10 Research questions and testing hypotheses

Soumyajit Patra

1. Objective

This module will teach you about the importance of research questions and hypothesis in social science research. At the end of this module, you will find some digital resources and a bibliography for your further study.

2. Introduction

Knowledge is contextual and much of it depends on agreement. It is contextual as it has a time-space dimension – knowledge varies from time to time, region to region and from society to society. Take the example of the imperishable plastic bags. These were thought to be very useful even a few years back. But today its harmful effects on environment are known to all and we have started thinking the other way round. The old knowledge has been replaced by the new one.

Knowledge is a matter of agreement as well. A significant portion of what we know is a matter of agreement and belief. Little of it is based on our personal experience and discovery. As we grow up through the process of socialization, we learn to accept (to take it for granted) what everybody around us already knows. If we start questing everything we are taught and try to test instead of accepting what is given, life would be impossible to bear with (Babbie 2004: 5). We have to learn where and how to raise ‘questions’ in our everyday life. This is equally true in case of scientific knowledge. We need to clarify, in this context, the distinction between scientific knowledge and common sense as well as the purpose of scientific research.

2.1 Scientific Knowledge and Common Sense

The distinction between scientific knowledge and common sense would be relevant here. The former is based on logic and is verifiable. The foundation of scientific knowledge is systematic and critical questioning, observation and reasoning. But as Majumdar (2005: 10) defines it, common sense does not ‘take us beyond what are observable. It limits us to events and conclusions that are widely believed as true.’ So, for obvious reasons, knowledge gathered through scientific inquiries may oppose the common sense. Social scientists often emphasize upon the explanatory nature of science that, to a large extent, involves refined and fundamental questioning of the existing knowledge.

2.2 Scientific Research

The purpose of scientific research is to modify or contribute to the existing stock of knowledge through proper inquiry directed by properly framed research questions and reasoning. It is widely believed that there is no ‘absolute’ in science. Scientific knowledge is inclusive and is always open to further investigation and revision. In the words of Das (2004: 21), ‘research may be described as systematic and critical investigation of phenomena towards increasing the stream of knowledge.’ In a similar way Majumdar (2005: 25) writes: ‘The obvious function of research is to add new knowledge to our existing store. Therefore, scientific research is a cumulative process. Since new insights are obtained into the problem investigated, we need to review or modify our earlier beliefs and postulates’.

3. Learning Outcome

This Module will help you understand different types of research questions and hypotheses that give rise to reliable scientific knowledge. You will also learn how to formulate them.

4. What are Research Questions?

Researchers have many queries and curiosities in their mind and they try to reach at some satisfactory and valid answers and solutions of these after a careful and meticulous analysis of the relevant data collected through appropriate methods. Research questions are specific questions framed during the initial phase of the research, the answers of which a researcher tries to find out. The research questions set the direction of the entire research process. We can argue following Bryman (2012: 9) that

“A research question is a question that provides an explicit statement of what it is the researcher wants to know about. A research purpose can be presented as a statement … but a question forces the researcher to be more explicit about what is to be investigated”.

4.1 How can Research Questions be framed?

Mere selection of research topic does not direct a researcher to the actual methods to be followed and the specific areas to look at for collecting data. As Patrick White (2009: 33) has argued,

“It is usually much easier to decide upon a topic or area of interest than it is to produce a set of well-structured questions”.

It may be worth noting here that research involves certain definite stages and a researcher starts framing research questions and hypothesis after selection of topic and review of existing literature. The following diagram shows the stages of research before and after the research questions:

Selection of research topic, which is the elementary task of any research, is, however, not an easy task. One has to go through the existing literatures to find out gaps in research. Indian Council of Social Science Research (ICSSR) however publishes trend reports of research done on important themes in sociology and makes us aware of what has so far been done and what needs to be done (Singh 2013). Despite such literature, a researcher has to be very precise in focusing his or her attention while framing research questions. When, after going through the existing literatures on the concerned area, the researcher finally specifies the objectives of the study, he or she is better able to frame his or her research questions. Research questions can also be framed on the basis common sense.

• De Vaus (2002) has provided us with some examples that can guide us in developing research questions particularly for descriptive researches. These are:
• What is the time frame of our interest?
• What is the geographical location of our interest?
• Is our interest in broad description or in comparing and specifying patterns for subgroups?
• What aspect of the topic are we interested in?
• How abstract is our interest?

The research questions of explanatory studies mostly focus on delving the causal relationships between different variables. Naturally the ‘why-questions’ are more important in explanatory researches than the ‘what-questions’, which forms the basis of descriptive studies. According to Babbie (2008: 99), descriptive studies answer questions of what, where, when and how, exploratory studies questions of why? However, the suggestions of Ramkrishna Mukherjee (1993) can be helpful for formulating research questions for any type research in social sciences.

Calling his approach as ‘inductive-inferential method’ Mukherjee (Ibid.) argues that a social scientist should try to find out the answer of the following questions:

What is it?

How is it?

Why is it?

What will it be?

What should it be?

For obvious reasons when a researcher deals with the first two questions, i.e. ‘what is it?’ and ‘how is it?’ the orientation of her research is descriptive and classificatory (see also Bose 1997). As soon as she incorporates the question ‘why is it’, her work becomes more explanatory in its spirit. When a social scientist’s research questions include the fourth and fifth questions as well, it becomes a diagnostic study. There can be a reasonable debate among the positivists regarding the inclusion of the fifth question as they believe that the questions like ‘what should it be?’ involve value judgements. However, you canunderstand that a comprehensive research should be based on all the questions mentioned above. In many cases, the researchers deal with a number of research questions, but do not clearly state which questions are more important, or how the questions are related. Such a multiplicity of questions can lead to the problem of lack of focus (Andrews 2003). The researchers should select the number of research questions for her or his study considering the time-cost-labour components of the work. Time, labour and cost of the study would proportionately be increased with the increase in the number of research questions. Too many research questions are difficult to manage as well.

4. 2 Features of Research Questions?

Good research questions that lead to proper research findings have some important features (White 2009; De Vaus 2002; Andrews 2003). The following are the most important among them:

 Research questions should be interrogative – Research questions should be interrogative in nature, it should not be declarative. For example it should be like this: ‘What is the relationship between educational level and attitude towards the freedom of media?’ A statement like: ‘There may be some relationship between educational level and attitude towards the freedom of media’ is not a research question.

 Research questions should be based on the objectives of the study – Research questions should not divert the attention of the researcher from the basic objectives of the study. It should rather try to delve deep into the problem.

 Research questions should be specific – There should not be any ambiguity in the research questions. It should be easily understandable and precise as much as possible.

 It should be simple but well-structured – Much of the success of a research depends on the research questions. It should be focused and precisely framed. The ‘fallacy of many questions’ (i.e. aiming at ‘more than one inquiry in a single question’) should be avoided (White 2009). The questions should be structured in such a manner that they help the researcher unveil a specific dimension of the problem.

Self-check Exercise -1:

1. What do you mean by research?

Research is a scientific process of inquiry by which the existing stock of knowledge is either enriched or modified.

2. Distinguish between scientific knowledge and common sense.

Scientific knowledge is based on logic and is verifiable. The foundation of scientific knowledge is systematic and critical questioning, observation and reasoning. But common sense is gathered from direct day to day experience. Although common sense is not gathered through scientific inquiry it can be helpful in many research works.

3. What are research questions?

Research questions are specific questions emerged out of the broad problems of research, the answers of which a researcher tries to find out. The research questions set the direction of the entire research process.

4. What are the features of good research questions?

Research questions should be interrogative. It should not be a statement. A research question should also be specific, understandable and well-structured. Good research questions are based on the objectives of the study.

5. Distinguish between descriptive and explanatory research.

Descriptive research tries to describe a phenomenon or a situation or a problem. It generally deals with ‘what’ and ‘how’ questions. Explanatory research, on the other hand, tries to explain the ‘cause-effect’ relationships between different variables. This type of research also involves ‘why’ questions along with the ‘what’ and ‘how’ questions.

5. Research questions and hypothesis

Both research questions and hypotheses are useful in social science research. According to White (2009), the difference between them is that while research questions are interrogative in its form, hypotheses are declarative statements which are intended to be tested during the course of research. Hypotheses can be restructured in the form of questions. But then one should not call it hypothesis.

5.1 What is hypothesis?

When a researcher conceptualizes her research problems, she thinks about it in general terms. Research questions or hypotheses help look at the specific aspects of the problem. So hypotheses or research questions enable us to carry out meaningful analysis. Hypotheses are specific statements about the problem made at the initial stage of the research, which may be proved right or wrong also include things such as households, cities, organizations, schools, and nations. If an attribute does not vary, it is a constant” (Bryman 2012: 48).

Once a hypothesis assumes a relationship between two or more variables, the validity of such assumption, made on the basis of the personal experiences, knowledge and insights of the researcher, is tested through suitable statistical techniques. Hence, hypothesis is ‘[a]n informed speculation, which is set up to be tested, about the possible relationship between two or more variables’ Bryman (2012: 712). If the primary assumptions are proved correct after the analysis of data, they become part of the theory. So it is said that ‘hypothesis provides the link between the empirical world and the theory’ (Majumdar 2005: 78). Hypothesis formulation and testing are closely associated with the quantitative approach to study social phenomena (Jupp 2006).

5.2 Features of a good hypothesis

‘A hypothesis is a specified testable expectation about empirical reality that follows from a more general proposition’ (Babbie 2004: 44). It is the assumption made about the relationship between different variables on the basis of existing knowledge or common sense. But all declarative statements or assumptions are not hypotheses. Let us discuss some examples:

Examples:

1. ‘The rate of dropout is higher among the girl students’.

2. ‘The rate of dropout varies with gender with the girl students having a higher dropout rate’.

The first assumption is not an example of a good hypothesis as it does not clearly state the two variables. But, the second one is a better one because it clearly mentions gender and rate of dropout as two variables and a relationship between them is assumed.

The features of good hypotheses are as follows:

• Hypothesis generally states (predicts) the relationship between two variables.
• It is expressed as a statement and not as a question (Payne and Payne 2005: 112)
• Hypothesis should be clearly stated, specific and conceptually clear.
• It should be consistent with the known laws of nature (Majumdar 2005)
• Hypothesis is testable (after the final analysis it may prove to be correct or incorrect).

5.3 Soures of hypotheses

Hypotheses are not ordinary or casual statements about the empirical reality. They emerge through a systematic and logical process. According to Goode and Hatt (1981), there are four possible sources from which hypotheses can emerge. These are:

 Culture can furnish hypotheses – Every human society has some distinctive cultural traits. Many social science researches focus on human behaviour or on meaningful social actions. Folkways, mores, values, customs, belief patterns can help formulate hypotheses in these studies. at the end of the analysis (Henn et al 2006).

Hypotheses are formulated at the third stage of the research process (see Diagram 1). According to Goode and Hatt (1981: 56), ‘[a] hypothesis states what we are looking for.’ They write ‘[i]t is a proposition which can be put to a test to determine its validity.’ Hypotheses are primary assumptions about the interrelations of different variables which set the direction of the entire research process. It may be noted that “a variable is simply an attribute on which cases vary. ‘Cases’ can obviously be people, but they can

Hypotheses can emerge from the science itself – In the backdrop of any research there should be one or more theories. Hypotheses are often deducted from a theory to verify or modify some of its basic conclusions. Goode and Hatt (ibid.) opine that the socialization process, that a student of a particular discipline undergoes, teaches her/him about the promising areas, paradigms, laws, analytical categories, concepts and methods of that particular discipline. This knowledge can help the student to assume some possible causal relationships between some variables that she or he can put to a test for verification.

 Hypotheses can be formulated from analogies – Analogies between human society and nature, between two different types of communities are often a fertile source of hypotheses. For obvious reasons, the researcher should take care in making such analogies. Analogies should not be illogical, it should, on the other hand, be consistent with the known laws of nature.

 Hypotheses can come out from idiosyncratic, personal experiences of the researcher – The scientist lives in a particular culture or she can encounter some cultural traits of some other cultures. Her personal experiences can help her formulate effective hypotheses.

5.4 Types of hypothesis

Hypothesis can be classified in many ways. Goode and Hatt (1981) categorize them into three types on the basis of the level of abstraction.

1. pothesis that state the existence of empirical uniformities – Generally these hypotheses are framed when the researchers want to test the ‘common-sense propositions’. In other words, sometimes the researchers are interested to establish the parallels between what people think about a phenomenon and what actually exists. These often lead to the observations of simple differences. In these hypotheses, sometimes, common sense ideas are put into well-defined concepts and then the hypotheses are statistically verified.
2. Hypothesis that is concerned with complex ideal types – These hypotheses try to focus on the logically assumed relationships existing among empirical uniformities. In particular, these hypotheses ‘lead to specific coincidences of observations’ (ibid.: 62). For obvious reasons, these types of hypotheses deal with a higher level of abstraction than the hypotheses that are concerned with the existence of empirical uniformities.
3. thesis that is concerned with the relation of analytical variables – According to Goode and Hatt (ibid.) these hypotheses deal with the highest level of abstraction. In this case, the researcher analytically formulates a hypothesis that shows a relationship between changes in one aspect of the phenomenon with the actual or assumed changes in another aspect.

Majumdar (2005) has categorized hypothesis into two types – eliminative (or analytic) induction and enumerative induction. In the former case hypotheses are formulated as ‘universal generalization’ and the presence of any contrary evidence leads to its rejection. In case of enumerative induction, a complete enumeration is required to accept or reject the hypothesis. Look at the following examples:

Examples

Hypothesis I: Female students score better in Research Methodology course than the male students.

This is an example of eliminative or analytic induction. If any male student is found, who has scored more than the female students, there would be no reason to accept the hypothesis.

Hypothesis II: Ten Percent female students score better in Research Methodology course than the male students.

This is an example of enumerative induction. To accept or reject the hypothesis a complete enumeration is necessary.

5.5 Hypothesis Testing

A researcher formulates a number of hypotheses (sometimes called experimental hypotheses) and all these hypotheses are tested on the basis of data collected for the study. When a researcher wants to test the hypothesis with the help of some statistical techniques, he or she frames what is called null hypothesis. According to Babbie (2004: 49), in connection with hypothesis testing and tests of statistical significance, the hypothesis that suggests that there is no relationship among the variables under study is null hypothesis. Sometimes null hypothesis states that there is no difference between two variables.

Examples

Null Hypothesis (denoted by H0): There is no difference between the percentage of male students and the percentage of female students who have got 60 per cent marks in Research Methodology course.

If the data collected for the study show, for example, that in reality there are differences between the percentage of male students and percentage of female students who have scored 60 per cent in Research Methodology course, there are statistical techniques to determine whether the difference found is statistically significant, or we can ignore the difference attributing it simply to chance factors and accept the null hypothesis (H0). If the difference obtained from the collected data is statistically significant the researcher rejects the null hypothesis and accepts the alternative hypothesis. For obvious reasons there may be more than one alternative hypotheses (denoted by: H1, H2, H3 etc) the researcher has to select any one from among the alternatives if the null hypothesis (H0) is rejected. The following are the examples:

Alternative Hypothesis (H1): There is significant difference between the percentage of male students and the percentage of female students who have got 60 per cent marks in Research Methodology course.

Or,Alternative Hypothesis (H2): The percentage of male students is higher than the percentage of female students who have got 60 per cent marks in Research Methodology course.  Or,

Alternative Hypothesis (H3): The percentage of female students is higher than the percentage of male students who have got 60 per cent marks in Research Methodology course.

It is not always easy to accept a hypothesis from among the alternatives. The researchers often has to find out what is called crucial instance to take a final decision regarding the acceptance of a hypothesis from among a number of options (alternative hypotheses). Sometimes they have to go through an experiment to decide what actually would be the alternative hypothesis (in the above example whether H2 is correct or H3 is correct. It should be noted that both H2 and H3 cannot be correct at the same time.) The experiment which finally helps to come to a final decision regarding which one should be accepted reasonably from among the hypotheses is called experimentum crucis (Babbie 2004; Majumdar 2005). There are a number of statistical techniques like Z-test, t-test, χ2-test etc to test the null hypothesis.

Self-check Exercise – 2:

1. Distinguish between research question and hypothesis.

Both research questions and hypothesis are framed at the initial stage of the research and both help to look at the research problem in a very specific manner. But while the research question is a specific question the answer of which is sought, hypothesis is a declarative statement framed on the basis of the initial assumptions the validity of which is tested with the help of some statistical techniques. Hypotheses are formulated mainly in case of quantitative research.

2. State any two features of a good hypothesis.

A good hypothesis is specific and it indicates the relationship between two variables.

3. What is null hypothesis?

In case of hypothesis testing, the hypothesis that states that there is no difference or relationship between two variables under study is called null hypothesis. It is denoted by Ho. The statistical testing of hypothesis starts with null hypothesis. If the test-result tells the researcher to reject the null hypothesis, the researcher accepts alternative hypothesis.

4. What are alternative hypotheses?

Alternative hypotheses are formulated against the null hypothesis that states that there is some relationship or difference between the variables. These are denoted by H1, H2 etc.

6. Type I error and Type II error

Although in case of quantitative research, the researcher specifies the variables and puts the null hypothesis to test using some statistical techniques, there are dangers of reaching at wrong decisions even if the researcher resort to scientific techniques in the testing of hypotheses. He or she can commit two types of errors – Type I error and Type II error. When the researcher accepts a hypothesis when it is actually incorrect it is called Type I error. In case of Type II error the test-result tells the researcher to reject a hypothesis when it is actually correct. Let us look into the following examples:

Examples

Suppose you are interested to know whether there is any relationship between the education of the mother and that of their daughters. Sociologists generally use χ2 test to determine such relationships statistically. The null hypothesis of such a test would be like this:

H0: There is no relationship between mothers’ education and daughters’ education.

The alternative hypothesis would be:

H1: There is a definite relationship between mothers’ education and daughters’ education.

Suppose the calculated value of χ2 forces the researcher to accept the null hypothesis and come to the conclusion that there is no relationship between the education of the mother and that of their daughters. But in reality these two are highly related. This is the example of Type I error.

Now, suppose you are interested to know about the relationship between distance of home from the nearest high road and the number of children of the married women.

The null hypothesis of such a test would be like this:

H0: There is no relationship between distance of home from the high road and the number of children.

H1: There is significant relationship between distance of home from the high road and the number of children.

Suppose the calculated value of χ2 forces the researcher to reject the null hypothesis and come to the conclusion that there is significant relationship between distance of home from the nearest high road and the number of children married women have. It is not difficult to understand that any such relationship between these two is absurd. This is the example of Type II error.

7. Hypothesis and qualitative research

It has been said that hypothesis is generally associated with quantitative research. But it would be wrong to assume that in qualitative studies, they are irrelevant. According to Jupp (2006), some qualitative researches aim at describing the nature, contexts and consequences of social interactions, social relationships and the process of creations of meanings. These studies also start with some assumptions about the social realities which can be treated as hypotheses. Obviously, these hypotheses do not indicate the relationships between variables. Hypothesis testing in ‘qualitative research is a continuous process, involving the search for cases or contexts that do not square with the assertions being made, rather than a once-and-for-all event’ (Jupp 2006: 138). This is the process of analytical induction and when contrary evidences or what is called crucial instances challenge the conclusions of previous study they are modified or rejected. New hypotheses are then framed in the light of new information or experiences and again their validity is checked.

8. Summary

The topic of any research, which the title of the dissertation signifies, indicates the broad area of research. It is often easy to decide about a topic of research. But, a researcher has to be precise in focusing his or her attention while framing his/her research questions. Research questions and hypotheses are framed to specify the areas in which the researcher concentrates. Research questions are interrogative whereas hypotheses are declarative statements. When the researcher finalizes the specific objectives of the study, he or she is better able to frame his research questions or hypotheses. The researcher tries to find out the answers of the research questions framed at the beginning of the study. Hypotheses or the assumptions.

References

• Andrews, R. Research Questions. London: Continuum, 2003.
• Babbie,  E. The Practice of Social Research. Australia: Thomson Wadsworth, 2004.
• Bose, P. K. “Problems and Paradoxes of Inductive Social Science: A Critique of Ramkrishna Mukherjee”.
• Sociological Bulletin 46, no. 2 (1997): 153-171.
• Bryman, A, Quantity and Quality in Social Research. London: Routledge, 1988.

• …….. Social Research Methods. Oxford: Oxford University Press, 2012. Das, D. K. L. Practice of Social Research. Jaipur: Rawat Publications, 2004.

• Goode, W. J. and Hatt, P. K. Methods in Social Research. Auckland: McGraw – Hill International Book Company, 1981.
• Henn, M.  et. al. A Short Introduction to Social Research, London: Sage Publications, 2006.
• Jupp, V. The Sage Dictionary of Social Research Methods. London: Sage Publications, 2006.
• Majumdar, P. K. Research Methods in Social Science. New Delhi: Viva Books Pvt. Ltd., 2005.
• Mukherjee, R. Systematic Sociology. New Delhi: Sage Publications, 1993.
• Payne, G. and Payne, J. Key Concepts in Social Research. London: Sage Publications, 2005.
• Singh, Yogendra. ICSSR Research Surveys and Explorations: Indian Sociology. Box Set, Vols 1-3. New Delhi: Oxford, 2013.
• Vaus, De D, Surveys in Social Research. London: Routledge, 2002.
• White, P. Developing Research Questions: A Guide for Social Scientists. New York: Palgrave Macmillan, 2009.