Variables, Sampling, Hypothesis, Reliability, and Validity | Sociology UPSC Notes

Print Friendly, PDF & Email
Variables, Sampling, Hypothesis, Reliability, and Validity

The Variables:

• A variable is something that can have more than one number. It’s not always the same. It is something that a lot of people, groups, events, things, etc. have in common.

• The degree to which each case has the trait is different from case to case. So, age (young, middle-aged, old), income class (low, middle, upper), caste (low, intermediate, high), education (illiterate, less educated, highly educated), work (low status, high status), etc. are all variables.

• The variables chosen for study are called explanatory variables, and all other variables are called extraneous variables. Variables that are not part of the set of explanatory variables are either managed or uncontrolled.

Controlled factors, also called control variables, are kept the same or kept from changing during the study. This is done so that the research can focus on one thing. For example, all men and women under the age of 18 could be left out of the study. This would mean that the theory doesn’t care about any particular subgroups.

Variables can be:

Dependent and Independent Variables:

• A variable that changes based on how another variable changes is called a “dependent variable.” A variable that changes without affecting another variable is called an independent variable. In a controlled experiment, the experimental variable that is not given to the control group is the independent variable.

• The variable that is changed by the experimenter is the independent variable. For example, a teacher might want to know if the talk method, the question-and-answer method, the visual method, or a mix of two or more of these methods is the best way to teach. In this case, the way of teaching is an independent variable that the teacher can change. The “effect on the students’ understanding” is the dependent variable. The thing we are trying to explain is called the “dependent variable.” In this experiment, apart from the ways of teaching, other independent factors could be the students’ personalities, their social class, how they are motivated (by rewards and punishments), the atmosphere in the classroom, how they feel about the teacher, and so on.

Experiment and factors that were measured:

• The experimental variables describe how the experiment was done, while the observed variables describe how the experiment was done. For example, rural development (a measured variable) could be looked at in terms of income growth, literacy rate, infrastructure, access to medical care, social security, and so on. In another study, we could look at how the absence or presence of books, libraries, good teachers, the use of visuals, and so on affects how well students do in school. All of these will be things that the researcher will change or change in an experiment. When planning and doing study, it is important to know the difference between these two types of variables.

Active factors and what they are:

• Variables that are changed or tried out will be called “active variables,” and variables that are measured will be called “assigned variables.” In other words, an active variable is one that can be changed, and a given variable is one that can’t be changed.

Qualitative and numeric variables:

• A quantitative variable is one whose values or groups are made up of numbers and whose differences between groups can be shown with numbers. Thus, age, income, sizes are quantitative factors. The qualitative variable is one that is not made up of numbers but of clear groups. This variable has at least two different groups that can be told apart. Class (low, middle, or high), caste (low, middle, or high), gender (male or female), and religion (Hindu or not Hindu) are all examples of qualitative factors.

• (Singleton and Straits) Relationships between quantitative factors can be either good or bad. When the value of one variable goes up, the value of the other variable also goes up, or when the value of one variable goes down, the value of the other variable also goes down. In other words, the two things always change in the same way. For example, if a father is taller, his kid will also be taller. A negative relationship between two variables arises when the value of one variable goes down and the value of the other variable goes up. For example, as age goes up, life expectancy goes down.

• Therese Baker has called qualitative and quantitative variables “categorical” and “numerical” variables, respectively. The former (e.g., occupation, religion, caste, gender, education, income) are made up of sets of categories (or attributes) that must follow two rules: one, the categories must be different from each other, i.e., they must be mutually exclusive; and two, the categories must be exhaustive, i.e., they should cover all the possible ranges of variation in a variable. After putting oneself into the categories of educated (the other being illiterate) in the field of education, one can put oneself into the subcategories of student, graduate, postgraduate, etc.

Variables can also be either binary or continuous. For example, sex is a binary variable, but ability is a continuous variable. Only a few factors are true dichotomies most of the time. Most variables can take on numbers that don’t change. Still, it’s good to keep in mind that converting continuous variables to dichotomous or trichotomous variables is often useful or required.


• A sample is a small group of people taken from a bigger group. It will be a good representation of the whole population only if it has the same basic traits as the whole population. In sampling, what we care about is not what kinds of units (people) will be interviewed or watched, but how many units of a certain kind and by what method should be picked.

“A sample is a part of the population that is studied to draw conclusions about the whole population,” says Manheim. In order to define the group from which the sample is taken, the “target population” and the “sampling frame” must be named. The target community is the group of people about whom information is needed. For example, drug-using students at one university, voters in one village/constituency, and so on. In order to define the group, the criteria for which cases are included and which are not must be stated.

• For example, for a study on how well women in one town know their rights, the target population is all women between the ages of 18 and 50, whether they are married or not. If the unit is an institution, like Vidya Mandir, it needs to say what kind of structure it has, how big it is based on the number of students in the school and college sections, and how many teachers and employees work in the professional classes.

• The sampling frame needs to be built so that the target group can be used. This is the set of all the cases from which the real sample is taken. It’s important to remember that the sampling frame is not a sample. Instead, it’s the operational description of the population that gives the sampling a place to start.

For example, in the Vidya Mandir case, if students in school and college are taken out of the sample, only students in professional classes (MBA, Computer Science, B.Ed., Home Science, and Biotechnology) are left. So, the sample frame cuts down on the total number of people and tells us who we want to study (i.e., only students of professional classes).

There are two reasons why sampling is done.

• Estimation of variables

• Putting a theory to the test Estimation of parameters: The main goal is to predict certain population parameters, such as how many office clerks worked overtime.

So, the study tries to pick a sample and figure out the important numbers (such as the average and the proportion). He can use this number as a guess to say something about how accurate it is in terms of standard errors and draw a conclusion about how many people live there in terms of probability.

Putting a theory to the test: The second goal of sampling could be to test a statistical theory about a group, such as the theory that at least 60% of households in the town of Kurukshetra have TVs.

The experts might choose a sample of homes and then figure out how many of those homes have TVs. Now, the problem is to figure out if the sample result is enough to say “no” to the theory or “yes” to it. To figure out how to solve this problem, the researcher needs to find a way to figure out how far off the sample result is from the hypothetical number.

Why we take samples,

Sarantakos has said that sampling is done for the following reasons:

• In many cases, the population may be so big and spread out that it may not be possible to cover everyone.

• It is very accurate because it only works with a small number of people. Most of us have had blood samples taken, sometimes from our fingers, sometimes from our arms, and sometimes from somewhere else on our bodies. The idea is that the blood is pretty much the same all over the body and that a sample is enough to figure out what the blood is like. Singleton and Straits have also said that looking at all cases will give a less true picture of the population than looking at a small sample.

• Valid and similar results can be found in a short amount of time. When data is collected over a long period of time, some of it is usually out of date by the time it is all collected. For example, getting information about how voters feel about different issues during election time, or demanding action against police officers who beat up women at protests or blind a lot of accused people in the police jail. Also, views at the time of an event and opinions a few months later are likely to be different. So, the results are likely to change if data is collected over a long period of time, which means not just taking a small sample but looking at the whole community.

ALSO READ  UPSC CSE Topper Mains Answer [Part 3]

Sampling is easier on investigators because it only needs a small part of the target group. It is also cheaper because it only needs a small number of people. A big population would mean hiring a lot of interviewers, which would raise the cost of the survey as a whole.

Many study projects, especially those that test quality control, require that the things being tested be thrown away. If the company that makes electric lights wants to know if each one meets a certain standard, there wouldn’t be any left after testing.

How to Pick a Sample:

The main idea behind sampling is that we try to learn about all the units (called the population) by looking at a small number of units (called the sample) and then applying what we learn from the sample to the whole population. With a cutter, we can take a small sample from the middle of a bag of wheat to tell if the wheat in the bag is good or not. But it’s not always true that a sample study gives us a good picture of the whole community.

If only a few people in a village agree that family planning is a good idea, that doesn’t mean that everyone in the village agrees. Opinions can be different based on religion, amount of education, age, wealth, and other things. From the study of a small number of people, the wrong conclusion or generalisation can be drawn because they are not a good representation of the whole community.

The study of a sample is needed because a study of a very big population would take a long time, a lot of interviewers, a lot of money, and the data collected by so many investigators might not be accurate. With a group, it is easier to plan an observation or study.


The most important rules of sampling are:

• Sample units must be picked in a methodical and fair way.

• Sample units should be well-defined and easy to find.

• Each bit of a sample must be able to stand on its own.

• During the whole study, the same number of sample units should be used.

• The process for choosing winners should be based on good criteria and avoid mistakes, bias, and other kinds of misinformation.

The benefits of sampling:

Some of the benefits of sampling are shown by the reasons and rules we’ve talked about so far.

These things:

• It’s not possible to study a lot of people who are spread out over a big area. By taking samples, their number will go down.

• Time and money are saved.

• It saves destruction of units.

• It makes the data more accurate because you can keep track of the small number of people.

• It gets more people to answer.

It gets more help from those who answer.

It’s easy to keep an eye on a small number of interviewers in a sample, but it’s hard to keep an eye on a large number of interviewers in a study of the whole community. The researcher can stay out of sight.

Why sampling is important:

There are many reasons why selection is important when collecting statistics.

Only Possible, Fast, and Cheap Way: It’s quick and cheap and might be the only way to do it. In a factory, samples are used to check the quality of the goods. If the product is not good enough after being tested, it is either remade or thrown away. So, tasting is the only way to figure out how good something is. In the same way, instead of observing all things, it is faster and cheaper to pick a sample from the world and figure out what those things are like from that sample. It is a very useful tool for researchers and practitioners who want to keep certain characteristics of a community within certain limits.

Representativeness and Sample Size: The Problem of Sample Representatives When choosing a group, the main goal is to make sure it is as representative of the whole world as possible. Clearly, the size of a group does not always affect how representative it is.

So, if a small sample is chosen scientifically, it may be more accurate than a large sample chosen at random. Samples should be chosen so that every item in the population being studied has the same chance of being a good representation of the population as a whole.

A biassed sample is one that doesn’t show what the community is like as a whole.

Yule and Kendal point out that “the human being is a very poor tool for making a random selection.” When observers have the chance to make their own decisions or choices, bias is almost certain to creep in. Studies that use skewed sampling are inherently wrong and false. This is true of a number of studies in behavioural science that are based on surveys that were mailed out and only some of which were filled out and returned. The original mailing list of possible responders could, of course, be a good sample. But the surveys that were sent out may be very different from the ones that were actually sent back.

Sample Size Problem:

A scientific group is one that is both representative of the whole population and has enough cases to make sure the results are accurate. The question of whether a sample is enough is very complicated. According to Hagood and Price, the size of the sample can be determined by three things: the parameters that will be studied, the range of reliability that can be used in estimates, and a good idea of how far apart the studied characteristics are.

Different kinds of sampling:

Two types of sampling: probability sampling and non-probability sampling.

1. In probability sampling, every unit of the community has the same chance of being chosen for the sample. It is a good representation of the whole.

2. Non-probability sampling doesn’t claim to be representative because it doesn’t give every unit a chance to be chosen. The researcher is in charge of choosing the sample units.

Probability Sampling:

Probability sampling is still the best way to choose large, representative samples for business and social science studies. Black and Champion say that the following conditions must be met for probability sampling to work:

• There is a full list of all the things to study;

• The size of the universe must be known;

• The desired sample size must be stated; and

• Each element must have an equal chance of being chosen.

It means that they should use some kind of chance in at least one of their stages. Leabo divides probability samples into five groups: samples, random samples, sorted samples, and samples chosen at random.

Simple Random Sampling: Simple random samples are not used very often, but they are the base for other sampling methods. A simple random sample of n items is a sample taken from a community so that each possible combination of n units has the same chance or probability of being chosen.

Simple random sampling has the following benefits:

1. It saves time. Full coverage is more expensive than sampling, but the cost per unit is higher.

2. It saves time and money on labour. With sampling, fewer people are needed to gather, record, and handle the data. So, it saves a lot of work. 3. It saves time. Because of these benefits, sampling was used for the first time in the 1951 census of people.

This method saves a lot of time.

4. It improves accuracy. The overall amount of accuracy is higher when a sample is used. It makes it possible for the field to be better, with more checks for correctness, careful editing and analysis, and more detailed information.

Random sampling with a plan:

For these samples, the population is split into groups that are similar and a random sample is taken from each group. People can be put into groups based on what we know about them and how a certain feature group affects them. People can be put into groups based on what is known about them and how a certain trait affects the estimate that needs to be made.

The pros of stratified random sampling are as follows:

This process makes sure that each group and chance sample are properly represented. The way people were put into groups or classes was based on the type of problem being studied.

For example, if the problem is to figure out how much the average person makes in a certain area, occupational groups can be used as biases to divide the community. If it is done right, the stratified random sample is better than the sample random sample. In fact, the reliability of the data for a given sample size goes up as the range of all possible sample averages gets smaller. This means that a stratified random sample of the same size is more reliable than a simple random sample of the same size.

Non-probability sampling:

Probability sampling is hard to do and shouldn’t be used in many research situations, especially when there is no list of people to study (e.g., husbands who beat their wives, widows, people who own Maruti cars, people who use a certain type of detergent powder, alcoholics, students and teachers who skip class a lot, migrant workers, etc.). Non-probability sampling is the best way to do these kinds of studies. Non-probability sampling methods don’t use the rules of probability theory, don’t claim to be representative, and are usually used for qualitative exploratory analysis.

There is no randomness in these samples, so they can be called quota sampling, purposeful sampling, accidental sampling, or snowball sampling.

• Quota sampling is used to do study in marketing. It is a structured, but not random, sampling. In this type of sampling, the population is split into two or three groups based on their traits. Then, a quota is set, and each interviewer is asked to ask a certain number of people from each area. The interviewer might choose a person from the crowd who is easy to reach. Because of this, bias is likely to show through. The bias can be lessened by making it harder for him to do what he wants. This method is useful when you need rough estimates instead of exact results. In fact, the answers are just a rough guess, so they can’t be tested to see if they are reliable.

ALSO READ  Swati Sharma Biography, Age, UPSC Marksheet, Rank, Optional Subject, Preparation Strategy

• Purposeful Sampling: This involves using your best judgement and making a concerted effort to get a sample that is representative of most places or groups.

Namjoshi’s work is a good example of what a purposive sample is. In this study, two types of people were asked to take part: 1. married men and women 2.Men and women who are not married. This method was used to choose both samples so that there would be enough people from higher and lower castes, different social groups, and both sexes. A group of 400 married men and women and a group of 400 single boys and girls were chosen as samples.

• Accidental sampling:– This is the weaker type of sampling because it uses the samples that are already out there. This type of sampling can be used if there are no other options.

• Snowball sampling: It refers to a set of procedures in which the first respondents are chosen by chance, and then more respondents are chosen based on the information they give. Referral is used with this method to find members of rare groups. For example, a company wants to sell a wood croquet set for serious adult players. Since the market for this product is small, researchers need to use this method to get the job done as cheaply as possible.

A theory is a guess about how variables are related to each other. It is a guess about what the research problem is or what the research results will be. Before starting the study, the researcher has a vague, even muddled idea of what the problem is. It might take the expert a long time to say what questions he was trying to answer. So, it is very important to have a good statement about the study problem.

• Theodor son and Theodor son: “A hypothesis is a tentative statement that suggests a link between certain facts.”

• Ker linger: “A hypothesis is a guess about the link between two or more variables.”

• Black and Champion have called it a “possible statement about something whose truth is often unknown.” This claim is meant to be tried in the real world to see if it is true or false. If the statement is not backed up by enough evidence, it is not a scientific rule.

Webster defines a hypothesis as “a tentative assumption made in order to find and test its logical or empirical consequences.” “Test” in this case means “either prove it wrong or prove it right.” Since statements in Hypothesis must be tested in the real world, the definition of hypothesis excludes all statements that are just opinions (e.g., getting older makes you sicker), value judgements (e.g., modern politicians are corrupt and have a vested interest to serve), or normative (e.g., everyone should take a morning walk). A normative statement is a statement about how things should be. It is not a statement of fact that can be proven right or wrong through research.

Some examples of theories are as follows:

• Group study improves performance in the upper level.

• Hostels use more.

• More crimes against women happen to young girls (15–30 years old) than to middle-aged women (30–40 years old).

• Men from the lower class commit more crimes than men from the middle class.

• The risk of suicide goes down when people spend less time alone.

• Women with less education have more trouble adjusting to marriage than women with more education.

• Most kids from broken homes grow up to be bad people.

• Unemployment lessens juvenile delinquency.

• People with more money have fewer children than people with less money.

Criteria for Building Hypotheses:

Hypothesis is never written out as a question. Bailey, Becker, Selltiz, and Sarantakos have all said that a theory must meet a number of criteria.

It should be able to be tested in the real world to see if it is right or wrong. It should be specific and clear.

• There shouldn’t be any contradictions in the claims in the hypothesis.

• It should list the factors that will be used to figure out the relationship.

• It should only talk about one thing.

You can write a theory in either a descriptive or a relational way. In the first, it explains what happened, while in the second, it shows how variables are related. A theory can also be in the form of a direction, a lack of direction, or the word “null.”

What hypotheses are:

The following must be true for a scientifically valid hypothesis:

• It must be true to the sociological facts that are important.

• It can’t go against what is known to be true in other science fields.

• It must take into account what other experts have found.

You can’t say whether a hypothesis is true or wrong. They can only be about the topic of the study or not. For example, poverty in a village can be caused by:

• Low growth of agriculture, which is caused by a lack of irrigation, sandy soil, unpredictable rain, and the use of traditional farming tools.

• Poverty is caused by a lack of infrastructure, like power, roads, and markets.

• Obstacles to rural growth include lack of resources (water, soil, minerals), lack of support (rain, irrigation, livestock), and problems with the social system (lack of credit, infrastructure, wasteful spending, and market problems).

The important hypotheses could be:

1. Rural poverty is linked to the availability and ease of getting loans.

2. Lack of infrastructure is the cause of rural poverty.

• People who are poor tend to spend a lot of money on social activities.

• Water, land, and mineral shortages are linked to rural poverty in a bad way.

Different kinds of hypotheses:

Hypotheses can be broken down into working hypotheses, study hypotheses, null hypotheses, statistical hypotheses, alternative hypotheses, and scientific hypotheses.

1. A working hypothesis is a researcher’s first guess about the topic of the research. It is used when there isn’t enough knowledge to make a hypothesis, or as a step towards making the final research hypothesis. Working theories are used to make the final research plan, put the research problem in the right context, and narrow down the research topic to a manageable size. In the area of business administration, for example, a researcher can come up with a working hypothesis like “promising a bonus makes a product sell more.” Later, after getting some preliminary data, he changes this idea and comes up with the study hypothesis that “promising a good bonus makes a product sell more.”

2. A scientific hypothesis has a statement that is based on or comes from enough theoretical and real-world facts.

3. An alternative hypothesis is a set of two hypotheses (research and null) that say the opposite of the null hypothesis. In statistical studies of null hypotheses, when Ho (the null hypothesis) is accepted, the alternative hypothesis is rejected, and when Ho is rejected, the alternative hypothesis is accepted.

4. A researcher’s study hypothesis is a statement about a social fact that doesn’t talk about its details. Researchers think it’s true and want to disprove it. For example, they think that Muslims have more children than Hindus or that upper-class students who live in dorms or rented rooms are more likely to use drugs. Theories can lead to research hypotheses or research hypotheses can lead to theories.

5. The study hypothesis is the opposite of the null hypothesis. It is a theory that there is no link. Null hypotheses are not real, but they are used to test other hypotheses in study.

6. According to Winter (1962), a statistical hypothesis is a statement or fact about statistical populations that a person wants to prove or disprove. Things are turned into numbers, and choices are made based on these numbers, like the difference in income between two groups: More money is in Group A than in Group B. The null hypothesis will be that Group A is not richer than Group B. Here, factors are turned into numbers that can be measured.

Goode and Hatt have identified three kinds of theories based on how abstract they are:

• That explains a claim in terms of common sense; that already has some common sense observations about it; or that tries to test claims that make sense. For instance, bad parents raise bad kids, managers who work hard always make money, and rich students drink more booze.

• Which are a little bit complicated, i.e., which describe a relationship that is a little bit complicated. • Religious polarisation leads to riots between different groups.

• Cities grow in rings that move outward from the centre (Burgess).

• Unstable economic conditions make it hard for a business to grow.

• Sutherland says that different relationships lead to crime.

• Living in poverty is linked to youth crime, according to Shaw.

• Healy and Bronner say that mental illnesses are the cause of bad behaviour.

• Which are very complex, meaning that they describe the relationship between two factors in more complicated terms. For example, high fertility is more common among people with low incomes, who are conservative, and who live in rural areas than among people with high incomes, who are modern, and who live in cities. In this case, “fertility” is the dependent variable, and “income, values, education, place of residence, etc.” are the independent variables. The other case is that Muslim women have more children than Hindu women. To test this theory, we have to keep a number of things the same. This is a vague way to deal with the issue.

Problems with coming up with hypotheses:

Goode and Hatt say that there are three main problems with making hypotheses:

1. Not being able to put the idea in the right way.

2. There isn’t a clear theoretical framework or the researcher doesn’t know about it. For example, a woman’s awareness of her rights relies on her personality and her environment (her schooling).

3. Not being able to use the theory framework in a logical way, such as with workers’ commitment, role skills, and learning roles.

4. You can tell if a theory is good or bad by how much information it gives about the thing being studied. Take the following theory, which is given in three different ways:

• X is linked to Y.

• X depends on Y.

• When X goes up, Y goes down. The third form is the best one to explain what’s going on.

ALSO READ  G20 or Group of Twenty : Countries, Importance, Agenda, G20 Summit 2023 UPSC

Things that make a hypothesis useful:

Goode and Hatt have said that a good theory has the following traits:

• The main idea must be clear. This means that ideas should be explained in clear terms. These ideas should be put into action. Most people should agree with these. These should be easy to share. The idea behind the statement “as institutionalisation goes up, production goes down” is not easy to explain.

• It should be based on real-world examples. This means that the factors should be able to be tested in the real world. They shouldn’t just be moral judgements. For example, capitalists take advantage of their workers, officers take advantage of their subordinates, young people are more bold in their ideas, and good management leads to good relationships in a business. These ideas can’t be thought of as useful ideas.

• It should be clear, like “Vertical mobility in industries is decreasing” or “Exploitation leads to agitation.”

• It should be linked to techniques that can be used. This means that not only should the researcher know about the techniques, but they should also be able to be used. Take the statement, “Changes in infrastructure (means of production and relationships of production) lead to changes in social structure (family, religion, etc.).” With the tools we have, we can’t test this kind of theory.

• It should fit into a body of knowledge.

Where Hypotheses Come From:

1. The cultural ideals of a society. For example, American culture values individualism, mobility, competition, and equality, while Indian culture values tradition, collectivism, karma, and not being attached to anything. So, Indian traditional values let us come up with and test the following ideas:

• The number of Indian families who live together has gone down, but they still work together.

• Divorce is the last thing a woman will do to end her marriage.

• The way Indians vote depends on their caste.

• An Indian family usually includes not only first and second cousins, but also third cousins and even distant cousins.

2. Past study: Hypotheses are often based on research that has already been done. For example, a researcher looking into student unrest might use the results of another study that say “students who have been in college or university for two or three years are more interested in student problems on campus than freshmen” or “students with high ability and high social status are less likely to take part in student agitations than students with low ability and low social status.” Such theories could be used to either confirm the results of previous studies or change the ideas that the supposed correlation does not exist.

3. Folk wisdom: Sometimes, researchers get an idea for a hypothesis from a widely held belief, such as that a person’s caste affects how they act, or that geniuses have unhappy marriages, or that married women without children are less happy, or that young, uneducated married girls are more likely to be exploited in joint families, or that being an only child makes it harder for a child to develop certain personality traits, and so on.

Discussions and conversations: People’s random observations and thoughts about their own lives shed light on events and problems.

5. Personal experiences: Researchers often see proof of certain patterns of behaviour in their everyday lives. Intuition: Sometimes, investigators have a gut feeling that something is connected to something else. The investigator makes a guess about a link between the two things and then does a study to see if his or her guess is right. For example, someone who has lived in a hotel for a few years might think that “lack of control leads to bad behaviour.” So, he or she begins to study the subculture of hostels.

What hypotheses do or how important they are:

Sarantakos has pointed out that theories do the following three things:

1. To help organise and run social research. 2. To give a temporary answer to the research question. 3. To make statistical analysis of variables easier in the context of hypothesis testing.

You can also say the following about how important theories are:

1. Hypotheses are important for scientific study and inquiry because they come from theory or lead to theory.

2. The facts (in theories) have a chance to show that the likely truth is true or that it is not true.

Hypotheses are used to learn more because they are not based on people’s ideals or opinions.

4. Hypotheses help social scientists come up with theories that could explain and predict what will happen.

5. Hypotheses describe what is going on. The fact that the theory was tested tells us something about the thing it is about. In a nutshell, hypotheses are mostly used to test theories, suggest new theories, and explain social events.

The secondary functions are to help make social policy, such as for rural communities, prisons, slums in cities, educational institutions, and solutions to different kinds of social problems; to help disprove some “common sense” ideas, such as the idea that men are smarter than women; and to show that systems and structures need to change by giving new information.

Criticisms of Hypotheses:

• Some experts have said that every study needs a hypothesis. The development of a theory can help not only exploratory and explanatory research, but also descriptive research. But this view has been criticised by other experts. They say that hypotheses do not help the study process in any way. On the contrary, they may affect how the experts collect and analyse data. They might limit what they do and how they do it. They might even be able to predict how the research study will turn out.

• Qualitative researchers say that theories are important tools in social research, but they shouldn’t come before the research. Instead, they should come out of the research.

Even though these two points are at odds with each other, many researchers use hypotheses, whether they say so or not. The biggest benefit is that they not only help you set goals for your study, but they also help you focus on the most important parts of your topic by letting you skip over the less important ones.

Reliability is how consistent your measurements are, or how much a tool measures the same way every time it is used in the same way with the same people. In short, it’s how often you can measure the same thing. If a person gets the same score on the same test twice, we say that the measure is reliable. It’s important to remember that reliability is not tested, it’s estimated.

Most of the time, there are two ways to figure out how reliable something is: Test/Retest and Internal Consistency.

Test/retest: Test/retest is the more

safe way to figure out how reliable something is. The idea behind test/retest is that your score on test 1 should be the same as your score on test 2. The three most important parts of this method are:

• Use your measuring tool on each subject twice at different times;

• Figure out the relationship between the two readings; and

• Assume that the underlying condition or trait you are trying to measure hasn’t changed between test 1 and test 2.

2.Internal Consistency: Internal consistency measures reliability by grouping questions in a questionnaire that measure the same idea. For example, you could write two sets of three questions that measure the same thing, like class involvement, and then compare the answers to see if your instrument is accurately measuring that thing.

The main difference between test/retest and internal consistency measures of reliability is that test/retest involves giving the measurement tool twice, while the internal consistency method only requires giving the measurement tool once.

Validity is how strong our findings, inferences, or statements are. In a more formal way, Cook and Campbell (1979) say that it is the “best available approximation to the truth or falsity of a given inference, proposition, or conclusion.” To put it simply, were we right? Let’s look at a simple case. Say we are looking into how strict attendance rules affect how many people show up to class. In our case, we saw that more people did come to class after the strategy was put in place. Each type of validity would show a different part of the link between our treatment (a strict policy on attendance) and the result we saw (more people in class).

Types of Validity; In social study, there are four main types of validity:

1. Conclusion truth asks if there is a link between what was done and what was seen. Or, in our case, is there a link between the policy on attendance and the fact that more people showed up?

2. Internal validity asks if there is a cause-and-effect link between the programme and the results we saw. For instance, did the attendance rules make more people come to class?

3.Construct validity is the most difficult for me to understand. It asks if there is a connection between how I operationalized my concepts in this study and the real causal relationship I’m trying to study. Or, in our case, did our treatment (the policy on attendance) match the concept of attendance? Did our measured outcome (more people in class) match the concept of participation? Overall, we are trying to apply how we think about treatment and results to a wider range of similar ideas.

4. External validity means that the results of our work can be used in other places. In our case, are our results something that could be used in other classrooms?

Marriage between Value and Trustworthiness

• The main thing that makes the difference between dependability and validity is how you define them. Reliability is a way to figure out how consistent your measurements are, or how much an instrument measures the same way every time it is used in the same way with the same people. Validity, on the other hand, is about how well you measure what you’re meant to measure, or, in other words, how accurate your measurement is. I think validity is more important than reliability because if an instrument doesn’t measure what it’s meant to measure accurately, there’s no point in using it, even if it measures the same thing every time (reliably).

So, what is the link between truth and reliability? These two things don’t always go together. At best, we have a way to measure that is both accurate and valid. It gives the same results every time, and it’s a good representation of what we want to stand for.It is possible for a measure to be very reliable but not very true. This means that it consistently gives wrong information or misses the mark. It is also possible for one to be inconsistent and not on goal, with low reliability and validity.

Lastly, a measure can’t be both unreliable and accurate. If your measure changes all over the place, you can’t really find out what you want or what you’re interested in.