Research Methodology

Sampling design involves five major steps: (i) defining the target population: (ii) selection of sampling frame; (iii) determinations of sampling methods; (iv) determination of sample size; and (v) selection of sampling unit and survey administration (Zikmund, 2003, Maholtra, 1999, David, 2005). Each of these is discussed in more detail below. The first step of any sampling process is to define a target population from within the wider population of potential subjects in accordance with design the relevant parameters of a study and the potential generalization of any finding of the research (Maholtra, 1999, Neuman, 2016b) (Zikmund, 2003, Maholtra, 1999). Given that the primary objective of the present study was to identify the factors that determine the choice of a medical destination by prospective medical tourists, the first criterion for inclusion in the target population was people who have an interest in medical tourism.Any samplingframe is subject to potential error by over-representing the actual population within the given parameters (Davis, 2005, Zikmund, 2003). The selection of a sampling frame is thus crucial for any quantitative research project because a mismatch of the sampling frame and the actual population could result in distorted findings (Zikmund, 2003).Sampling methods can be broadly divided into two types: (i) probability sampling (in which members of the population have a known chance of being included in the sample); and (ii) non-probability sampling (in which members of the population do not have a known chance of being selected (Zikmund, 2003).In determining the sample size, the three key considerations are: (i) the degree of accuracy; (ii) the degree of diversity of the population; and (iii) the number of variables to be examines (Neuman, 2006b) in most cases these factors are positively correlated with sample size. However, as Maholtra (1999) has noted, this does not necessarily mean that huge sample sizes are routinely required; indeed, different purposes require different numbers of respondents.The term survey administration refers to data collection using a survey instrument to obtain information pertinent to the research question from a large number of respondents (Davis, 2005) (De Vaus, 1995). When choosing a survey-administration method, five issues must be considered; (i) obtaining an appropriate response rate; (ii) obtaining a representative sample; (iii) design of a questionnaire; (iv) quality of responses; and (v) implementation problems.Four methods of survey administration are commonly applied: (i) personal interview; (ii) telephone survey; (iii) postal-mail survey; and (iv) Internet survey.

Sampling design involves five major steps: (i) defining the target population: (ii) selection of sampling frame; (iii) determinations of sampling methods; (iv) determination of sample size; and (v) selection of sampling unit and survey administration (Zikmund, 2003, Maholtra, 1999, David, 2005). Each of these is discussed in more detail below. The first step of any sampling process is to define a target population from within the wider population of potential subjects in accordance with design the relevant parameters of a study and the potential generalization of any finding of the research (Maholtra, 1999, Neuman, 2016b) (Zikmund, 2003, Maholtra, 1999). Given that the primary objective of the present study was to identify the factors that determine the choice of a medical destination by prospective medical tourists, the first criterion for inclusion in the target population was people who have an interest in medical tourism. Any samplingframe is subject to potential error by over-representing the actual population within the given parameters (Davis, 2005, Zikmund, 2003). The selection of a sampling frame is thus crucial for any quantitative research project because a mismatch of the sampling frame and the actual population could result in distorted findings (Zikmund, 2003). Sampling methods can be broadly divided into two types: (i) probability sampling (in which members of the population have a known chance of being included in the sample); and (ii) non-probability sampling (in which members of the population do not have a known chance of being selected (Zikmund, 2003). In determining the sample size, the three key considerations are: (i) the degree of accuracy; (ii) the degree of diversity of the population; and (iii) the number of variables to be examines (Neuman, 2006b) in most cases these factors are positively correlated with sample size. However, as Maholtra (1999) has noted, this does not necessarily mean that huge sample sizes are routinely required; indeed, different purposes require different numbers of respondents. The term survey administration refers to data collection using a survey instrument to obtain information pertinent to the research question from a large number of respondents (Davis, 2005) (De Vaus, 1995). When choosing a survey-administration method, five issues must be considered; (i) obtaining an appropriate response rate; (ii) obtaining a representative sample; (iii) design of a questionnaire; (iv) quality of responses; and (v) implementation problems. Four methods of survey administration are commonly applied: (i) personal interview; (ii) telephone survey; (iii) postal-mail survey; and (iv) Internet survey.

Data preparation involved five activities: (i) checking questionnaire responses and excluding unsatisfactory surveys; (ii) editing of partially incomplete questionnaires; (iii) coding of data; (iv) cleaning of data; and (v) statistical adjustment (Maholtra, 1999). With regard to the first of these, some returned questionnaires were excluded from the sample after data collection had been concluded. These included: (i) incomplete questionnaires with insufficient data for further analysis: and (ii) disqualified questionnaires from respondents who chose option that would consider travelling abroad for medical reasons). With regard to editing the questionnaires, those with partially incomplete data were assigned missing values (as appropriate) and checked for consistency of responses. Inconsistent surveys were also disqualified and excluded from the sample (Maholtra, 1999). Coding then proceeded by assigning values to the responses reported by respondents. These were then recorded in accordance with Microsoft Excel. After coding, each of variables was checked for a normal distribution. Outliers for nominal variables were detected using boxplot, and cases outside the three-box range were excluded from the analysis for that particular nominal variable (Manning and Munro, 2007). Multivariate outliers for interval and ratio variables were by calculating Mahalanobis distance; these fixed as appropriate (Manning and Munro, 2007). It was decided that the statistical techniques to be adopted in this study would include: (i) Pearson’s correlation coefficient; (ii) analysis of variance; and (iii) principal component analysis. Variables of interval measurement that satisfied an assumption of normality were tested with parametric statistical techniques-including Pearson product moment correlations, analysis of variance, and multiple correlation coefficients. Variables of interval measurement that failed to satisfy an assumption of normality were treated with non-parametric statistical techniques-including Spearman’s rank order correlation the Krush-wallis test, and the Mann-Whitney U test.

As a form of social science research, business research deals with human respondents who must be treated with respect, justice, integrity, and beneficence. When potential respondents are asked to participate in such research, researchers must be mindful of the fact that they are actually seeking permission to intrude on the privacy of these people, and obtaining informed consent to participate its therefore crucial. In this regard, Bouma (2000) proposed five ethical principles that should guide researchers in conducting a study such as the one reported in this dissertation.