Quantitative and qualitative methods in social science are commonly recognised as two
distinct research methods. The dichotomy between the two methods not only refers to
the techniques applied in each method, but also reflects the two philosophical positions
underlying the two methods (Creswell & Plano-Clark, 2006; Punch, 2005; Spratt,
Walker, & Robinson, 2004). Each of the two methods has its own strengths and neither
method is superior to the other.
Quantitative methods are believed to reflect positivism and post-positivism (Neuman,
2006; Spratt et al., 2004). Positivists believe that there is absolute true knowledge which
can be discovered through scientific methods. This true knowledge is objective and
ultimately measurable. This philosophy of science emphasises evidence and sees social
research which involves reducing ideas into a small, discrete set of ideas to test. Based
on this worldview, all hypotheses and theories must be tested deductively against
observations of the natural world. This approach conceptualizes reality in terms of
variables that comprise hypotheses and research and the relationships between those
variables, emphasising objectivity in data collection and relying on measurement.
Quantitative data enable standardized and objective comparisons and the measurements
provide overall descriptions of phenomena in a systematic and comparable way.
In contrast, qualitative methods are believed to reflect interpretive, naturalistic and
constructivist paradigms (Migiro & Magangi, 2011; Neuman, 2006; Spratt et al., 2004).
These philosophical paradigms comprise the systematic analysis of socially meaningful
action through the direct detailed observation of people in natural settings in order to
arrive at understandings and interpretations of how people create and maintain their
social world. The purpose behind the qualitative methods is to explore a topic or
discover the underlying meanings and patterns of relationships, and inductively
generating hypotheses and theories. Qualitative research is especially effective for
obtaining culturally specific information about the values, opinions, behaviors, and
social contexts of particular populations without involving the use of mathematical
models.
distinctions between the two methods lie in the nature of the data and in the methods
used for collecting and analysing data. They have been considered to be like the two
ends of a continuum, with the third methodology, mixed methods, situated in between.
The mixed methods approach represents a philosophy of pragmatism articulated by
many researchers (James, 1907; Maxcy, 2003; Peirce, 1904/1997). The pragmatists
pursue answers to research problems by utilising any methods available to obtain
knowledge about the problems regardless of the underlying circumstances. This
pragmatic perspective rejects the notion that the use of any single method can
effectively access knowledge; instead, a combination of qualitative and quantitative
methods within a single study works best to understand a particular problem (Migiro &
Magangi, 2011).
The mixed methods approach refers to the mixing of quantitative and qualitative
methods or forms of data in a single study or in multiple studies (Creswell &
Plano-Clark, 2006). This kind of integration is also referred to as the “multi-strategy
approach” in which a qualitative method is used to examine the processual aspect, while
a quantitative method is used to acquire structural features (Punch, 2005). The mixed
methods approach, based on a pragmatic philosophical stance, has been widely adopted
According to Creswell and Plano-Clark (2006), a mixed methods research design
involves four decisions that influence the design choice: (a) a timing decision (whether
the two methods are implemented concurrently or sequentially); (b) a weighting
decision (whether the two methods have equal priority or one has a greater emphasis
than the other); (c) a mixing decision (at what stage the two methods are integrated);
and (d) a theorising decision (the choice of a theoretical perspective that guides the
mixed methods inquiry). Researchers can choose any combination of timing, weighting,
and mixing decisions in their mixed methods design. These decisions, combined with
different research purposes, lead to different design choices such as triangulation,
embedded, explanatory and exploratory mixed methods designs.
Researchers have recognised a number of strengths in the mixed methods approach
which shows the superiority of this design over any single method design (Creswell,
2003; Johnson & Onwuegbuzie, 2004; Morgan, 2007; Tashakkori & Teddlie, 1998).
First, the mixing of methods can provide answers to research questions with a broader
scope than the single method. The mixed methods approach can better serve research
that aims to answer exploratory questions about how a predicted relationship actually
happens. In addition, the combination capitalizes on the strengths of both methods and
compensates for their respective weaknesses (Punch, 2005). Integrating a variety of data
sources and analytical techniques, the mixed methods approach can yield more
methods designs allow for diverse perspectives which lead to greater insight and deeper
understanding of a phenomenon. These advantages of the mixed methods approach may
produce more complete knowledge necessary to inform theory and practice and increase
the generalizability of the results.
However, when constructing mixed methods designs, it can be somewhat difficult for
researchers to decide how to mix the two methods appropriately. One factor that must
be taken into account is the weighting of the two methods. Morse (1991) suggests that
the priority of the methods can be gauged by the theoretical drive, the research purposes
and questions, the use of procedures and the resources for the methods. These practical
considerations, which were carefully assessed in the design of research methods for this
study, may help to decide whether to assign equal weight to both methods or prioritise
one over the other. Another concern involves deciding at what stage to integrate the two
methods, choosing between the stages at which the research questions are conceived, or
at the stages pertaining to sampling, developing instruments, analysing data, or
interpreting findings. Finally, it can be more difficult and time consuming for a single
researcher to carry out mixed methods when they have to be conducted concurrently.
The decision to adopt mixed methods for this study was grounded both in the state of
the art in L2 research on CMC and in the purpose of this research, i.e., to extend a
interactive face-to-face and online discussion tasks. As reviewed in Chapter 2, a number
of studies on students’ attitudes and perceptions have mainly adopted quantitative
methods (Chen, 2005; Skinner & Austin, 1999; Warschauer, 1996a; Yildiz, 2009; Yildiz
& Bichelmeyer, 2003) that have helped to empower researchers in capturing the nature
of psychological constructs by collecting a large sample of data. In quantitative research,
measurement is generally accomplished through statistical methods using scale items.
This research employs a survey questionnaire to explore students’ perceptions of their
learning gains and influential factors in order to complement and validate the qualitative
data.
An increasing number of qualitative studies situated in educational settings are
emerging to examine students’ social presence, interactional patterns, discourse
functions, and critical thinking (Chiu, 2006; Kung, 2004; Liang, 2010; Shin, 2006;
Sotillo, 2000). This emergence may reflect a recognition of the possibility that
quantitative methods alone are inadequate to evaluate the quality of students’ written
language and content, learning processes, and individual in-depth perspectives.
Responses to discrete questionnaire items may not suffice to reveal the complexity of
students’ learning in blended face-to-face and online discussions; in order to understand
the processes of student interaction and meaning construction in such an innovative
The choice of a mixed methods design for this research was guided by “methodological
purposiveness” (Richards & Morse, 2007), which means that the research purposes and
questions were the deciding factors in selecting the most suitable approach. The present
mixed methods approach was equivalent to a QUALquan design (Creswell, 2003) as
shown in Figure 4.2. The qualitative methods were followed by the quantitative
methods and both of them were complete in themselves, with more weight assigned to
the qualitative. The integration of the two methods occurred at the final interpretation
stage. The quantitative methods served to complement, triangulate, and expand on the
qualitative methods; this mixing of methods thus takes advantage of both the in-depth,
contextual nature of qualitative findings and the representativeness and generalizability
of quantitative findings.