3 Types of Research and Designing Them

Biswajit Ghosh

epgp books

Contents

 

1. Objective

2. Introduction

3. Learning Outcome

4. Types of Research

4.1. Exploratory Research

4.2. Descriptive Research

4.3. Relational Research

4.4. Explanatory Research

5. Stages of Research Plan

Self Check Exercise 1:

6. What is Research Design?

7.  Distinction between Design and Method

8. Purpose and Usefulness of Research Design

Self-check Exercise 2

9. Types of Research Design

9.1. Experimental Research Design

9.2. Cross-Sectional Research Design

9.3. Longitudinal Research Design

9.4. Case Study Research Design

9.5 Comparative Research Design

10. Criteria of a Good Research Design

Self-check Exercise 3:

11.  Summary

12.  Bibliography

 

1.   Objective

 

In this module you will learn about the nature, types and uses of the research designs for sociological research. At the end are given some digital resources for use as also a brief bibliography pertaining to designing a research.

 

2.      Introduction

 

·         All research involves a strategy which provides general orientation to the conduct of social research.

·     Different kinds of research design allow researchers the opportunity to pay attention to research strategy and different frameworks for collection and analysis of data.

 

Research is a systematic process by which one acquires dependable and useful information about a phenomenon or a process. Social research is normally understood to be concerned with conceptual and theoretical issues. But the practice of social research is equally associated with the vision of how social reality should be studied. Social research is not just writings about any one’s feelings and opinions in news paper. It is a serious endeavour and calls for application of well thought plan. As Bryman (2008: 4) argues, ‘Methods are not simply neutral tools: they are linked with the ways in which social scientists envision the connection between different viewpoints about the nature of social reality and how it should be examined’. For instance, a qualitative aspect of social reality, say intensity of any relationship, should be studied qualitatively. On the other hand, we normally apply quantitative research tools to discover, say numerical estimates. Despite valid triangulation of research techniques by social scientists, the fact remains that research involves a strategy, a plan that is to be carried out objectively to justify the findings.

 

Moreover, research findings are also connected to social, economic, political or cultural issues. These connections between a particular research method and the issue being investigated are often critically reviewed as methods are not neutral. It is to kept in mind that three major questions of reliability (whether results are repeatable), replicability (the degree to which the results can be reproduced) and validity (whether conclusions are integrated) haunt researcher all the time. A well devised research strategy helps him/her to look into these criteria used for the evaluation of social research. In other words, both theoretical and practical concerns guide the contours of social research and herein the importance of research strategy becomes pertinent.

 

The term ‘research strategy’ refers to a general or broad orientation to the conduct of social research. It is a step-by-step plan of activity that is devised by the researcher for directing his/her thoughts and concerns. It enables him/her to conduct research systemically so that the results remain valid and reliable. A research strategy also allows him to stay focused, augment quality, save time, energy and money and thereby lessen invisible disappointments at a later stage. For instance, qualitative research demands a particular type of research strategy. The opposite is true about quantitative research. A researcher, therefore, needs to first decide about what he/she plans to do so as to make the strategy full proof. Any failure on the part of any researcher to devise such strategy results in failure to comprehend and analyse social reality.

 

The first step towards devising a research strategy is to take a decision regarding research design. Research design refers to the “framework for the collection and analysis of data” (Bryman 2008: 31). As there are various types of research designs, choice for a particular one by any researcher would reflect his/her decision for the criteria and the research questions in which the researcher is interested. Obviously, different kinds of research design allow researchers the opportunity to pay attention to different research stratagems and different frameworks for collection and analysis of data.

 

·         Selections of research design and research methods are two key decisions (among others) that affect the outcome of any research.

The outcome of a research depends to a large extent on the selection of research design and research method. A comparative study, for instance, must select a comparative research design. Again, if a researcher, for instance, expect qualitative data, the technique chosen for the study would also clearly reflect such choices. Normally the techniques chosen for any qualitative study are observation, case study, focused group discussion or participatory rural appraisal. Similarly, survey methods are used to collect quantitative data. These selections are often done concurrently as there are strong association between design and method. Many a time, certain designs have declared preference for particular method. Nonetheless, it is to be noted that despite their closeness, research methods and research designs are not the same. We would discuss differences between the two later.

 

3.   Learning Outcome

 

This module would acquaint you with the types of research in social sciences in general and the issue of research design in sociological research in particular along with its various types and their usefulness. It will also provide an introduction to the more important resources needed for devising the research strategy.

 

4.   Types of Research

 

Social research is normally classified into qualitative and quantitative types. Qualitative research usually emphasizes on sentiments, feelings and processes rather than on quantification for the collection and analysis of data. A qualitative researcher is open to the respondent’s views, actions and experiences. On the other hand, quantitative research stresses on quantifiable data for the collection and analysis of research findings. This viewpoint prefers observable variables or concrete expressions of variables that meet pre-defined standards of reliability and validity.

 

The classification of research into qualitative and quantitative is not complete. Another common way of classification is pure and applied research (Goode and Hatt 1981: 29-40). Pure research (also known as basic or fundamental research) is concerned with collection of data that aims at formulation, expansion or evaluation of a theory. Its primary concern is creation of knowledge solely for the sake of knowledge. For example, Max Weber’s research on the relationship between Protestant Ethics and Capitalism is a contribution to pure research. On the contrary, applied research is directed towards the solution of an immediate, specific and practical problem. Through such research, developmental professionals are able to solve problems at the appropriate level. For instance, field research on the issue of child marriage, trafficking and dowry in West Bengal was designed specifically to devise action plan against them by the state government (Ghosh 2007). But applied research does contribute to theoretical knowledge and vice versa; hence they are complementary and not opposed to each other.

 

The classification of research into some dichotomous types (qualitative and quantitative, pure and applied) may however appear to be grossly inadequate as one may also identify typology of research based on research objectives. From the point of view of research strategy, in particular, such classification is very useful. Following Selltiz, Wrightsman and Cook 1976), we may identify the four main research types:

 

4.1. Exploratory research (sometimes called formulative research) aims to gain familiarity with or to achieve new insights into a phenomenon, often in order to formulate a more precise research question or to develop hypothesis. Such a research helps researcher to understand the points of view of actors in the setting, identify worthy research questions, articulate and operationalise variables of interests and allow theory to emerge (Palys 1997: 79). Obviously, strategy of such research demands flexibility, breadth of coverage and open-ended techniques. Well-executed exploratory research helps to identify useful variables and provide hints about valid ways to measure them.

 

4.2. Descriptive research aims to accurately portray the characteristics of a particular individual, situation, sample, population or a group. Descriptive research is concerned with ‘what’ and ‘how’ questions. For instance, what is the extent of literacy in India? How do students use internet facilities? Accuracy is paramount in such research. Once the researcher has gained a preliminary understanding and identified variables that appear important, more systematic description of the situation and the corresponding variables is called for. Such a research tries to minimise bias while maximizing the representativeness of the sample thereby ensuring that results are valid and reliable.

 

4.3. Relational research (sometimes called correlational research) aims to determine how two or more variables are related within a given sample or population. Such research is often considered to be the first step towards explaining phenomena as theoretical questions very often guide researcher to look for the relationship between two or more specific variables. Though measures adopted for such a research must be valid and reliable, the sampling technique may or may not be representative. The scope of such research remains limited as the integral and causal connections between two particular variables might be caused by some unseen factors.

 

4.4. Explanatory (also called experimental) research aims to investigate causal relationship or other patterned conduct that is thought to characterise social processes. Such research requires the researcher to focus very clearly on the research questions, identify dependent and independent variables, develop hypothesis and test them by manipulating independent variable. Such research also use highly structured methods designed specifically for the research purpose. Control, manipulation and observation are the three essential characteristics of explanatory research.

  1. Stages of Research Plan

 

Once the research problem is formulated and the subject of study is identified, the researcher must try to determine the actual means for carrying out the tasks. This is the proposal-writing stage when the overall plan for carrying out the research is set out in logical order so that it works effectively. The initial plan may however go through many modifications and change over time as the research progresses and insights into it deepen. The working out of the plan consists of making certain “decisions regarding the what, where, how much, by what means, constitute a plan of study or a study design” (Young 1966: 12-13). The stages of a research plan include the following aspects:

 

a) Define the research topic,

b) Review existing literature,

c) Identify the nature and objective of study and develop hypotheses, if required,

d) Collect sources of information about the field,

e) Clarify concepts and their measurement,

f) Operationalise concepts and design the research instrument,

g) Determine sample size and sampling procedure,

h) Decide about appropriate data collection method,

i) Enter into the field and collect the data,

j) Process and analyse data collected.

k) Present the results.

 

Self Check Exercise 1:

  1. Why does research involves a strategy?

 

All research involves a strategy because it is a systematic process by which one acquires dependable and useful information about a phenomenon or a process. The practice is also associated with the vision of how social reality should be studied. It is a serious endeavour and therefore calls for application of well thought plan, i.e., a strategy that can be carried out objectively to justify the findings.

  1. What does a research strategy entails?

Research strategy, which refers to a general or broad orientation to the conduct of social research, enables a researcher to conduct the research systemically so that the results remain valid and reliable. It also allows him to stay focused, augment quality, save time, energy and money and thereby lessen invisible disappointments at a later stage.

  1. What are the objectives of Exploratory Research?

 

The aim of exploratory research is to gain familiarity with or to achieve new insights into a phenomenon. Its basic objective is to formulate a more precise research question or to develop hypothesis. It also helps researcher to understand the field situation, articulate and operationalise variables of interests and allow theory to emerge.

  1. What is Research Design?

 

In simple sense, the term research design refers to a framework for the strategy of enquiry, methods of data collection from field and the analysis of data collected. It involves a game plan through which one can gather information/data that are required for the purpose of the research in a simple, elegant and systematic way. It includes everything, from priorities and objectives of research to questions regarding who, what, when, where and how. According to Selltiz, Wrightsman and Cook (1976: 90) “a research design is the arrangement of conditions for collection and analysis of data in a manner that aims to combat relevance to the research purpose with economy in procedure”. It therefore situates the researcher in the empirical world and connects them to field location, people, institutions including documents and materials. By doing so, it allows the researcher to address the two critical issues of representation and legitimation in research (Denzin and Lincoln 1994).

 

A research design, thus, identifies and defines the study type, research questions, hypotheses, variables, data collection method, sampling, time frame and analysis plan. The work plan of a research therefore flows from the design as it suggests types and sources of data to be collected and analysed keeping in mind the objectives of research. Hence, it helps the researcher to collect relevant information that is needed to answer the research questions objectively, adequately and effectively. In short, it is plan, structure, and strategy for the collection, measurement and analysis of data. It describes the structure of an enquiry in a logical manner so that chances of drawing incorrect casual inferences from data can be minimised.

  1. Distinction between Design and Method

 

Research design is different from the method by which we collect data. In fact, a research design includes the types of methods to be followed for the collection of data. Hence, these are conceptually different and one should not confuse between them. Thus, research design is a logical structure of the enquiry while research method is a mode or a technique of data collection (read discussion on module 11-13 of this paper). For instance, interview is a tool of data collection and it is used mostly in cross-sectional research design. But it is not a research design itself. It should also be noted that a research design may opt for combination of two research methods for meeting the research objectives. There is therefore nothing intrinsic about any research design that demands a particular method of data collection. Technically speaking, any research design can use any type of data collection method and can use either qualitative or qualitative data. In other words, the questions regarding how the data for a study are collected is irrelevant to the logic of any design. Failure to distinguish between the two results in poor evaluation of designs. This is because all methods have their strength and weaknesses. The success or failure of any design is not therefore evaluated by success or failure of any method. Contrarily, the ability of a design to draw objective, valid and reliable conclusions is evaluated.

  1. Purpose and Usefulness of Research Design

 

Research design has two basic purposes: a) to provide answers to research questions and b) to control variance (Kerlinger 1978: 300). Though all these are done by the investigator, a design helps him/her to obtain answers to the questions of research while helping to control the experimental, extraneous and error variances of the particular research under study.

 

When a researcher plans to carry out a study, he/she would obviously look for data that are valid, objective, accurate and economic to the extent it is possible. Obviously, a research design is deliberately planned to see to it that these basic research objectives could be achieved. For instance, the issue of research known as ‘research problem’ is normally stated in the form of hypotheses. The researcher is supposed to test these hypotheses empirically in the course of the study. Now theoretically, several research designs are available to carry out this task. It may be noted that research problems suggest research designs. Hence, a researcher has to select the particular variety of research design that would yield dependable and valid answers to the research questions shown in the hypotheses. As the results of a research depends on how the observations and inferences are made, the researcher needs to adequately plan and execute his research design.

 

Now, the question is, how does design accomplish this? As Kerlinger argues, “research design sets up the framework for ‘adequate’ tests of the relations among variables” (Ibid. 301). Design suggests us what observations to make, how to make them and how to analyse the quantitative data collected. An adequate design also outlines possible conclusions to be drawn from the statistical analysis.

 

Let us take an example to explain the process. A researcher, for instance, takes up the task to evaluate the changing nature of a village life. The essential question he/she is asking is: what is the impact of a new government scheme (say Mahatma Gandhi National Rural Employment Guarantee Act or MGNREGA) on rural people? For this research, he may select a village where the scheme is implemented and compare the findings with a ‘control group’, that is another village where the scheme is not implemented. To conduct such research, the researcher has to take the help of a quasi-experimental design. This is because he/she cannot carry out any true laboratory experiment on village people though an attempt is made to compare the life pattern of two segments of population. Similarly, a researcher may make use of comparative design if he/she is attempting to study two contrasting cases using more of less identical methods. In this way, research designs are helpful to provide answers to research questions. We would discuss these and many other designs in details in the next section.

 

So far as question of variance control is concerned, a research design indeed works as a control mechanism in three ways: a) maximizing systematic variance, b) controlling extraneous variance and c) minimizing error variance. The term ‘maximising systematic variance’ means that the investigator is able to manipulate and control independent variable(s) to such an extent that he/she is in a position to measure the reaction of the dependent variable(s) as a result of such control. For instance, an investigator may try to test the hypothesis that says that ‘educational performance of college students depends more on intelligence than on class background’. Here, the investigator is aware of the fact that other possible independent variables like sex, social capital, educational background of family members or previous experience might also influence the educational performance of a college student. But, he/she may avoid these variables as due to their inclusion there would be little chance to separate the variable effect of intelligence and class background on educational performance. In other words, selection of only two independent variables would allow the research to manipulate the dependent variable, that is educational performance. At a later stage, he/she may include other independent variables to develop a comparative perspective on the issue.

 

Similarly, an investigator needs to control variable that is extraneous to the purpose of the study. For instance, in the study cited above, the investigator may avoid the influence of race by selecting students only from a similar race. Here the unwanted variable is just eliminated. The other way to control such extraneous factor is to accept it as an independent variable and yield additional research information about its effect on educational performance.

 

Finally, error variance can be minimised by a) reducing the errors of measurement through controlled conditions, and b) by increasing reliability of measures. The concern for minimising errors in social sciences is related to achievement of ‘validity’ of findings, meaning integrity of the conclusions that are generated from a piece of research. The main types of validity are: measurement validity, internal validity, external validity and ecological validity. Let us explain these types in brief. Thus, the criterion of measurement validity applies primarily to quantitative research and it tries to find out whether a selected concept, say family background, is really able to measure variations in intelligence. As various other factors do influence intelligence, the measures chosen for such a test also should reflect those. Internal validity relates mainly to the issue of causality and it seeks to test whether, for instance, x causes y and can we be sure that only x is responsible for variation in y. The issue of external validity is concerned with the question of whether the results of a study can be generalised beyond the specific research context. Finally, ecological validity is concerned with the question of whether social scientific findings are applicable to people’s everyday setting (for a detail discussion on these types of validity, read Bryman 2008: 31-33)

 

Apart from our concern for validity, reliability of research findings is also stressed and it refers to accuracy of a set of scores. The findings of a study is said to be ‘reliable’ if the results are repeatable. Reliability is particularly required for any quantitative research. For instance, the results of two IQ test conducted on the same set of persons should be the same. If they are not, then the data are not reliable. In qualitative studies done under field conditions, however, such exercise is difficult. Yet, efforts must be made to maintain uniformity and objectivity in collecting data and analysing them by excluding extraneous variables.

 

It appears that the basic purpose of research design is to allow the investigator to eliminate potential threats to validity and reliability in advance. Though it is often difficult to recognise and develop designs in advance that ere efficient and appropriate, a good design also presumes that the investigator knows the relevant variables and is able to handle them. He/she cannot take the chance of any rival hypothesis based on any rival independent variable to dominate over his findings. As a result, a good design begins with conceptualisation, which against starts with careful consideration of the nature of the problems and careful review of existing literature. Even in exploratory research, one must have some guiding idea of what is to be explored and what not though such tentative guidelines may change later on.

 

Self-check Exercise- 2

  1. What is the research design?

 

Research design refers to a framework for the strategy of enquiry, methods of data collection from field and the analysis of data collected. It involves a game plan through which one can gather information/data that are required for the purpose of the research in a simple, elegant and systematic way. It includes everything, from priorities and objectives of research to questions regarding who, what, when, where and how.

  1. How do research designs differ from research methods?

 

Research design is a logical structure of the enquiry while research method is a mode or a technique of data collection. A research design includes the types of methods to be followed for the collection of data.

  1. What is the usefulness of a research design?

A research design is planned for achieving the basic objectives of a research. It helps the researcher to collect valid, objective, accurate data by spending minimum time and energy. Design suggests us what observations to make, how to make them and how to analyse the quantitative data collected. An adequate design also outlines possible conclusions to be drawn from the statistical analysis. It also works as a control mechanism in three ways: a) maximizing systematic variance, b) controlling extraneous variance and c) minimizing error variance.

  1. Types of Research Design

Social scientists differ as to what really constitute the types of research design. Often artificial distinction is maintained to classify a type, while in other cases different designs are combined to form a type. There are also attempts on the part of social scientists from a particular branch to argue for a design that is suitable to say evaluative or action research. Despite these differences, we may identify the following five types based on Alan Bryman (2008: 35-61) classification. While discussing these types, we would also try to analyse their relative usefulness with suitable examples. It should be kept in mind that choice of a particular research design over others is basically governed by purpose and objective of a research.

 

9.1 Experimental Research Design

 

A research design that rules out alternative explanations of findings from its research by using a) at least one experimental group for treatment, and a control group which is not used, and b) random assignments to the two groups is called experiment. In an experimental design, the researcher tries to change the situation, condition or circumstances of the participants called ‘manipulation’, which may lead to changes in the behaviour of the participants of the research. Experiment, therefore, entails a comparison of results obtained from two groups under certain conditions. True experiments are, however, very rare in sociology. This is because, in a true experiment independent variables are manipulated in order to determine whether it has any impact on the dependent variables. But, in social research, presence of too many of independent variables makes the situation complex. Again, independent variables like caste, class, religion, gender cannot be manipulated or changed. Yet, field experiments (not laboratory experiment) are conducted in related fields like Social Psychology, organizational and evaluative studies including research on social policy. Moreover, findings of non-experimental research are often assessed by using true experiment as yardstick. In other words, experimental designs are treated as benchmark for being very strong in terms of internal validity. As a result, students of sociology are also taught about this design.

 

In experiment, two groups are chosen for experiment. The group that receives treatment (manipulation), is called experimental group; but the group that does not is called control group. The results are measured before and after the experiment. For instance, one may presume that computer aided teaching (independent variable) improves the performance of students (dependent variable). Now a time bound experiment to prove this relation may be done on two groups of students of the same class who have not been exposed to computer aided teaching earlier. The group A, called experimental group, will be introduced to computer aided teaching for six months, while the group B, called control group, will not. The results of these groups will then be compared.

 

There are however some aspects of threat in this type of study. It is difficult to eliminate the influence of other factors in influencing the result. For instance, in the example stated above, the intelligence level of the students might also influence their relative performance. Availability of suitable participants for such experiments is also a significant issue. Hence, questions of validity and replicability are often raised if the researcher fails to eliminate the external threats.

 

Given the constraints of conducting field research on social issues, some researchers have also tried to frame quasi-experiments – those having certain characteristics of experimental design but that do not fulfil all of the internal validity requirements. Quasi-experiments may however take different forms (Cook and Campbell 1979). An interesting type is concerned with ‘natural experiments’ in which the social setting is manipulated as a part of a naturally occurring attempt. The researcher however cannot randomly assign subjects to experimental and control groups. But because of lack of artificial intervention in social life, these studies are considered very strong ecologically. In evaluation research, in particular, quasi-experimental design has become very popular. Koeber (2005), for instance, has attempted to find the effects of introduction of new technology on students. One group of students was taught with the use of multimedia and a course website (experimental group), but the other group did not experience these. The findings reveal that the students of experimental group had a very favourable view about the course as compared to the control group students. But, there was hardly any difference in the final grades of the two groups of students. Attempts have also been made to find out the effects of television violence on the children by using a natural experiment (see Bryman 2008: 42, for details).

 

9.2 Cross-Sectional Research Design

 

Cross-sectional design is a non-experimental design that does not involve any manipulation of the situation or condition. It is often called a ‘survey design’ just because the research methods associated with survey are frequently used to conduct such research though many other tools are also used. The design entails the collection of data on more than one case and at a single point in time in order to collect a body of quantitative data in connection with two or more variables, which are then examined to detect pattern of association (Bryman 2008: 44). This definition makes it clear that cross-sectional design has the following features:

 

a) It employs more than one case as the researcher is interested in variation and finer distinctions between cases with large sample base;

b) It is done at a single point of time because the data collected for such a research through, say interview, is done simultaneously;

c) It provides quantifiable data in order to establish variation between cases as well as to examine relations between variables. By doing so, such design provides standardised measurement of social situations, objects and events.

d) It is also possible to study pattern of relationships only between variables by such design. But from the findings, one cannot say which variable affects the other as the researcher does not collect the data over time and does not manipulate any of the variables. Hence, it is not possible to draw certain inferences about causality based on a cross-sectional design.

 

As a corollary, such design lacks the internal validity that is normally the characteristics of experimental research. But, replicability is likely to be present in most cross-sectional research if the procedures of data collection and analysis are spelt out clearly. Similarly, external validity of such research remains strong when data are collected from randomly selected samples. But, ecological validity will certainly be disturbed when the researcher uses a research instrument to intervene in the normal setting of social or individual life. But, it has been noted earlier that social researchers cannot manipulate or control many aspects of social life in which they are interested. For instance, anyone interested in finding a casual relationship between alcoholism and economic background cannot select his/her sample of alcoholic and non-alcoholic persons randomly from a neighbourhood. One also cannot ask people to become alcoholic for the purpose of research. Manipulating a non-manipulable variable (say ethnic group identity, caste or sex) raises serious methodological concern in social research.

 

The practical and ethical concerns of any research, therefore, forces social scientists to employ a cross-sectional research design. More interestingly, much beyond quantitative research, such design is also used in qualitative research with the help of unstructured interview schedule, structured observation, content analysis, official statistics and diaries. Without bothering about questions of validity and replicability, such research is able to contribute significantly to the body of sociological literature.

 

9.3 Longitudinal Research Design:

 

A longitudinal research design tries to examine variables such as academic performance of a group of students or groups of students over a period of time. It stresses on collection of data from the sample(s) on at least two occasions. When the same group of people is interviewed at regular intervals, it enables researcher to track changes over time and help establish the direction and magnitude of change. A longitudinal research design, therefore, allows some insight into time order of variables and is able to allow causal inferences between them.

 

There are two ways in which such research can be conducted: the panel study and the cohort study. In panel study, a group of people most often selected randomly is asked for a specific set of information on at least two occasions. It is also possible to collect data from different types of case within a panel study framework. The British Household Panel Survey, for instance, had attempted to interview 10,264 people of 5538 households annually since 1991 on six main issues (household organization, labour market behaviour, income and wealth, housing, health, and socio-economic values). As compared to this, either an entire cohort of people or a random sample of them is selected for the cohort study. The cohort is made of people who share a certain characteristics or experience, such as born on the same month, or being unmarried or employed. Often such studies are not planned originally, but planned at a later stage. In India, several cohort studies have so far been conducted on health and developmental issues.

 

Panel and cohort studies share similar features though there are some differences between them. Thus, a panel study that is done for many years can distinguish between age effects and cohort effects. But a cohort study can only identify aging effects. Again, unlike the cohort study, a panel study conducted particularly at household level needs to frame rules regarding entry of new or exit of old members.

 

There are also some critical issues related to these studies. Thus, a researcher following this design will have to face the problem of simple attrition through death, refusal, migration and unavailability due to illness or unknown reasons. Many of the known longitudinal study do not involve the same set of people on each occasion (read Bryman 2008: 51). Those who leave the study may differ in certain respect with those remaining. So the question of whether the second group truly represents the original group of informants remains valid. One may also raise questions about the timing of research. The influence of any extraneous variable on the findings of subsequent survey cannot be ruled out. Longitudinal studies also need to be planned carefully as huge amount of data are collected by such survey. There is also evidence to suggest that a panel conditioning may occur due to continuous participation of the same respondents thereby influencing their opinion.

 

Despite these limitations, these studies are basically concerned with estimating the volume of change as well as improvement over a period of time on say smokers, children, or HIV infected patients. No other design is able to shed proper light on this aspect of social living. Though they are able to locate change, they cannot address the issue of the direction of the cause and effect accurately because of the limitations mentioned above. It is to be noted that even some qualitative study incorporates some elements of longitudinal design for better explanation of social reality. An ethnographer staying in the field for a considerable period of time takes note of the changes. Eminent Indian sociologist M.N. Srinivas (1996), for instance, had noticed qualitative changes in the nature of caste or social life in village through field studies conducted over a period of time. Herein lay the strength of longitudinal research design.

 

9.4 Case Study Research Design

 

A case study examines a current phenomenon in its real-life situation, using whichever research strategies are necessary to address the problem in hand (Yin 2003). It is an in-depth study of a particular research problem instead of a general survey. It is not a specific method of social research, but a research strategy which focuses on a single organization, institution, event, decision, policy or group. It entails the detailed and intensive analysis of that single case. It is often used to narrow down a very broad field of research into one or a few manageable research problem. There are many brilliant examples of eminent sociologist studying a single community, a single family, a single organization or even a single person (read Bryman 2008: 51-52, and Baker 1994: 299-300).

 

Case studies are employed both for quantitative and qualitative studies though there is often a tendency to associate it with only the latter type. It is however true that qualitative technique like participant observation and unstructured interview are particularly helpful for intensive and detail examination of a case. But due to large scale use of this strategy for quantitative studies, it is also not easy to differentiate whether a study is based on case study or cross-sectional design. Illustrating on the difficulties on case studies, Bryman has argued, “I would prefer to reserve the term ‘case study’ for those instances where the ‘case’ is the focus of interest in its own right” (2008: 53).

 

It appears that a specific research illustration may contain features of more than one research design. What makes case study so particular is its capacity to elucidate the unique features of the case. Robert Yin (2003) thinks that case studies are appropriate when the research question to be addressed asks how and /or why. They may be largely exploratory or descriptive. The reason behind studying a particular case is to find out why certain situation prevails or how a group/organization has succeeded. The design is useful for testing whether a specific theory and model actually applies to phenomena in the real world. By doing so, it helps us to explain a complex issue or phenomena by referring to the context and location in which the argument holds good.

 

It is pertinent here to distinguish between several types of case studies. Yin (Op.cit. 2003) has identified five types of case studies. He called the first type ‘critical’ where the researcher, backed by a well developed theory, chooses the field. The case would then allow a better understanding of the situation leading to acceptance or rejection of the hypothesis. The second one is ‘extreme’ or unique type which is found in clinical studies. Margaret Mead’s study of growing up in Samoa, for instance, is motivated by her belief that the case is unique (1928). The third type is called ‘representative’ or typical. Such cases exemplify a boarder category of events that is helpful to answer certain research questions. These cases also allow researcher to examine key social processes. The fourth type is ‘revelatory’ in the sense that it was previously inaccessible to scientific investigation. The last type is called ‘longitudinal’ as it provides opportunity to study an issue over time. Notwithstanding such classification, it should also be noted that a single case may also represent two different types. Hence, cases should be subjected to detail scrutiny for revealing their true character and inner strength.

 

Case study design is also not free from limitations. Time and again questions regarding its external validity or generalizability are raised. How can a single or small number of cases offer reliable and representative data? To what extent, for instance, the findings of a research about a village in South India hold good for the whole of even South India? Exponents of case study however feel that the objective of their research is not to generalise their findings to other cases or to populations beyond the case. Rather, they aim to conduct an intensive examination of a single case so as to identify the quality of the theoretical reasoning. This makes case study associated with both theory generation and theory testing. It is possible that case study researchers do not delude themselves and try to identify typical cases that may be argued to be representative. They are also in a position to generalise by drawing on findings from comparable cases investigated by others. The tradition of village studies in India, for instance, has yielded much rich data and concept and there is no second opinion that these findings have enriched the state of sociology in India.

 

9.5 Comparative Research Design

 

Comparative design involves the study of two contrasting cases using more or less identical methods. By doing so, it tries to explain the existing theory or generate new insights as a result of contrasting findings revealed through comparison. The basic logic behind this design is that social phenomena can be better understood if they are compared in relation to two or more meaningfully contrasted cases or situations.

 

The scope of such design is wide as it can be used in both quantitative and qualitative types of research. In case of quantitative research, data are collected from at least two situations or cases within a cross-sectional format. In cross-cultural or cross-sectional research, particular issues from two or more countries are compared using the same research instrument. The aim is to seek similarities and differences or to understand deeply the social reality of two different national contexts. For instance, there have been attempts to estimates the volume of human trafficking from counties of South Asian region. Similarly, such research can also be conducted at international level. United Nation’s Human Development Index (HDI), for instance, provides estimates of level of human development of different countries of the world based on certain given yardsticks. These estimates are very often based on secondary data supplied by respective countries. Researchers have also used the tool of survey research to collect data on say activities of self-help groups in Bangladesh and India. Such research helps to reduce the complaint that social science research is confined to cultural setting. Moreover, findings of such research are useful to focus beyond cultural specificities of different countries.

 

Comparative research can also be applied to a variety of situations and not just between two nations. Thus, conditions of labour markets in urban and rural context can be contrasted within a country. Comparisons may also be in time, between conditions before the study and conditions after introduction of a measure, say rural development programme. When applied to any qualitative research strategy, comparative research takes the form of multiple-case study. For instance, comparative study of two or more villages provided impetus for generation useful concepts in Indian sociology. Srinivas’s comparison of different cases led him to replace the term Brahminization by Sanskritization (Srinivas 1966). In organization studies, for instance, multiple-case studies are utilised to build a theory. J.B.P. Sinha (1990) has studied the contrasting work culture of public and private sector units of two steel plants, two fertilizer companies and two nationalized banks to generate useful hypothesis. Recently a comparative study of 12 tea gardens was undertaken to estimate the nature and intensity of vulnerability of children in the plantation industry (Ghosh 2012). These examples suggest that qualitative comparison is helpful to generate concepts and build theory.

 

Conducting cross-cultural research is not, however, without problems as management of resources and collection of fund at two different national levels may appear to be a very hard task. Also the data collected should be comparable in terms of categories and data collection techniques. Along with language barrier, cultural specificity might pose some challenge for viable comparison. Despite these challenges, cross-cultural research has the ability to go beyond the limited context and such widening of the field areas are very helpful for generation of viable concepts and theories.

  1. Criteria of a Good Research Design

Design is a data discipline. The implicit purpose of all research designs is to impose controlled restrictions on observation of natural phenomena. A research design tells the investigator what to do and what not, what to ignore and what not. It is the blueprint of the research architect and engineer (Kerlinger 1978: 327). Hence, if the design is poorly conceived structurally, the ultimate product with be faulty. It is therefore pertinent to discuss about the criteria of a good research design. These criteria will act as guiding principles for the investigator. The following are the most significant criteria of a research design:

 

a) Answer Research Questions: The main criterion of any research design is whether it is able to answer the research questions. Or to put it in a question frame: does it adequately test the hypothesis? A good research design must ensure that there is congruence between the research questions and hypothesis. Often the variables being matched are found to be irrelevant to the research purpose.

 

b) Control of Extraneous Independent Variable: The second criterion is control of independent variables. The researcher must identify the extraneous and unwanted independent variable which may influence the dependent variables and try to minimise their influence on the outcome of the study. In order to control independent variable, the investigator should select subjects at random and assign subjects to groups at random.

 

c) Generalizability: The third criterion of generalizability demands that the findings of the research should also be applicable to other subjects, groups and conditions. In order to find out the extent of generalization of a research outcome, one should ask questions like: How much can we generalize the results of the study? or To whom and what can we generalize the results of the study?

 

d) Reliability: Reliability  is  concerned  with  the  question  of  whether  the  results  of  a  study  are repeatable. Synonyms for reliability are: dependability, stability, consistency, predictability, and accuracy. In quantitative research, in particular, consistency and accuracy of the results is checked by looking into the question of whether the measures adopted (or instruments used) for the study are stable or not. Hence, reliability can simply be defined as the relative absence of errors of measurement in a measuring instrument.

 

e) Replication: The idea of replication is very close to the notion of reliability. The degree to which the results of a study can be reproduced is known as replicability. In order to do so, a researcher should follow the procedures adopted by another researcher on the same subject/field to find out similarities and differences of their findings. Though in social research replication is very rare, such quality is highly valued in the quantitative tradition.

 

f) Validity: This is one of the most important criteria of research and is concerned with integrity of the conclusions. It is a technical aspect of any research and basically relates to the question of what is being measured. As has been discussed earlier, the main types of validity include measurement validity, internal validity, external validity and ecological validity (read, for details, section 8: Purpose and Usefulness of Research Design).

 

It appears from the foregoing discussion that a good design is flexible, appropriate, efficient, accurate and economical too. It also should minimise bias and maximise the reliability of the data. But, the point to be kept in mind is that the need for a design is also related to the objective of research as well as the nature of problem to be studied. A design found to be suitable in one case/context, may turn out to be defective in other case/context. In other words, a single design cannot serve the purpose of all types of research and hence several types of designs are evolved to cope with divergent research stratagems.

 

Self-check exercise 3

  1. What are the types of research design?

Social scientists differ on the issue of typology of research design. Yet, we can distinguish between five major types, namely experimental, cross-sectional, longitudinal, case study and comparative types. As the goal of each of these designs is different, choice of a particular research design over others is basically governed by purpose and objective of a research.

  1. What are the limitations of experimental research design?

Though experimental designs are treated as benchmark for being very strong in terms of internal validity, they have certain limitations. It is difficult to eliminate the influence of extraneous factors on the findings of the research. Hence, questions of validity and replicability are often raised if the researcher fails to eliminate the external threats. Availability of suitable participants for such experiments is also a significant issue. True experiments are very rare in sociology as independent variables like caste, class, religion, or gender cannot be manipulated or changed to locate their impact on dependent variables.

  1. Why are questions of reliability and validity significant in social research?

 

Research is a systematic process by which one acquires dependable and useful information about a phenomenon or a process. Obviously, one would expect the findings of any research to be reliable, replicable and valid. These criteria of research design improve trustworthiness of any research findings and allow others to use the findings for comparative analysis. A piece of research must not only be relevant, it must also meet the minimum requirements for ‘craft’ of research.

  1. Summary

 

This module would have helped you to develop familiarity with some theoretical and practical concerns of social research. You have learnt about the types of research, the importance of research strategy, stage of a research plan and the notion of research design. This module would also make you aware about the various types of research design and their uses with suitable examples. You have seen the context in which a particular research design becomes more useful. You have also become acquainted with the limitations and shortcomings of each of these designs and the criteria to be followed for preparing a good research design.

 

you can view video on Types of Research and Designing them

 

References

  1. Baker, Therese, L. Doing Social Research (2nd edition). New York: McGraw-Hill, Inc., 1994.
  2. Bryman, Alan. Social Research Methods (3rd edition).  Oxford University Press, 2008.
  3. Cook, T. D., and Campbell, D.T. Quasi-Experimentation: Design and Analysis for Field Settings. Boston: Houghton Mifflin, 1979.
  4. Denzin, N.K and Lincoln, Y.S (Eds.). Handbook of Qualitative Research. Thousand Oaks: Sage, 1994.
  5. Ghosh, Biswajit. Trafficking in Women & Children, Child Marriage and Dowry – A Study For Action Plan in West Bengal (Edited & Compiled), Published by UNICEF and Govt. of West Bengal, Kolkata, 2007.
    1. ——. Study on Vulnerability of Children in Sick & Closed Tea Gardens in Jalpaiguri District, West Bengal, Save the Children and UNICEF, 2012.
  6. Goode, W and Hatt, P.K. Methods in Social Research. Auckland: McGraw-Hill Book Company, 1981.
  7. Kerlinger, Fred. N. Foundations of Behavioural Research (2nd Edition). Delhi: Surjeet Publication (Indian Reprint), 1978.
  8. Koeber, C. “Introducing Multimedia Presentations and a Course Website to an Introductory Sociology Course: How Technology Affects Student Perceptions of Teaching Effectiveness”. Teaching Sociology (2005), vol. 33: 285-300.
  9. Mead, M. Coming of Age in Samoa. New York: Morrow, 1928.
  10. Palys, Ted. Research Decisions – Quantitative and Qualitative Perspectives (2nd Edition). Canada: Harcourt Brace, 1997.
  11. Salkind, Neil J. Exploring Research. New York: Macmillan College Publishing Company, 1994
  12. Selltiz, C., Wrightsman, L.S., and Cook, S. W. Research Methods in Social Relations (3rd Edition). New York: Holt, Rinehart and Winston, 1976.
  13. Sinha, Jai, B. P. Work Culture in the Indian Context. New Delhi: Sage, 1990.
  14. Srinivas, M. N. Social Change in Modern India. New Delhi: Orient Longman, 1966.
    1. —–. Village, Caste, Gender and Method, New Delhi: OUP. 1996
  15. Yin, Robert K. Case Study Research: Design and Methods (3rd Edition). Thousand Oaks, Calif.: Sage, 2003.
  16. Young, P. V. Scientific Social Survey and Research. New Delhi: Prentice Hall of India, 1979.