study satisfies the expectations of scientific research methods. For construct validity, a registered statistician reviewed the structured questionnaire and the research
sample attributes to ensure that all three variables being analysed (staff interpersonal communication, staff interaction and staff cohesion) were adequately captured. Validity was further verified during the data analysis through employing significance statistical proof in factor analysis, the Kaiser-Meyer-Olkin (KMO) test, the Bartlett test of Sphericity and the t-statistical significance test. These indicated that sample size and method of analysis was appropriate and adequate. Cronbach’s alpha and a pre-test questionnaire pilot study ensured reliability of the questionnaire.
The next chapter will present the findings obtained from a number of statistical tests carried out on the primary data collected from the survey questionnaires to establish empirical results.