Social research methods derive their epistemological foundations from those of sciences as it is recognised that social research is fundamentally a scientific enterprise. Though the subject matter differs, social sciences share with other scientific disciplines in terms of their basic propositions, methodologies and the logic of enquiry. Four different strategies are adopted for different purposes of research viz., experiments, surveys, field research and the use of available data. The experiments are conducted usually by psychologists while sociologists carry out surveys. Anthropologists rely on field research while the existing data is used for analytical purposes by historians, economists etc. While different strategies are adopted for different purposes, all the methods have an underlying commonality in their methods that are built upon a common epistemological edifice.
Origin & Background of Scientific Method
Scientific method originated in the method of reason. Francis Bacon who was hailed as the father of modern scientific method differentiated two ways of discovering truth viz., rationalism and empiricism (Hollis, 1994, p23-40). While rationalism aims at generalizing through theoretical reasoning, empiricism highlights practical experimenting to discover truth. The qualitative and quantitative research methods that emerged eventually built their methodologies essentially on these two ways of discovering truth.
The method according to Comte is unitary for all the disciplines and essentially contains observation, experiment and comparison. Quantitative methods are developed based on this positivistic tradition. Both Quantitative and qualitative research methods derive their foundation from the inductive logic of observation. More than the qualitative method, the quantitative method is built on induction. The two fundamental prerequisites of induction are that there should be large number of observations made and under a wide variety of conditions for any inference to be drawn. Both the approaches seem to have problems in the sense that deduction is built on a pre-existing theoretical edifice and aims at proving it and if that edifice crumbles the entire research structure crumbles. Induction on the other hand insists on variety of conditions and infinite number of observations without knowing how many conditions and observations. Quantitative research derives its method both from the above and thus overcomes their limitations. It starts from a theoretical premise and to prove or disprove that premise number of observations are carried.
Quantitative research methods aim at generating and analysing quantifiable data. It is usually based on the positivistic tradition that knowledge is scientifically decipherable. Its main tools are observation and experimentation and these emphasize objectivity. These methods are built on a pre existing theoretical foundation that trains the researcher to conceptually understand the social reality on empiricist tradition. The theoretical grounding helps in making assumptions about the social reality and thus a problem is hypothesized to begin with. The hypothesized reality is then tested in the field through different tools and techniques like observation, verbal exchange and documentation. The different tools of data collection employed in the quantitative tradition are interview schedule, structured questionnaires, etc that are constructed based on the pre-structured set of assumptions obtained through conceptual understanding of the phenomenon.
Prior to the data collection, a universe is selected where study is being conducted within which sample is identified. The fundamental units are identified within the universe. The universe could be geographical or social. For instance a particular locality or a borough can be identified for the universe or a particular organisation or an institute depending on the study. Within the chosen universe units of the study are identified. The units could be people, organisations, institutes, libraries, or any thing pertaining to the nature of the problem identified. Then a sample is drawn from the universe for the purpose of research and the observations carried out of the sample are ultimately extrapolated to the universe.
Different types of sampling are generally in vogue viz., random sampling, stratified sampling, cluster sampling etc. by which the fundamental units of the study are sampled for the purpose of detailed study. A set of variables is identified for each parameter prior to the study and this classification of variables would be of help while analysing the data. The responses of the respondent are scaled through various scaling techniques like Lickert scaling, Guttman scaling, Thurstone scaling etc. Scaling essentially helps in quantifying the non-quantifiable responses of the respondent. Dependent and independent variables are identified among the parameter that helps in understanding the several factors that contribute the observing phenomenon. The identification of variables as dependent and independent is done based on the pre existing knowledge obtained through theoretical understanding. The responses are quantified in terms of data that is then analysed to get a quantitative description of the phenomena.
Analysis of data under quantitative tradition aims at drawing statistical inferences from the observations made. The data are described through descriptive statistical tools like frequency of occurrences of a particular phenomena and uses the tools like different measures of central tendency (mean, median, mode), measures of dispersion (variance, standard deviation) and measures of association (Lambda, Gamma etc) and employs different statistical tests for measuring the significance of the data. Inferences are drawn from the data with the help of statistical tools like correlation, regression etc.
While quantitative research starts with a problem identification, qualitative research approaches the existing problem. Qualitative research aims at describing social reality qualitatively through much more socially sensitive methods. Even in the qualitative tradition, the fundamental premise is based on a pre- existing theoretical understanding of social reality. But it is carried out without any prior assumptions. In that sense, theoretical grounding only helps in conceptualising the social reality and does not aid in method of observation.
Qualitative research thrives mostly on its techniques of ethnographies, life histories, participant observation, case study approach, content analysis, etc which target basically the sympathetic introspection of the researcher while dealing with the social objects. The methods like participant observation, ethnography especially facilitate the closeness of the researcher to his/her social object and the social reality to understand not only the manifest functions but also the latent, tacit, under currents.
Qualitative approach emerged at the beginning of the twentieth century through the works of anthropologists like Bronislaw Malinowski, Edward Evan-Pritchard, and Margaret Mead et al. The ethnomethodological tradition of these anthropologists made it explicit later the necessity to interpret the reality away from the context and required a critical reassessment of the methods of ethnology adopted by the anthropologists. As a result of this a new qualitative tradition of ethnographic research emerged. As Baszanger et al delineate (1997, P 8), ethnographic studies are identified to have three requirements viz.
- the need for an empirical approach
- the need to remain open to elements that cannot be codified at the time of the study
- a concern for grounding the phenomena observed in the field.
This requires the researcher to be objective and balance all the three requirements essential for ethnographic research. The threat in this is that while sharing the reality with the social objects, the researcher can tend to get in to subjective notions. There is also a criticism for techniques like ethnographies that they are unscientific, too limited for generalisations and they fail to consider inherent representational practices.
Comparison between Quantitative and Qualitative Methods
Both the quantitative and qualitative traditions use interviewing as a method. But, as Miller et al (1997, P99) observe, while positivists insist on a pure interview method enacted in such a way to bring out a mirror reflection of the reality, radical social constructionists believe that no interview can bring out the reality.
While it is agreed that quantitative research methods facilitate generalisations due to their large scale of coverage, qualitative research methods aid in thick description of the reality with much more in-depth understanding. Thus qualitative research helps in addressing larger issues due to their intensive collection of information rather than quantitative methods that can handicap the research findings due to their superficiality and numerical reduction. Techniques like life histories, ethnographic methods are especially useful in studies concerning feminist and Marxist discourses as they can bring out the reality through reflexivity and indexicality.
But qualitative research methods also face the problem of reliability and validity of their data. The ultimate aim of social research is to achieve objectivity. Objectivity is unbiased observation and is a combination of overcoming inter and intra subjectivity. Reliability is defined by Kirk and Miller (as quoted in Perakyla, 1997, p. 203) as the ‘degree to which the finding is independent of accidental circumstances of the research’. Perakyla (ibid) says ‘in ethnographic research the reliability of research entails whether or not the ethnographer would expect to obtain the same finding if he or she tried again in the same way’. As Silverman (1993: 146-8) observes, checking the reliability is closely related to assuring the quality of field notes and guaranteeing the public access to the process of their production.
Similarly validity of research is identified by Kirk and Miller, Silverman et al (as quoted in Perakyla, 1997, p. 207), as that ‘concerning the interpretation of observations; whether or not the researcher is calling what is measured by the right name’. Validity in quantitative research has an underlying assumption that the raw observations have some back ground issues or aspects that they represent. Thus the responses collected to questionnaires in quantitative research represent the underlying attitudes and values of the respondents. But in qualitative research, especially in conversation analysis, the responses are to be analysed at their face value and not for the underlying processes. This adherence to the naturalistic description of the interaction at times raises the problem of validity and raises the issues of tranparence of analytical claims, validation through next turn, etc.Reliability and validity in quantitative research are achieved through scaling techniques as described earlier during observation. And during analysis, they are achieved by conducting several statistical significance tests, through several iterations of the data.