Variables, Sampling, Hypothesis, Reliability, and Validity
(Relevant for Sociology optional for UPSC CSE)
Paper-1 ,Unit-3 : Variables, Sampling, Hypothesis, Reliability, and Validity
The Variables:
- A variable is a characteristic that takes on two or more values. It is something that varies. It is a characteristic that is common to a number of individuals, groups, events, objects, etc.
- The individual cases differ in the extent to which they possess the characteristic. Thus, age (young, middle-aged, old) income class (lower, middle, upper), caste (low, intermediate, high), education (illiterate, less educated, highly educated), occupation (low status, high status), etc., are all variables.
- The variables selected for analysis are called explanatory variables and all other variables are extraneous variables. Extraneous variables which are not part of the explanatory set are categorized as controlled or uncontrolled variables.
- Controlled variables, commonly called control variables, are held constant or prevented from varying during the course of the study.
- This is to limit the focus of the research. For example, in age, all males and females under 18 years of age may be excluded from the study. This would mean that the hypothesis is not concerned with specific sub-groups.
Types of Variables:
Dependent and Independent Variables :
- A dependent variable is one which changes in relationship to changes in another variable. An independent variable is one whose change results in the change in another variable. In a controlled experiment, the independent variable is the experimental variable, i.e., one which is withheld from the control group.
- In experiments, the independent variable is the variable manipulated by the experimenter. For example, a teacher wants to know which method of teaching is more effective in the students’ understanding: lecture method, question-answer method, visual method or combination of two or more of these methods. Here, teaching method is independent variable which is manipulated by the teacher. The “effect on students’ understanding” is the dependent variable. The dependent variable is the condition we are trying to explain. In this experiment, besides the methods of teaching, other independent, variables could be personality types (of students), social class (of students), types of motivation (reward and punishment) class atmosphere, attitude towards teacher, and so on.
Experimental and measured variables :
- The experimental variables spell out the details of the investigator’s manipulations while the measured variables refer to measurement. For instance, rural development (measured variable) may be assessed in terms of increase in income, literacy level, infrastructure, availability of medical facilities, availability of social security and so forth.
- In another study on factors affecting student’s achievement (high or low marks), we may examine the absence/availability of books, libraries, good teachers, use of visuals and so on. All these will be experimental variables or experimental manipulations for the researcher. It is important when planning and executing research to distinguish between these two types of variables.
Active and assigned variables :
- Manipulated or experimental variables will be called active variables, while measured variables will be called assigned variables. In other words, any variable that is manipulated is an active variable and variable that cannot be manipulated is an assigned variable.
Qualitative and quantitative variables:
- The quantitative variable is one whose values or categories consist of numbers and if differences between its categories can be expressed numerically. Thus, age, income, sizes are quantitative variables. The qualitative variable is one which consists of discreet categories rather than numerical units. This variable has two or more categories that are distinguished from each other. Class (lower, middle, upper), caste (low, intermediate, high) sex, (male, female), religion (Hindu, non-Hindu) are all qualitative variables.
- Relationships among quantitative variables may be either positive or negative (Singleton and Straits). A positive relationship exists if an increase in the value of one variable is accompanied by an increase in the value of the other, or if decrease in one is accompanied by decrease in the other. In other words, the two variables constantly change in the same direction, e.g., the taller a father, the taller will be his son. The negative relationship between variables exists if the decrease in the value of one variable is accompanied by an increase in the value of the other, e.g., as age increases, the life expectancy decreases.
- Therese Baker has used the terms categorical and numerical variables for qualitative and quantitative variables, respectively. The former (e.g., occupation, religion, caste, gender, education, income) are made up of sets of categories (or attributes) which must follow two rules: one, the categories must be distinct from one another, i.e., they must be mutually exclusive; two, the categories must be exhaustive, i.e., they should cover all the potential range of variation in a variable. After putting himself in the categories of educated (other being illiterate) in the field of education, one can put himself in the sub-category of undergraduate, graduate, postgraduate, etc.
The variables can also be dichotomous or continuous.
- While sex is dichotomous variable, intelligence is continuous variable. Ordinarily, only a few variables are true dichotomies. Most variables are capable of taking on continuous values. Nevertheless, it is useful to remember that it is often convenient or necessary to convert continuous variables to dichotomous or trichotomous variables.
Sampling:
- A sample is a portion of people drawn from a larger population. It will be representative of the population only if it has same basic characteristics of the population from which it is drawn. Our concern in sampling is not about what types of units (persons) will be interviewed/observed but with how many units of what particular description and by what method should be chosen.
- According to Manheim, “a sample is a part of the population which is studied in order to make inferences about the whole population”. In defining population’ from which the sample is taken, it is necessary to identify ‘target population’ and ‘sampling frame’. The target population is one which includes all the units (persons) for which the information is required, e.g., drug abuser students in one university, or voters in one village/constituency, and so on. In defining the population, the criteria need to be specified for explaining cases which are included and excluded.
- For example, for studying the level of awareness of rights among women in one village community, the target population is defined as all women–married and unmarried–in the age group of 18-50 years. If the unit is an institution (say, Vidya Mandir), then the type of its structure, size as measured by the number of students in school section, college section, and in professional courses the number of teachers and employees needs to be specified.
- For making the target population operational, the sampling frame needs to be constructed. This denotes the set of all cases from which the sample is actually selected. It should be noted that sampling frame is not a sample; rather it is the operational definition of the population that provides the basis for sampling.
- For example, in the above example of Vidya Mandir, if students studying in school and in college are excluded, only students of professional courses (MBA, Computer Science, B.Ed., Home Science and Biotechnology) are left out from which the sample is to be drawn. Thus, the sample frame reduces the number of total population and gives us the target population (i.e., students of professional courses only)
There are two objectives of sampling
- Estimate of parameters
- Testing of hypothesis Estimate of parameters:
The major objectives is to estimate certain population parameters (e.g. the proportion of clerk did an office working overtime).Thus, the researches attempts to select a sample and calculate the relevant statistics (i.e. average and proportion. He can use this statistic as an estimate to make a statement about its precision in terms of standard errors and conclude about its population in terms of probability.
Testing of hypothesis: The second objective of sampling may be to test statistical hypothesis about a population (i.e. the hypothesis that at least 60 percent of the household in Kurukshetra town have T.V sets).The researchers may select a sample of household and then calculate the proportion of household possessing T.V sets. The problem now is to assess whether the sample result is such as to reject the hypothesis or whether it supports the hypothesis. To resolve this problem, the researcher has to find out a criterion by which he can determine the precise deviation of the sample result from the hypothetical value..
Purposes of Sampling,
Sarantakos has pointed out the following purposes of sampling:
- Population in many cases may be so large and scattered that a complete coverage may not be possible.
- It offers a high degree of accuracy because it deals with a small number of persons. Most of us have had blood samples taken, sometimes from the fingers and sometimes from the arm or another part of the body. The assumption is that the blood is sufficiently similar throughout the body and the characteristics of the blood are determined on the basis of a sample. Singleton and Straits have also said that studying all cases will describe population less accurately than a small sample.
- In a short period of time, valid and comparable results can be obtained. A lengthy period of data collection generally renders some data obsolete by the time the information is completely in hands. For example, collecting information on the attitudes of voters’ preferences during election period, or demanding action against police personnel responsible for using violence against women demonstrators, or for making a large number of accused persons in the police lockup blind. Besides, opinions expressed at the time of incidence and those expressed after a few months are bound to be different. The findings are thus bound to be influenced if long period is involved in data collection, i.e., not taking a small sample but studying the entire population
- Sampling is less demanding in terms of requirements of investigators since it requires a small portion of the target population.
- It is economical since it contains fewer people. Large population would involve employing a large number of interviewers which will increase the total cost of the survey.
- Many research projects, particularly those in quality control testing, require the destruction of the items being tested. If the manufacturer of electric bulbs wishes to find out whether each bulb met a specific standard, there would be no product left after the testing.
Principles of Sampling:
The main principle behind sampling is that we seek knowledge about the total units (called population) by observing a few units (called sample) and extend our inference about the sample to the entire population. For purchasing a bag of wheat, if we take out a small sample from the middle of the bag with a cutter, it will give us the inference whether the wheat in the bag is good or not. But it is not necessary that study of sample will always give us the correct picture of the total population.
If few people in a village are found in favor of family planning, it would not mean that all people in the village will necessarily have the same opinion. The opinion may vary in terms of religion, educational level, age, economic status and such other factors. The wrong inference is drawn or generalization is made from the study of few persons because they constitute inadequate sample of the total population.
The study of sample becomes necessary because study of a very large population would require a long period of time, a large number of interviewers, a large amount of money, and doubtful accuracy of data collected by numerous investigators. The planning of observation/study with a sample is more manageable.
The important principles of sampling are (Sarantakos):
- Sample units must be chosen in a systematic and objective manner.
- Sample units must be clearly defined and easily identifiable.
- Sample units must be independent of each other.
- Same units of sample should be used throughout the study.
- The selection process should be based on sound criteria and should avoid errors, bias and distortions.
Advantages of Sampling:
The above mentioned purposes and principles of sampling point out some advantages of sampling.
These are:
- It is not possible to study large number of people scattered in wide geographical area. Sampling will reduce their number.
- It saves time and money.
- It saves destruction of units.
- It increases accuracy of data (having control on the small number of subjects).
- It achieves greater response rate.It achieves greater cooperation from respondents
- It is easy to supervise few interviewers in the sample but difficult to supervise a very large number of interviewers in the study of total population.
- The researcher can keep a low profile.
The Significance of Sampling:
There are various reasons for the signifance of sampling in colleting statistical data.
Only Possible, Quick, Economic Method: Perhaps it is the only possible method; it is quick and economic. In a manufacturing unit, quality of products is tested with the help of sample. After testing, if the quality of the product is unsatisfactory, it is reprocessed or scrapped. Thus, there is no alternative to sampling for measuring quality. Likewise instead of observation of all items, selection of a sample from the universe and inferring its characteristics from that sample forms the quick and economical method. It is a highly useful device for the researchers and the practitioners concerned for interring within limits certain characteristics of a population.
Representativeness and Size of Sampling:– Problem of representatives of sample The basic point in the selection of a sample is to ensure that it is as representative of the universe as possible. Explicitly, the size of sample does not necessarily determine its representivess.Thus, if a relatively small sample is scientifically selected , it may be more reliable than an arbitrary selected large sample. The process of sample selection should be such that every items in the population under study has the same chance representative of the population.
A sample which does not represent the population is called biased sample .As Yule and Kendal observes, “the human beings is extremely poor instrument, for the conduct of a random selection. Whenever there is any scope for personal choice or judgment on the part of the observes, bias is almost certain to creep in, The studies based on biased sampling are intrinsically inaccurate and misleading. This is true of several studies in behavioral science which are based on mailed questionnaires involving incomplete and distorted returns. Of course, the original mailing list of prospective respondents any be representative sample However ,the questionnaires actually received may be extremely in view of operation of selective factors.
Problem of Sample Size:-A scientific sample is one which in conjunction with representing the population also consists of enough cases to ensure reliable results. The issue of adequacy of a sample is highly complex. As indicate by Hagood and price ,the size of the sample can be determine by the following items of information :the designation of parameters which one intends to study, the range of reliability permissible in estimates and a cride estimate of the dispersion of studied characteristics.
Types of Sampling:
Two types of sampling: probability sampling and non-probability sampling.
- Probability sampling is one in which every unit of the population has an equal probability of being selected for the sample. It offers a high degree of representativeness.
- Non-probability sampling makes no claim for representativeness, as every unit does not get the chance of being selected. It is the researcher who decides which sample units should be chosen.
Probability Sampling:
Probability sampling today remains the primary method for selecting large, representative samples for social science and business researches. According to Black and Champion, the probability sampling requires following conditions to be satisfied:
- Complete list of subjects to be studied is available;
- Size of the universe must be known;
- Desired sample size must be specified, and
- Each element must have an equal chance of being selected.
It means use some kind of randomization in one or more of their phases. Leabo classifies probability samples in five categories-sample random samples, stratified samples.
Simple Random Sampling:
Although simple random samples are not used widely, they form a basis for other types of sampling. A simple random sample of n items refers to a smple which has been selected from a population in such a manner that each possible combination of n units has the same chance or probability of being selected.
The Advantages Of Simple Random Sampling:-
- Its saves time- As against complete coverage, sampling is cheaper of course, per unit cost is higher.
- It saves labour- Sampling includes a smaller number of staff for the collection, tabulation and processing of the data. Thus it saves labour considerably.
- It saves time-Because of these advantage, sampling was first used with the census of population in 1951.This procedure save a of time.
- It improve accuracy: A sample coverage provides a higher overall level of accuracy. It permits a higher quality of the field, more checks for accuracy, more care editing and the analysis and more elaborate information.
Stratified Random Sampling:
These samples involve division of population into similar groups and selection of a random sample from each other. The population can be divided into groups in the light of the knowledge about it and effect of a certain characteristic group. The population can be divided into groups in the light of the knowledge about it and effect of a certain characteristic upon the estimate to be made.
The Advantages Of Stratified Random Sampling:
This procedure ensures proper representation from each group and probability sample. The basis for division into groups or strata related to the nature of the problem to be studied .For e.g. if the problem involves the estimates of the average income in an area occupational groups can be used as biases for dividing the population. The stratified random sample, if properly carried out, forms improvements upon the sample random sample. Indeed, the reliability of the results for a given size increases with the smaller range of all possible sample averages accordingly, it can said that a properly stratified random sampling is more reliable the a simple random sample of the same size.
Non-probability sampling:
In many research situations, particularly those where there is no list of persons to be studied (e.g., wife battering, widows, Maruti car owners, consumers of a particular type of detergent powder, alcoholics, students and teachers who cut classes frequently, migrant workers, and so on), probability sampling is difficult and inappropriate to use. In such researches, non-probability sampling is the most appropriate one. Non-probability sampling procedures do not employ the rules of probability theory, do not claim representativeness and are usually used for qualitative exploratory analysis.
These samples do not use randomization and can be classified as quota sampling, purposive sampling, accidental sampling, and snowball sampling.
- Quota sampling: It is used in marketing research. It is a stratified sampling but of the non-random type. In this sampling, the population is divided into two or three parts in terms of characteristics. Quota is then fixed up and interviewer is asked a specified number from each division. The interviewer may select a member of the population who is conveniently available. Because of this convenience, bias is likely to color. The bias can be reduced by restricting his convenience. This method is useful where merely rough estimates (rather than results)are needed Indeed, it results are only a rough estimate and thus ,
cannot be tested for reliability. - Purposive Sampling: It involves use of judement and a concerted attempt to obtain representative under the impression of including typical areas or groups in this sample.A study conducted by Namjoshi exemplifies the nature of purposive sample. In this study two types respondants were selected 1.married males and females 2.Unmarried males and females. Both the samples were selected by this procedure in order to get sufficient representation of respondents from higher and lower castes, socioeconomic groups and from both the sexes. A sample of 400 married male and female respondents and a sample of 400 unmarried boys and girls were selected.
- Accidental sampling:– This involves use of available , samples and , is therefore, the weaker type of sampling. This kind of sampling can be used if no other types of sample are available.
- Snowball Sampling:- It related to set of procedure whereby the initial respondents’ are selected by probability methods and thereafter, additional respondents’ are obtained on the basis of information provided by them. This technique is used to identify elements of rare populations by referral. For e.g, a manufacture is interested in marketing a mahogany croquet set for serious adult players as the market for this product is small, the researches is required to use this technique in order to accomplish the task economically.
Hypothesis:
A hypothesis is an assumption about relations between variables. It is a tentative explanation of the research problem or a guess about the research outcome. Before starting the research, the researcher has a rather general, diffused, even confused notion of the problem. It may take long time for the researcher to say what questions he had been seeking answers to. Hence, an adequate statement about the research problem is very important.
- Theodor son and Theodor son, “a hypothesis is a tentative statement asserting a relationship between certain facts.
- Ker linger describes it as “a conjectural statement of the relationship between two or more variables”.
- Black and Champion have described it as “a tentative statement about something, the validity of which is usually unknown”. This statement is intended to be tested empirically and is either verified or rejected. If the statement is not sufficiently established, it is not considered a scientific law.
- Webster has defined hypothesis as “a tentative assumption made in order to draw out and test its logical or empirical consequences”. ‘Test’ here means “either to prove it wrong or to confirm it”. Since statements in Hypothesis have to be put to empirical investigation, the definition of hypothesis excludes all statements which are merely opinions (e.g., aging increases ailments), value judgements (e.g., contemporary politicians are corrupt and have a vested interest to serve), or normative (e.g., all people should go for a morning walk). Normative statement is a statement of what ought to be, not a factual statement that can be shown through investigation to be right or wrong.
Following are a few examples of hypotheses:
- Group study increases higher division achievement.
- Hostlers use more.
- Young girls (between 15-30 years) are more victims of crimes against women than middleaged women (between 30-40 years).
- Lower-class men commit more crimes than middle-class men.
- Suicide rates vary inversely with social integration.
- Educated women have more adjustment problems after marriage than illiterate women.
- Children from broken homes tend to become delinquents.
- Unemployment decreases juvenile delinquency.
- Upper-class people have fewer children than lower-class people.
Criteria for Hypotheses Construction:
Hypothesis is never formulated in the form of a question. Bailey, Becker, Selltiz and Sarantakos have pointed out a number of standards to be met in formulating a hypothesis:
- It should be empirically testable, whether it is right or wrong.
- It should be specific and precise.
- The statements in the hypothesis should not be contradictory.
- It should specify variables between which the relationship is to be established.
- It should describe one issue only.
A hypothesis can be formed either in descriptive or relational form. In the former, it describes events, whereas in the latter, it establishes relations between variables. A hypothesis can also be formed in the directional, non-directional or null form.
Nature of Hypotheses:
A scientific justified hypothesis must meet the following criteria:
- It must accurately reflect the relevant sociological fact.
- It must not be in contradiction with approved relevant statements of other scientific disciplines.
- It must consider the experience of other researchers.
Hypotheses cannot be described as true or false. They can only be relevant or irrelevant to the research topic. For instance, the causes of poverty in a village can be explored in terms of:
- Low development of agriculture (caused by lack of irrigation, sandy soil, erratic rainfall and use of traditional agriculture implements) causes poverty.
- Lack of infrastructure (electricity, roads, markets) causes poverty.
- Barriers in rural development are resource barriers (water, soil, minerals), support barriers (rainfall, irrigation, livestock) and social system barriers (credit, infrastructure, extravagant expenditure and market barriers).
The important hypotheses could be :
- Rural poverty is positively co-related with availability of and accessibility to credit.
- Rural poverty is the result of lack of infrastructural facilities.
- Poverty is associated with extravagant social expenditure.
- Rural poverty is adversely related to resource barriers (water, soil, minerals).
Types of Hypotheses:
Hypotheses are classified as working hypotheses, research hypotheses, null hypotheses, statistical hypotheses, alternative hypotheses and scientific hypotheses.
- Working hypothesis is a preliminary assumption of the researcher about the research topic, particularly when sufficient information is not available to establish a hypothesis, and as a step towards formulating the final research hypothesis. Working hypotheses are used to design the final research plan, to place the research problem in its right context and to reduce the research topic to an acceptable size. For example, in the field of business administration, a researcher can formulate a working hypothesis that “assuring bonus increases the sale of a commodity”. Later on, by collecting some preliminary data, he modifies this hypothesis and takes a research hypothesis that “assuring lucrative bonus increases the sale of a commodity”.
- Scientific hypothesis contains statement based on or derived from sufficient theoretical and empirical data.
- Alternative hypothesis is a set of two hypotheses (research and null) which states the opposite of the null hypothesis. In statistical tests of null hypotheses, acceptance of Ho (null hypothesis) means rejection of the alternative hypothesis; and rejection of Ho means similarly acceptance of the alternative hypothesis.
- Research hypothesis is a researcher’s proposition about some social fact without reference to its particular attributes. Researcher believes that it is true and wants that it should be disproved, e.g., Muslims have more children than Hindus, or drug abuse is found more among upper-class students living in hostels or rented rooms. Research hypothesis may be derived from theories or may result in developing of theories.
- Null hypothesis is reverse of research hypothesis. It is a hypothesis of no relationship. Null hypotheses do not exist in reality but are used to test research hypotheses.
- Statistical hypothesis, according to Winter (1962), is a statement/observation about statistical populations that one seeks to support or refute. The things are reduced to numerical quantities and decisions are made about these quantities, e.g., income difference between two groups: Group A is richer than Group B. Null hypothesis will be: Group A is not richer than group B. Here, variables are reduced to measurable quantities.
Goode and Hatt have given the following three types of hypotheses on the basis of level of abstractness :
- Which presents proposition in common sense terms or, About which some common sense observations already exist or, Which seeks to test common sense statements. For example: Bad parents produce bad children, or Committed managers always give profits, or Rich students drink more alcohol.
- Which are somewhat complex, i.e., which give statement of a little complex relationship. For example:
- Communal riots are caused by religious polarization.
- Growth of cities is in concentric circles (Burgess).
- Economic instability hampers development of an establishment.
- Crime is caused by differential associations (Sutherland).
- Juvenile delinquency is related to residence in slums (Shaw).
- Deviant behaviour is caused by mental disorders (Healy and Bronner).
- Which are very complex, i.e., which describe relationship between two variables in more complex terms, e.g., high fertility exists more in low income, conservatives and rural people than in high income, modern and urban people. Here dependent variable is ‘fertility’ while independent variables are income, values, education and residence, etc. The other example is: Muslims have high fertility rate than Hindus. We have to keep number of variables constant to test this hypothesis. This is abstract way to handle the problem.
Difficulties in Formulating Hypotheses:
According to Goode and Hatt, three main difficulties in formulating hypotheses are:
- Inability to phrase the hypothesis properly.
- Absence of clear theoretical framework or knowledge of theoretical framework, e.g., awareness of rights among women depends upon personality, environment (education).
- Lack of ability to utilize the theoretical framework logically, e.g., workers’ commitment and role skills and role learning.
- Evaluating whether a hypothesis is good or bad depends upon the amount of information it provides about the phenomenon. For example, let us take the following hypothesis, given in three forms:
- X is associated with Y.
- X is dependent on Y.
- As X increases Y decreases. Of these three forms, third form explains the phenomenon better.
Characteristics of A Useful Hypothesis:
Goode and Hatt have described the following characteristics of a good hypothesis:
- It must be conceptually clear. This means that concepts should be defined lucidly. These should be operationalized. These should be commonly accepted. These should be communicable. In the hypothesis, “as institutionalization increases, production decreases”, the concept is not easily communicable.
- It should have empirical referents. This means that it should have variables which could be put to empirical test, i.e., they should not merely be moral judgements. For example, capitalists exploit workers, or officers exploit subordinates, or young people are more radical in ideas, or efficient management leads to harmonious relations in an establishment. These hypotheses cannot be considered useful hypotheses.
- It should be specific, e.g., vertical mobility is decreasing in industries, or exploitation leads to agitation.
- It should be related to available techniques, i.e., not only the researcher should be aware of the techniques but these should be actually available. Take the hypothesis: “Change in infrastructure (means of production and relations of production) leads to change in social structure (family, religion, etc)”. Such hypothesis cannot be tested with available techniques.
- It should be related to a body of theory.
Sources of Deriving Hypotheses:
- Cultural values of society : American culture, for example, emphasizes individualism, mobility, competition and equality, while Indian culture emphasizes tradition, collectivism, karma and unattachment. Therefore, Indian cultural values enable us to develop and test the following hypotheses:
- Residential jointness in Indian family has decreased but functional jointness continues to exist.
- Divorce is used as a last resort by a woman to break her marriage.
- Caste is related to voting behaviour among Indians.
- Indian family comprises of not only primary and secondary kin but most often of tertiary and distant kin too.
- Past research : Hypotheses are often inspired by past research. For example, a researcher studying the problem of student unrest may use the finding of another study that “students having spent two or three years in the college/university take more interest in students’ problems in the campus than freshers; or that “students with high ability and high social status participate less in students, agitations than those who have low ability and low social status”. Such hypotheses could be used either to replicate past studies or revise the hypotheses that the alleged correlation does not exist.
- Folk wisdom : Sometimes researchers get the idea of a hypothesis from commonly held lay beliefs, e.g., caste affects individual’s behaviour, or that geniuses lead unhappy married life, or married women without children are less happy, or that young illiterate married girls are more exploited in joint families, or that being an only child creates barriers in child’s development of some personality characteristics, and so on.
- Discussions and conversations: Random observations during discussions and conversations and reflections on life as a person throw light on events and issues.
- Personal experiences: Very often researchers see evidence of some behaviour pattern in their daily lives. Intuition: Sometimes the investigators get a feeling from inside that certain phenomena are correlated. The suspected correlation leads the investigator to hypothesize a relationship and conduct a study to see if his/her suspicions are confirmed. For example, living in a hostel for a few years gives an idea to the hostler that “lack of control leads to deviant behaviour”. He/she therefore decides to study hostel sub-culture.
Functions or Importance of Hypotheses:
Sarantakos has pointed out following three functions of hypotheses:
- To guide social research by offering directions to the structure and operation;
- To offer a temporary answer to the research question; and
- To facilitate statistical analysis of variables in the context of hypothesis testing.
The importance of hypotheses can also be pointed out in following terms:
- Hypotheses are important as tools of scientific inquiry/research because they are derived from theory or lead to theory.
- The facts (in hypotheses) get a chance to establish the probable truth or falsify it.
- Hypotheses are tools for the advancement of knowledge as they stand apart from man’s values and opinions.
- Hypotheses help the social scientists to suggest a theory that may explain and predict events.
- Hypotheses perform a descriptive function. The tested hypothesis tells us something about the phenomenon it is associated with. In a nutshell, the main functions of hypotheses are:
- To test theories,
- To suggest theories, and
- To describe social phenomena.
The secondary functions are:
- To help in formulating social policy, say, for rural communities, penal institutions, slums in urban communities, educational institutions, solutions to various kinds of social problems;
- To assist in refuting certain ‘common sense’ notions (e.g., men are more intelligent than women); and
- To indicate need for change in systems and structures by providing new knowledge.
Criticism of Hypotheses:
- Some scholars have argued that each study needs a hypothesis. Not only exploratory and explanatory researches but even the descriptive studies can benefit from the formulation of a hypothesis. But some other scholars have criticized this position. They argue that hypotheses make no positive contribution to the research process. On the contrary, they may bias the researchers in their data collection and data analysis. They may restrict their scope and limit their approach. They may even predetermine the outcome of the research study.
- Qualitative researchers argue that although hypotheses are important tools of social research, they must not precede the research but rather result from an investigation.
- Despite these two contradictory arguments, many investigators use hypotheses in their research implicitly or explicitly. The greatest advantage is that they not only guide in goals of research but help in concentrating on the important aspects of the research topic by avoiding less significant issues.
Reliability;
Reliability is the consistency of your measurement, or the degree to which an instrument measures the same way each time it is used under the same condition with the same subjects. In short, it is the repeatability of your measurement. A measure is considered reliable if a person’s score on the same test given twice is similar. It is important to remember that reliability is not measured, it is estimated.
There are two ways that reliability is usually estimated: Test/Retest and Internal Consistency.
- Test/Retest: Test/retest is the more conservative method to estimate reliability. Simply put, the idea behind test/retest is that you should get the same score on test 1 as you do on test 2. The three main components to this method are as. follows
- Implement your measurement instrument at two separate times for each subject;
- Compute the correlation between the two separate measurements; and
- Assume there is no change in the underlying condition (or trait you are trying to measure) between test 1 and test 2.
- Internal Consistency: Internal consistency estimates reliability by grouping questions in a questionnaire that measure the same concept. For example, you could write two sets of three questions that measure the same concept (say class participation) and after collecting the responses, run a correlation between those two groups of three questions to determine if your instrument is reliably measuring that concept.
The primary difference between test/retest and internal consistency estimates of reliability is that test/retest involves two administrations of the measurement instrument, whereas the internal consistency method involves only one administration of that instrument.
Validity:
Validity is the strength of our conclusions, inferences or propositions. More formally, Cook and Campbell (1979) define it as the “best available approximation to the truth or falsity of a given inference, proposition or conclusion.” In short, were we right? Let’s look at a simple example. Say we are studying the effect of strict attendance policies on class participation. In our case, we saw that class participation did increase after the policy was established. Each type of validity would highlight a different aspect of the relationship between our treatment (strict attendance policy) and our observed outcome (increased class participation).
Types of Validity;
There are four types of validity commonly examined in social research :
- Conclusion validity asks is there a relationship between the programme and the observed outcome? Or, in our example, is there a connection between the attendance policy and the increased participation we saw?
- Internal Validity asks if there is a relationship between the programme and the outcome we saw, is it a causal relationship? For example, did the attendance policy cause class participation to increase?
- Construct validity is the hardest to understand in my opinion. It asks if there is there a relationship between how I operationalzed my concepts in this study to the actual causal relationship I’m trying to study? Or in our example, did our treatment (attendance policy) reflect the construct of attendance, and did our measured outcome – increased class participation – reflect the construct of participation? Overall, we are trying to generalize our conceptualized treatment and outcomes to broader constructs of the same concepts.
- External validity refers to our ability to generalize the results of our study to other settings. In our example, could we generalize our results to other classrooms?
Camparrision b/w Validity and Reliability
- The real difference between reliability and validity is mostly a matter of definition. Reliability estimates the consistency of your measurement, or more simply the degree to which an instrument measures the same way each time it is used in under the same conditions with the same subjects. Validity, on the other hand, involves the degree to which you are measuring what you are supposed to, more simply, the accuracy of your measurement. It is my belief that validity is more important than reliability because if an instrument does not accurately measure what it is supposed to, there is no reason to use it even if it measures consistently (reliably).
- So what is the relationship between validity and reliability? The two do not necessarily go hand-inhand. At best, we have a measure that has both high validity and high reliability. It yields consistent results in repeated application and it accurately reflects what we hope to represent.
- It is possible to have a measure that has high reliability but low validity – one that is consistent in getting bad information or consistent in missing the mark. It is also possible to have one that has low reliability and low validity – inconsistent and not on target.
- Finally, it is not possible to have a measure that has low reliability and high validity – you can’t really get at what you want or what you’re interested in, if your measure fluctuates wildly.