Comprehensive Guide to Research Methodology – Design | Methods | Best Practices

Research Methodology refers to the specific procedures or techniques used to identify, select, process, and analyze information about a topic. It includes the theoretical framework and methods chosen by a researcher to effectively answer a research question or test a hypothesis. This can involve various methods such as qualitative or quantitative research, and different approaches like surveys, experiments, observational studies, or case studies.

Table of Contents

1. Introduction
2. Steps in Research Process
3. Classification of Research Design

Check out Taxmann's Research Methodology – Methods | Techniques | Practices which simplifies and enriches research methodologies for management students and doctoral scholars authored by seasoned educators. It addresses the unique needs of Indian students by integrating relevant examples and case studies. Structured into four parts, it covers the basics of scientific research, quantitative and qualitative research foundations, and report writing. Pedagogical elements include clear learning outcomes, real-world examples, boxed summarisations, key terms, review questions, and practical demonstrations in SPSS and Excel. The book's clear and straightforward writing style makes complex ideas accessible without compromising depth.

1. Introduction

This article describes the research process and different research designs in detail. Management and social science research, like other forms of scientific inquiry, require a structured sequence of highly interrelated steps (Zigmund et al., 2010). The research process involves a series of steps or actions essential for the smooth conduct of any research. The figure below illustrates the sequence of the research process. It is to be noted that these steps are not a road map to all kinds of research. Basically, it is applicable for deductive or functionalist research, and it can or needs to be revised to suit the requirements of a specific project. The research process doesn’t need to be followed successively; rather, the steps overlap frequently and are interrelated. The research process offers a comprehensive guideline that can be referred to for any management and social science research. It may happen that later stages can be accomplished before the earlier stages.

The steps involved in the research process are neither mutually exclusive nor separate and distinct. The selection of a research topic at the outset, defining the research problem and objectives, influences the selection of a sample and data collection. The sample selection may affect the design of questionnaire items. For example, suppose an organization wants to know the cause of attrition among lower-category employees with low educational qualifications. In that case, the wording for the questionnaire will be easier than for people in top management positions with professional educational qualifications. The steps may differ based on the objectives of the research. However, research based on deductive logic should follow the steps outlined below:

2. Steps in Research Process

• Problem Identification
• Literature Review
• Formulating Research Questions
• Research Design
• Sampling
• Data Collection
• Data Analysis
• Conclusions and Report Writing.

The quest for research must always be triggered by the longing to explore and gain more knowledge and understanding. The management dilemma encourages the need for a decision. The need may arise owing to the cause that the researchers want to discover or reestablish certain relationships. The orientation might be to solve immediate management issues, discover something new, or have purely academic intentions. For instance, in an organization, the manager may want to know the reason for high attrition and lack of job satisfaction, or a retail store may survey the post-purchase satisfaction among the customers.

2.1 Research Problem Identification

Defining the research problem is the first step in the research process. The researchers get the proper direction to conduct their research by first understanding the research problems. Hence, a well-defined research problem is crucial. When the problem is discovered, researchers and management can take further steps to define the problem clearly and precisely. A problem defined with accuracy and conscience helps the researchers utilize the available resources effectively. It is imperative for researchers to explore what exactly is the problem and what are the objectives of the research. The rule generally followed to define the research problem is that the definition should permit the researchers to acquire all details required to address the managerial issues and show guidelines for finding a solution. The researcher should be careful not to define the problem too broadly or narrowly. Examples of broad managerial problems are defining a strategy for enhancing organizational performance and a strategy to elevate the organization’s brand equity. An example of a narrow definition of a problem is how to match competitors’ recruitment strategies. To overcome the possibility of both errors while defining the research problem, the researchers must define the problem with broad, popular terms and devise its components. The broad general statement helps the researchers get a sound perspective on the research problem and avoid the error of defining the problem narrowly. On the other side, the specific component helps to identify the key aspects of the research problem, extend a transparent guideline to proceed further and avoid the error of defining the problem too broadly. In management and social science research, broad management problems need to be converted to information-oriented research problems that focus more on the cause than the symptoms. Some examples of managerial problems converted to research problems are presented in Table below. The conversion of management dilemma to managerial questions and further to research questions can be carried out through exploratory research. Such research incorporates an examination of past research studies, a review of extant literature and organizational records and interviewing experts (Cooper et al., 2016).

2.2 Literature Review

Exploring the existing literature is critical in the research process. Researchers must explore and investigate extant literature to observe whether other researchers have already addressed the identified research problem. A literature review is a systematic search of published work, including periodicals, books, journal papers (conceptual and empirical), and reports, representing theory and empirical work about the research problem and topic at hand. A survey of existing literature is customary in applied research and is an elementary requirement of a basic research report. The internet, electronic databases, websites, and e-library help the researcher to carry out literature surveys systematically and easily.

The literature review aims to study the existing state of knowledge in the domain of interest, to picture key authors, theories, methods, topics, and findings in that domain, and to explore the gaps in knowledge in that domain. A literature review conducted systematically reveals whether initial research questions have already gained substantial attention in the extant literature, whether more interesting newer research questions are available, whether past studies have consistent findings or contradictions exist, flaws in the body of research that the researchers can address, and whether the initial research questions need to be revised as per the findings of the literature review. Furthermore, the review can answer the proposed research questions and help identify theories used in previous studies to address similar research questions. For example, for an organization interested in determining the true cause of turnover, the researcher will study extensively the existing literature on attrition and its causes. By studying relevant journal articles, books, and book chapters, the researcher will discover the causes of attrition in general, find out the existing gaps, and suggest the management carry forward the research to find causes specific to the organization.

As deductive research primarily involves theory testing, the researchers must identify one or more theories that can illuminate the proposed research questions. Through an extensive literature review, researchers may uncover various concepts and constructs related to the phenomenon of interest. A theory will extend support to constructs/variables that are logically relevant to the chosen phenomenon. In the deductive approach, researchers use theory/theories as the logical basis for hypothesis testing. However, researchers must carefully select the theories appropriate for the identified problem to be studied. The hypotheses need to be logically formulated and connected to the research objectives.

2.3 Formulating Research Questions

After problem identification and clarification, with or without an exploratory research approach, the researchers should derive the research objectives. Cautious attention to problem definition helps the researchers devise proper research objectives. Research objectives are the goal to be achieved through research. The research objective drives the research process further. A well-devised research objective enhances the possibility of gathering, relevant information and avoiding unwanted information. The research objectives can be properly developed with the consensus of the researchers and management on the actual managerial and business problems. The researcher should ensure that the research objectives are clearly stated, appropriate, and will yield germane information. The research objective may involve exploring the likelihood of venturing into a new market or may necessitate examining the effect of a new organizational policy on employee performance. The nature and types of objectives lead to choosing an appropriate research design.

Research Objectives: Research objectives represent the goal of the research the researchers want to accomplish.

2.3.1 Suitable Research Questions

Research questions are important to conduct effective research. Without a clear research question, the researcher may face the risk of unfocused research and will not be sure of what the research is about. Research questions are refined descriptions of the components of the research problem. These are questions related to behavior, events or phenomena of interest that the researchers search for answers in their research. Examples include what factors motivate the employees in an organization to apply the gained knowledge back to their jobs or what needs to be done to enhance the creativity of school-going students. Research questions can best state the objectives of the research. Each component of the research problem needs to be broken down into sub-parts or research questions. Research questions inquire about the information essential concerning the problem components. Properly answered research questions will lead to effective decision-making. While formulating research questions, researchers should be guided by the problem statement, theoretical background, and analytical framework.

Sources of Research Questions

• Extant Literature
• Personal experience
• Societal issues
• Managerial problems
• New theories
• Technological advancement
• Empirical cases
• Contradictory finding

2.3.1.1 Significance of Research Questions

Research questions are critical because they guide scientific and systematic literature search, the decision about appropriate research design, the decision about data collection and target audience, data analysis, selection of right tools and techniques and overall to move in the right direction.

The researcher can utilize different sources for formulating research questions, such as extant literature, personal experience, societal issues, managerial problems, new theories, technological advancement, and contradictory findings. The research question must portray certain attributes. Research questions in quantitative research are more specific compared to qualitative research. Sometimes, some qualitative research follows an open approach without any research questions. The main steps involved in formulating research questions are illustrated in Figure below.

Criteria of Effective Research Questions

• Simplicity
• Rateability
• Novelty
• Systematic and logical
• Significant
• Fascinating
• Logical association among variables

The sequence in selecting research questions suggests that the researchers are engrossed in a process of progressive focusing down when developing the research questions. It helps them to slide down from the general research area to research questions. While formulating the research questions, the researchers should understand that ending a research question with a question mark is essential. Without a question mark, a statement cannot be considered as a research question. It is quite possible that the researchers may not get answers to all research questions. The research questions need to be related to each other.

2.4 Planning the Research Design

After formulating research problems and literature surveys, the next stage in the research process is to develop the research design. Research design is the blueprint of research activities to answer research questions. It is a master plan that includes research methods and procedures for gathering and analyzing the relevant information with minimum cost, time, and effort. A research design extends a plan for carrying out the research. The researchers need to decide the source to collect information, the techniques of research design (survey or experiment), sampling techniques, and the cost and schedule of the research. The success of these objectives depends on the purpose of the research. Usually, research purposes are segregated into four types: exploration, description, diagnosis, and experimentation.

There are varied designs, such as experimental or non-experimental hypotheses testing (details of different research designs are outlined in section 2.3 in this chapter). There are four primary research methods for descriptive and causal research: survey, experiments, secondary data, and observations. The selection of an appropriate research method relies on the research objectives, available data sources, the cost and effort of collecting data, and the importance of managerial decisions. If the research objective is exploration, a flexible research design can extend better opportunities to investigate different aspects of the research problem. On the other hand, if the intention is simply to describe any situation or phenomena of interest to examine the relationship between two or more variables, the appropriate design should prioritize minimizing bias and maximizing reliability in data collection and analysis. For example, suppose a researcher wants to conduct exploratory research to know the different types of arthritis common in India. In that case, it may require a flexible design relying on secondary data from hospital records or discussions with doctors or other experts to reach conclusions. However, to invent COVID-vaccination and medicine for the COVID-19 virus, the researchers conducted varied experiments to reach a conclusion.

2.4.1 Hypotheses Development

Exploratory research helps the researchers define the research questions, key variables, and theoretical underpinnings and formulate hypotheses if required in the research. The hypotheses must be logically derived based on the research questions and linked to research objectives. A hypothesis is a tentative proposition regarding a research phenomenon. It may be a tentative statement that indicates an association between two or more variables, guided by any supportive theory, theoretical framework, or analytical model. It is a viable answer to the research questions framed by the researchers. Hypotheses are statements of relationships or propositions that are declarative and can be tested with empirical data. Some examples are:

H₁: Training influences organizational performance.

H₂: Training enhances employee performance.

For two more research questions i.e., “to what extent does brand love determine purchase intention?” and “does age and family background moderate the relationship?”, the hypotheses are:

H₁: Brand love is related to purchase intention.

H₂: Age and Family status moderate the association between brand love and purchase intention. Figure below provides a pictorial representation of the hypotheses drawn.

However, it is not always feasible for researchers to formulate hypotheses in all situations. Sometimes, researchers may lack all relevant information, and theoretical support may not be available to formulate the hypotheses.

2.5 Sampling Design

This stage of the research process involves an investigation of the population under study. A complete investigation of the population under study is known as a census inquiry. Usually, in census investigation, all units or items of the population are studied with high accuracy and reliability. However, it is usually not practicable and feasible for the researchers to study the entire population. Researchers usually prefer to investigate small, representative subgroups from the population known as sample. The procedure to select the sub-groups/samples is called sampling design. Sampling entails the process of drawing conclusions based on a subgroup of the population. Hence, the sample is a subset of the population. The first question that needs to be addressed in sampling is “who is to be included in the sample?” and this requires the identification of the target population under study. It is difficult for the researcher to define the population and sampling unit. For example, if a researcher wants to investigate the financial savings and vehicle loan association survey. In that case, individuals with existing accounts will be taken, and this sample unit represents the existing customers and not the potential customers. Hence, it is critical in sampling design to determine the specific target population.

Secondly, the issue that concerns the researchers in sampling design is selecting an appropriate sample size, and the third concern is selecting the sampling units. Researchers need to address these concerns to justify the research. Samples can be selected either using probability sampling techniques or non-probability sampling techniques. There are four types of probability sampling such as simple random, systematic, stratified, and cluster sampling. Non-probability sampling includes convenience, judgmental, quota, and snowball sampling. Depending on the objective, researchers should select the appropriate sampling techniques for their study.

2.6 Fieldwork and Gathering Data

After the formalization of the sampling plan, the fieldwork and data-gathering stage begins. The researcher gathers data after finalizing what to research, among whom, and which method to use. Data gathering involves the process of information collection. Different data collection instruments are available for researchers to collect information or data. Broadly, there are two ways to collect data, such as primary and secondary data collection methods. Primary data include data collected firsthand and are original. Varied methods are available for primary data collection, such as structured and unstructured interviews, focused group discussion, observation, and survey using a structured questionnaire. The data can be collected offline or online. Secondary data included information collected from published or unpublished sources that were already available. Some secondary data collection sources are articles, magazines, company records, expert opinion survey data, feedback of customers, government data, and past research on the subject. For example, to conduct a survey of job satisfaction in an organization, the researcher may circulate a printed questionnaire offline or mail the questionnaire to the selected respondents following an appropriate sampling technique.

Another example could be a study that investigates the purchase preference for luxury cars, and the base model demands primary and empirical information. However, another study that intended to describe the financial investment behavior of existing customers will use secondary data. At this stage, the researchers need to ensure the reliability and validity of the data obtained for the study.

2.7 Data Processing and Analysis

After data gathering, the data needs to be converted or properly coded to answer the research question under study. The information gathered in the data collection phase should be mined from the primary raw data. Data processing starts with data editing, coding, and tabulation. First, it is vital for the researchers to check the data collection forms for missing data, clarity, and consistency in categorization. The editing process involves problems associated with data, such as respondents’ response errors. Editing improves the quality of the data and makes the data usable for tabulation, analysis, and interpretation. Tabulation is a technical process in which classified data are presented in tables. Researchers use computers to feed data to a computer spreadsheet for data analysis. The preparation of a spreadsheet also requires lots of expertise and experience.

After coding the data, the next step is to analyze the data. Data analysis is the utilization of reasoning to make sense of data gathered. Ample statistical techniques are available for the researchers to analyze the data. Based on the research questions, objectives, study types, sampling framework used, data types, and degree of accuracy involved in the research, one can choose from parametric or non-parametric techniques for data analysis. Researchers may adopt univariate, bi-variate or multi-variate methods for data analysis. The analysis may include simple frequency analysis, multiple regression, or structural equation modeling. Different techniques are available for qualitative data, presented in Part 3 of this book.

2.8 Drawing Conclusion and Preparing a Report

After data analysis, the final stage in the research process is the interpretation of the results. The researcher requires analytical skills to interpret the statistical results, link the output with the research objectives, and state the implications of the result.

Research Design: Research design is the blueprint/systematic steps to carry out research smoothly

Finally, researchers must communicate the result in the form of a report. The preparation of the final report needs to be done with the utmost care. The final report should include the identified research questions, research approach, data collection method, data analysis techniques, study findings, and implications for theory and practice. The structure of the report will be discussed in the last section of this book. The report should be prepared comprehensively to be usable by management or organizations for decision-making.

3. Classification of Research Design

This section highlights the classification of research design. As mentioned in the previous section, research design is the framework for carrying out management and other research. After the identification of a problem, the researchers formulate the research design. A good research design ensures the effectiveness of the research work. The choice of selecting an appropriate design relies on the research objectives. The broad categorization of research design with sub-categorization is detailed in various sub-sections.

3.1 Exploratory Research Design

Methods to Conduct Exploratory Research

• Literature survey
• Secondary sources of data
• Experience survey
• Focused group discussions
• Observations
• Structured and unstructured interviews
• Pilot surveys
• Case Studies

Exploratory research design is the simplest form of research design. The researchers explore the true nature of the problem. When researchers aim to study a new area or examine a new interest, exploratory design is a good option. This research design is flexible and versatile in approach. The information required by the researchers is defined loosely and unstructured. Researchers carrying out qualitative research usually adopt exploratory research design. Exploratory research design serves three purposes (a) it helps the researchers to address their inquisitiveness and quest for better understanding (b) to assess the practicality of carrying out border research (c) and devise methods for further studies.

Methods to Conduct Descriptive Research

• Self-administered survey
• Phone survey
• Mail survey/online survey
• Observation
• Personal interview
• Telephone interview

Exploratory research design has paramount significance in management and social science research. They are crucial for researchers who want to study something new. To cite an example, during the COVID-19 pandemic, physical health, mental health, and safety of school and college-going children were a concern for most people. The online education system was the new normal at that time. Research studying the impact of digitalization, long time spent in online studies on students’ health and mental well-being during the COVID-19 pandemic, is of an exploratory kind. One of the disadvantages of exploratory research design is that researchers rarely get specific answers to the research questions.

3.2 Descriptive Research Design

The prime objective of descriptive research design is to describe certain situations or events. This type of design provides an extensive explanation of the research phenomena under study. In descriptive research, the researchers possess prior knowledge about the problem situations. The information is defined with clarity. This kind of research is preplanned and more structured than exploratory research. Researchers must formulate research questions properly and have clarity regarding the types of data needed and the procedure to be followed to achieve the research objectives. Researchers have the luxury of covering a large representative sample. Researchers must answer five Ws and one H – what, who, when, where, why, and how of research issues. What kind of information is required for the research, who are the target respondents, when the information will be collected, where to interact with the respondents, why information is collected from the respondents and how to collect data from the respondents. Descriptive research studies can be cross-sectional or longitudinal. The major objectives for the following descriptive research are given below.

• To explain the characteristics of certain groups such as the Indian population, employees, students, marketing personnel, organizations, sales persons. For example, a university to design a customized online higher studies course for working professionals needs a holistic profile of the interested population.
• To evaluate the portion of individuals in a specific population portraying a typical behavior. For instance, when a researcher is inclined to know the percentage of employees not interested in an online platform introduced for them in their organization.
• To predict for future. For instance, to know the future of physical retail stores due to the widespread expansion of online stores.
• To examine the extent to which management research variables relate to each other. For example, to what extent does work-life balance, salary, and conducive work environment enhance employee job satisfaction?

3.3 Causal Research Design

Usually, causal research design is adopted by researchers to explain causal relationships among phenomena under study. Causal research examines cause-and-effect relationships among variables. Causal research has certain criteria, as already discussed in Chapter 1. Causal research follows a planned and structured design like descriptive research. Though the magnitude of the relationship among variables is examined in descriptive research, the causal association cannot be explained through such research. Experimentation is one of the methods for carrying out causal research.

In causal research, the researchers usually examine the impact of one variable on another. The researchers try to explore the cause-and-effect relationship (nomothetic explanation). How can the researcher know whether cause and effect are associated? There are three criteria for a nomothetic causal relationship when (1) two or more variables are correlated, (2) the cause precedes the effect and (3) the absence of a plausible alternative explanation for the effect other than the proposed cause (Babbie, 2020). First, without establishing a correlation among two or more variables, causation cannot exist. Second, the cause should happen before the effect in time. For instance, it is more sensible to say that children’s religious affiliation is caused by their parents than to reflect that parents’ religious affiliation is due to children; even in some cases, it is plausible that children may convert to other religions later with their parent’s permission. The third significant condition for a causal relationship is that the effect cannot be attributed to any external third variable for establishing causation.

To cite one classic example, there is a causal association between sales of ice cream and death owing to drowning. Intake of more ice creams in summer does lead to a higher death rate due to drowning. The third intervening variable that causes higher death is season or temperature. In summer, higher deaths occur due to swimming and not because of taking ice-creams. The intervening variable season or temperature causes a higher death rate.

To establish a reliable causal relationship among two or more variables, other influencing variables must be controlled to neutralize their impact on the studied variables. For example, to study the effect of factors influencing training transfer in soft skill training, the other intervening variables such as age, gender, and educational qualification need to be controlled. This kind of research sometimes demands experimentation to establish causality. In most cases, causal research is quantitative and needs statistical hypothesis testing.

3.4 Experimental Research Design

Experimental research aims to examine the cause-effect relationship in a controlled setting by isolating the cause from the effect in time. The three criteria suggested by John Stuart Mill mirror in experimental research. In experimental research, the cause is administered to one group of subjects, known as the treatment group and not to the control group, and the researchers observe the difference in mean effect among the subjects of both groups. Whether variation in the cause is connected to variation in effect is observed. To be more specific, the researcher manipulates the independent variable and examines the change in the dependent variable, keeping other variables constant. Researchers used varied methods during the experiments to reduce the plausible effect of other explanations for the effect, along with ancillary methods to investigate the plausibility of those that cannot be ruled out. It is vital in experimental studies to control the extraneous and confounding variables while carrying out the experiments. Ignorance of such variables may lead to spurious relationships among studied variables. However, bringing many of the variables under experimental control is impossible. For example, personal characteristics of the subject like age, sex, intelligence, beliefs and persona. In such cases, the researchers must observe natural variations in the variables of concern. Then, statistical procedures are used to rule out the plausible impact of uncontrolled factors.

Experimental Research Design: An experiment is a method of collecting evidence to indicate the effect of one variable
on another.

Experimental research design can be conducted in a laboratory setting (laboratory experiment) or in a field setting (field experiments) where the phenomena of research interest happen. As an example, one of the most talked about and controversial experiments conducted on understanding human behavior has been the Stanford Prison Experiments, which took place at Stanford University in 1971. The experiments were funded by the US Office of Naval Research, and the principal investigator for the same was Prof Phillip Zimbardo. The major purpose of these experiments was to understand how norms develop and social expectations about roles shape group behavior. Experimental studies are segregated into four categories such as pre-experimental, true-experimental, quasi-experimental and statistical design.

3.4.1 Correlation, Causation and Cofounds

Correlation cannot be treated as causation, and correlation does not always prove causation. In correlation, it is unclear which variable comes first or whether any alternative explanation exists for the assumed effect. Two variables may be correlated due to chance. Correlation is symmetric, while causation is asymmetric. Two variables may be co-related, but their relationship may be affected by a third variable called cofounds. For example, let’s say that high salary and high educational qualifications are correlated. It is difficult to say with confirmation which comes first. Whether a high educational qualification leads to a high salary, or a high salary leads to a high educational qualification. Both possibilities can hold true and necessitate further investigation. Until researchers conclude through their investigation, a mere correlation among these two variables will not give a clear picture of their causal relationship. There is also the possibility of an alternative explanation for the relationship between high salary and high educational qualifications. The link may be due to a third variable called intellect, which results in high salary and high educational qualifications.

In management research, social science, and natural science, three significant pairs of components are required for experimentation: Experimental and control group, independent and dependent variable, and pre-test and post-test.

3.4.1.1 Experimental and Control Group

The group in which an experimental treatment is administered is known as the experimental or treatment group. In contrast, the group in which no experiment is administered is known as the control group. Using control groups enables the researchers to assess the experiment’s effects. For example, suppose a researcher wants to study the impact of rewards on employee productivity in an organization. In that case, the researcher can experiment with two groups of employees. One group will be given external rewards, known as the experimental group, and the other group (control group) will provide no external rewards. Then, the researcher can investigate the causal association between rewards on employees’ productivity through this experiment. The use of a control group is quite common in medical science research. In social science and management research, the use of control groups and experimental studies became popular with several experiments conducted in the late 1920s and early 1930s by F. J. Roethlisberger and W. J. Dickson (1939) to discover the changes required in working conditions to enhance employee satisfaction and productivity. Their series of experiments resulted in the Hawthorne effect.

3.4.1.2 Independent and Dependent Variables

In experimental research, the researchers study the impact of an independent variable on the dependent variable. Usually, experimental stimuli, whether present or absent, are considered independent variables. Independent variables are manipulated in the study, and their effects are assessed and compared. The researchers compare outcomes when the stimulus is present and not present. Hence, the independent variable is the cause, and the dependent variable is the presumed effect. It is to be noted that the independent variable in one study may serve as a dependent variable in another study. For example, an experiment intends to explore the causality between high salary and job satisfaction, job satisfaction is the dependent variable. However, in another experiment designed to explore the causality between job satisfaction and employee productivity, job satisfaction is the independent variable.

3.4.1.3 Pre- and Post-test

In an experiment, the experimenters measure the variable before conducting the experiment on the group known as the pre-test and measure the variable after conducting the experiments is called as post-test. Hence, subjects are exposed to a stimulus called a dependent variable (pre-testing), then exposed to a stimulus, i.e., independent variable, and again assessed with a dependent variable (post-testing). Any discrepancies between the two measurements of dependent variables are ascribed to the independent variable.