Article Preview
TopIntroduction
Systematic reviews continue to show that the quantitative research approach is the most frequently used technique in the world (Limaye, Limaye, Krause & Fortwengel, 2017; Hanlon et al., 2018). Quantitative studies we encounter in the literature in daily research activities enormously outnumber qualitative ones. As academics, we have seen most student researchers under our supervision opt for the quantitative design in their research work on the grounds that it was the most suitable approach for addressing their research problems. A few college students are also honest to link their choice of the quantitative approach to their interest and capabilities.
A large number of studies (Guo et al., 2017; Özer et al., 2017; Yağar & Dökme, 2017; Sibiya, Ngxongo & Beepat, 2018; Chappell et al., 2019; Cuéllar-Molina, García-Cabrera & Déniz-Déniz, 2019; Herrett et al., 2019; Sendoval et al., 2019; Medhekar, Wong & Hall, 2019) have used criteria-based selection (i.e. using a set of relevant criteria to select participants) to determine their accessible population or sample. In some other quantitative studies, researchers were constrained by research conditions to use an available population or predetermined sample, which researchers consider a ‘convenient’ population or sample (Chappell et al., 2019; Herrett et al., 2019). In fact, a careful analysis of top-tier quantitative papers (Chappell et al., 2019; Cuéllar-Molina et al., 2019; Herrett et al., 2019; Sendoval et al., 2019) revealed to us that quantitative researchers frequently use at least five different non-probability sampling methods. Obviously, the use of non-probability sampling methods in quantitative studies is a growing norm. Considering the fine reputation of many of the studies championing this tradition, it can be said that using non-probability sampling in quantitative designs is acceptable and unavoidable. We have nonetheless observed that the non-probability techniques used by quantitative researchers are either not well documented or portrayed as absolute probability sampling methods. We would want to ascribe this problem to non-availability of a formal typology for these non-probability selection procedures in the literature. That is, there is no acceptable standard for documenting these methods. The primary goal of this study is thus to develop a typology of these techniques to enable researchers to scientifically or plausibly document them, guide journal editors and reviewers to assess manuscripts and provide a standard for research critique.
A basic goal in quantitative research is to generalise sample statistics to the general population. The idea that qualitative studies can also generalise findings from a sample to the population is undeniable, but the quantitative design is renowned for applying more rigorous procedures that make its findings more generalizable (Krejcie & Morgan, 1970; Creswell, 2003). Central to these procedures is determining and using a representative sample, which is a sample that is large enough to give rise to findings that the population would have produced if entirely surveyed (Black, 2010; Saunders, Lewis & Thornhill, 2012). Sampling theory suggests that representativeness of the sample is necessary if findings are to be inferred to the population (Krejcie & Morgan, 1970; Williams, 2007). A sampling frame is required to determine a representative sample (Asiamah et al., 2017b; Krejcie & Morgan, 1970; Williams, 2007), but this list is often not available when the general population is very large. Similarly, the researcher needs a list that details characteristics of members of the general population to be able to determine the target and accessible populations.