Research Design and Methodology - Essay Example

Research Design and Methodology 4.1 Categorising Different Types of Research According to Collis and Hussey (2003), the variety of different types of research can be categorised along three general lines: According to the logic, or theoretical foundation of the research; according to the purpose of the research; and according to the research approach. 4.2.1 Differences between Deductive and Inductive Research Although the degree may vary, all research projects invariably rely on the use of theory, which can be described as a set of rationally connected concepts, definitions and propositions that together help to explain and predict phenomena or facts (Blumberg, et al., 2005). A theory can be made explicit in the design of the research; whether it is or not, it will almost always be explicitly presented in the findings and conclusions of the research (Saunders, et al., 2007: 117). Accordingly, research can be divided into two main types, deductive and inductive research (Blumberg, et al., 2005; Collis & Hussey, 2003; Saunders, et al., 2007). The deductive research approach is one in which a conceptual and theoretical structure is developed first, and then tested by empirical observation (Collis & Hussey, 2003; Saunders, et al., 2007). It is called ‘deductive’ because specific cases or phenomena in the research testing are deduced from general inferences (Collis & Hussey, 2003). The inductive research approach, by comparison, works in the opposite direction; data is collected, and the resulting data analysis leads to the development of a theory (Saunders, et al., 2007). In other words, the particular instances under research observation are empirically defined, and from this general inferences are induced to produce theory (Collis & Hussey, 2003). 4.2.2 Purposes of Research The reasons for doing research can are almost limitless, but three of the most common and useful purposes for research are exploration, description and explanation (Babbie, 2004; Saunders, et al., 2007). 4.2.2.1 Exploratory Research Exploratory research is conducted when the research problem or phenomenon confronting the researcher is a new topic, for which there is little to no earlier research that the researcher can draw upon for background information (Collis & Hussey, 2003: 10). Although exploratory research usually does not provide conclusive answers to research problems or develop strong theory, it introduces new topics and insights and provides direction for future research on the subject (Robson, 1993; Collis & Hussey, 2003: 11; Babbie, 2004). 4.2.2.2 Descriptive Research Research that seeks to describe phenomena as they exist is called descriptive research (Collis & Hussey, 2003). The objective of descriptive research is either to simply portray an accurate profile of persons, events or situations, as described by Robson (1993: 4), or it may involve an in-depth analysis of the interaction of two or more variables (Blumberg, et al., 2005). Descriptive research goes further than exploratory research because of the narrowness of its focus; it is conducted to describe in fine detail the characteristics of a single, or in some cases, a limited number of very similar issues, and as such may not have applicability outside its scope (Collis & Hussey, 2003: 11). One significant shortcoming of descriptive research, however, is that it cannot explain the reasons why particular variables interact in a particular way, or why a certain phenomenon has occurred ( de Vaus, 2002a; Blumberg, et al., 2005). 4.2.2.3 Explanatory Research Research that does establish causal relationships amongst variables (unlike descriptive research) is explanatory research (Saunders, et al., 2007). In simple terms, explanatory research explains why or how a particular phenomenon occurs; in this sense, explanatory research can be considered an extension of descriptive research (Collis & Hussey, 2003: 11); descriptive research provides accurate, detailed observation of phenomena, whereas explanatory research goes several steps farther to apply theory or hypothetical concepts to describe the causes behind what is being observed The researcher goes beyond merely describing the characteristics, to analysing and explaining the reasons for the phenomenon that the descriptive research has only observed (Collis & Hussey, 2003; Blumberg, et al., 2005: 10). 4.2.3 Research Approach The third basic way to categorise research is according to the approach adopted by the researcher (Collis & Hussey, 2003). There are three familiar research approaches: Quantitative, Qualitative, or Mixed (Creswell, 2003). 4.2.3.1 Quantitative Research Quantitative research requires the collection of empirically measurable data about the research subject (Hancock, 2002). Quantitative research is considered to be objective in nature because it draws conclusions from the analysis of numerical data, most often through the application of statistical tests (Collis & Hussey, 2003: 13). Quantitative research provides factual, descriptive information – answering questions such as how many, how often, and how much – according to tests that derive their credibility from being easily reviewed and repeated by other researchers (de Vaus, 2002a). 4.2.3.2 Qualitative Research Qualitative research is defined as the direct observation of social phenomena in natural settings, ‘natural’ in this sense meaning, ‘not controlled’ (Babbie, 2004; Creswell, 2003). The rationale for qualitative research is that direct observation of a social phenomenon in its natural context leads to a more complete understanding of it (Babbie, 2004). In contrast to quantitative research, qualitative research can be emergent; new ideas and directions for this research can emerge during the study, leading to refinements of the research questions (Creswell, 2003). Qualitative research is sometimes criticised, however, because it relies on some degree of subjective interpretation of the researcher, and usually does not have results that can be generalised, or replicated by other researchers, all of which can put the credibility of its results in question (de Vaus, 2002a). 4.2.3.3 Mixed-Method Approach Babbie (2004) argues that social researchers who limit themselves to a single approach may limit their ability to completely understand the subject of their research. The main rationale for an approach that adopts mixed methods is that the biases and limitations which are unavoidably a part of any single method can be reduced or eliminated entirely by combining the methods (Creswell, 2003; Saunders, et al., 2007). Saunders, et al. (2007) define mixed methods as an approach in which both quantitative and qualitative techniques of data collection and analysis are used, either in parallel (i.e., at the same time) or in sequence. The mixed-method approach is justified by pragmatic knowledge claims (Creswell, 2003; Saunders, et al., 2007); in other words, according to Creswell (2003), methods, techniques and procedures of research are chosen according to their relevance to the research subject and objectives, without being subjected to a predetermined philosophy or conception of reality. 4.3 Research Strategies In this section, some of the different research strategies that can be used by researchers are described. 4.3.1 Case Study A case study is a kind of descriptive or explanatory research, and is defined by Robson (1993: 40) as “...the development of detailed, intensive knowledge about a single ‘case’, or a small number of related ‘cases’.” The research approach used for collecting and analysing information may be any of the three approaches described in the previous section (Yin, 1994; Hancock, 2002). The main objective of a case study is not necessarily to compare cases, but to develop a complete understanding of a particular case and its attributes within its own context (de Vaus, 2002a). 4.3.2 Ethnography Ethnography is a research strategy that applies to social research, and involves studying a social or cultural group in its natural setting over a prolonged period of time in order to describe and explain the group’s behaviour (Hancock, 2002; Creswell, 2003). As a form of qualitative research, Ethnography requires the collection of in-depth information through the personal interaction of the researcher with the research subjects over a period of time, and thus relies mainly on a qualitative approach (Hancock, 2002: 9). The main objective of data analysis in ethnography is to describe and explain the group from its own perspective (Ibid.); in other words, interpreting the social environment of the research subjects in the same way they interpret it, and describing that perspective to others (Collis & Hussey, 2003; Saunders, et al., 2007). 4.3.3 Grounded Theory A grounded theory research strategy is described by Hussey and Hussey (1997) as an inductive/deductive approach, on that is ‘grounded’ in established theory (the inductive aspect), but which also attempts to develop new knowledge and new theories about the topic being investigated (Hancock, 2002). It is generally applied to social research, and attempts to develop new theoretical concepts by constantly comparing emerging observations to the existing theoretical foundation (Babbie, 2004). 4.3.4 Action Research Action research is another strategy applied to social research, one which assumes that the researchers themselves have an effect on the continuously-changing nature of the social world (Collis & Hussey, 2003: 67). It is usually described as being ‘cyclic’ in nature, because it requires a problem-solving process that may be repeated several times in the course of the research: Reflecting on the current situation, identifying potential problems and solutions, implementing changes, and then evaluating the effects of those changes (Hancock, 2002; Collis & Hussey, 2003). One area in which action research may be applied is in business and management studies. 4.3.5 Experimentation An experiment is a controlled investigation that attempts to observe and identify a cause-and-effect relationship amongst two or more variables, with the objective being the prediction of the outcomes of tests (Hancock, 2002: 7). Classic experimental design involves the manipulation of the natural progression of events with respect to the independent variable or variables (Sekaran, 1992: 117); this tests both the effect of the independent variable, as well as the effect of various interventions once all other factors that might have an impact on the result are identified and controlled (Creswell, 2003). 4.3.6 Survey A survey strategy is used to create a quantitative description of characteristics such as demographic indicators, attitudes, or actions of a given population, which is defined as some group of individuals who share certain attributes (Hancock, 2002: 8; Creswell, 2003). The most common methods for collecting survey data are through face-to-face interviews, interviews by some means of communication such as by telephone or the internet, and questionnaires provided to the research participants and returned when completed (Hancock, 2002; Mathers, et al., 2002; Collis & Hussey, 2003). Surveys can either be descriptive, counting the frequency of a specific population, or analytical, in which data on two or more variables are collected, with the objective being to identify and explain any relationship amongst the variables (Collis & Hussey, 2003: 66). Surveys can be further defined by the way in which they gather data. A survey in which data is gathered just once (even if the data collection process takes an extended period of time to complete) is called a cross-sectional study (Sekaran, 1992); A study in which data is gathered at several different times in order to compare changes over time called a longitudinal study (Collis & Hussey, 2003: 64). Because statistical data analysis techniques are available, surveys are usually directed to only a sample of the research population (Mathers, et al., 2002) with the objective being to reach a statistically-significant number from which the results can be generalised for the larger population (Fox, et al., 2002; Babbie, 2004). 4.4 Sampling In order to make generalisations about research results that a statistically applicable to a wider population, the researcher must apply reliable sampling techniques (Fox, et al., 2002; Saunders, et al., 2007). The sampling technique applied to a study can be one of two types: probability sampling or non-probability sampling (Saunders, et al., 2007). 4.4.1 Probability sampling Probability sampling is based on the principle that a fractional sample of a target population will be statistically representative of the entire population if every individual member of the population has the same probability of being selected for the sample (Blumberg, et al., 2005; Saunders, et al., 2007). Thus, the sample is selected through one of various random methods in order to reduce and hopefully eliminate sampling bias that may affect the results (Blumberg, et al., 2005). With potential bias reduced or eliminated, probability sampling can produce an estimate of the statistical likelihood that patterns observed in the sample are representative of patterns in the full population (de Vaus, 2002a). Depending on the nature of the research problem, the resources available to the researcher, the nature of the target population, the data-collection methods to be used, and the desired degree of accuracy of the results, the researcher can apply a simple random sampling technique, systematic sampling, stratified random sampling, or cluster sampling (de Vaus, 2002a: 71; Babbie, 2004; Blumberg, et al., 2005; Saunders, et al., 2007). 4.4.2 Non-Probability Sampling Any sampling technique that applies a selection criteria not suggested in some way by probability theory is a non-probability sampling technique (Babbie, 2004: 182). With non-probability sampling, statistical generalisations about the wider target population cannot be made, because the probability of selection for individuals in the population is not known; the researcher might still make generalisations about the population from the results, but these would be qualitative rather than quantitative in nature (Saunders, et al., 2007: 207). Situations in which non-probability sampling would be appropriate include circumstances in which probability sampling techniques are impractical or unnecessary, such as in exploratory research or in early phases of some research where survey techniques are being pilot-tested, or when the aim of the research is not concerned with generalising the results to a larger population (de Vaus, 2002a; Saunders, et al., 2007). Some non-probability sampling techniques include convenience sampling, judgemental sampling, quota sampling, and snowball sampling (Babbie, 2004, Saunders, et al., 2007). Read More