Methods of Enquiry in Psychology

Written and reviewed by Subhra Gupta, Humanities faculty at UniDrill (M.A. Sociology, University of Delhi · 7 years teaching Humanities). Based on NCERT Psychology Class XI (Introduction to Psychology), Chapter 2.

Smoke in the waiting room

Imagine you are alone in a waiting room, filling in a questionnaire, when smoke starts pouring through a wall vent. Within four minutes you can hardly see or breathe. Do you walk out and report it? In a famous 1970 study, 75 per cent of students sitting alone did. But when the student sat with two other people who calmly ignored the smoke, only 10 per cent reported it.

Nobody could have known this by asking "what would you do?" Psychologists had to build a situation, change one thing and measure what happened. That is the heart of this chapter: psychology is a science because of how it finds out, not because of what it studies.

Goals of psychological enquiry

Psychologists try to describe, predict, explain and control behaviour, and to apply what they learn, all objectively. The same study-habits example runs through all five goals.

Prediction is not explanation. Prediction needs only an association. Explanation needs a cause. A student who knows that tall children are better at basketball can predict who will play well without knowing why.

The scientific method

Science is defined less by what it studies than by how. The scientific method is objective, systematic and testable. Objectivity means that if two people independently study an event, they should reach more or less the same conclusion, as when you and a friend measure a table with the same ruler.

The cycle in one line: Conceptualise a problem (choose a topic and form a hypothesis) → Collect data (participants, methods, tools, procedure) → Draw conclusions (statistics) → Revise conclusions (confirm the hypothesis, or restate it or build a new theory) → back to the start.

A hypothesis is a tentative answer to the research problem. An example: "The more time children spend watching violence on television, the more aggression they display." If the data support it, it is confirmed. If not, the researcher revises it and tests again, so research never really ends.

In step 2 the researcher chooses a research design, which means four decisions: who the participants will be, which method to use, which tools (interview schedule, observation schedule or questionnaire) and the procedure.

Two views. For the exam: the scientific paradigm assumes that behaviour is caused by internal and external forces and can be observed, measured and controlled. For much of the twentieth century psychology therefore studied only overt behaviour. Another reading: the interpretive paradigm emphasises understanding over explanation and prediction. It says human behaviour is complex and variable, and that people give meaning to events. When people suffer through a tsunami, an earthquake or a long illness, "objective measurement is neither possible nor desirable". The researcher explores their experience without trying to disturb its natural flow, rather like an explorer mapping unknown land. Both traditions are used, and psychology also aims to help you understand your own experience through reflection.

What kinds of data?

Data (singular: datum) are information about overt and covert behaviour, subjective experiences and mental processes. Data are not independent of context: you behave differently alone and in a group, at home and in the office. The method and the respondents also shape what you get. A student may say he behaves one way in an interview and be seen doing the opposite. And data do not speak for themselves, because the researcher must infer their meaning.

Choosing the right method

Different questions need different methods, and each has a typical example.

Observation: seeing with a purpose

You see many things every day but observe few. Scientific observation differs from everyday seeing in three ways: selection (choose one behaviour to watch), recording (tallies, notes, symbols, photographs, video) and analysis (derive meaning). The two classifications run on different axes, so a study can be both naturalistic and participant.

Observation lets you study behaviour as it occurs, but it is labour intensive, time consuming and open to observer bias: "We see things as we are and not as things are." So record what happens as it happens, and interpret later.

The experiment: finding cause and effect

An experiment changes one factor, studies its effect on another and keeps everything else constant. The cause is what the researcher changes, and the effect is the behaviour that changes in response.

A variable is any stimulus or event that varies and can be measured. A pen is not a variable, but the colour or size of pens is. A room is not a variable, but its size is. Intelligence is a variable, and so is the presence or absence of people in a room.

Memory trick. The IV is the thing the researcher varies. The DV depends on it.

Experimental and control groups. The experimental group is exposed to the manipulated IV. The control group is treated exactly the same, except the manipulated variable is absent. Participants are assigned randomly, so everyone has an equal chance of being in any group. Without it, if one group were all boys and the other all girls, gender could explain the results.

Worked example: the smoke study. Students arrived singly at a Columbia University laboratory, thinking they would be interviewed. In the waiting room they sat alone, with two other naive students, or with two confederates (people working with the researchers) who were told to do nothing.

The speed and amount of smoke and the room conditions were the same for all groups, so presence of others was the likely cause. A common wrong answer: treating the smoke as the IV. The smoke was held constant, so it cannot be the cause being tested.

Beyond the textbook. For the exam: name the IV as presence or absence of others and the DV as frequency of reporting smoke. In reality: the finding became known as the bystander effect: the more people who are present, the less likely any one person is to act. Why it matters: it explains why a crowd on a station platform may stand watching an emergency, each person assuming someone else will respond.

Extraneous variables. The IV is not the only thing that can change the DV, so the others must be controlled. There are three kinds, and each has its own control technique.

Pause and predict: a researcher tests memory under two lighting conditions by giving every participant both conditions, in the same order. Which extraneous variable threatens the result, and how can she fix it?

Check your answer: a sequential variable. By the second condition participants are tired or practised. She can fix it with counter balancing, giving half the group the lights in the opposite order.

Strengths and limits of the lab experiment. It gives relatively convincing evidence of cause and effect. But laboratory results may not generalise to real situations, which means low external validity. Some studies cannot be done for ethical reasons: you cannot starve children to study the effect of poor nutrition on intelligence. And it is hard to know and control every relevant variable.

Correlational research: do two things move together?

Here the researcher does not manipulate anything. She measures two variables and asks whether they are associated, or covary, in order to predict. The strength and direction come as a number, the correlation coefficient, that ranges from +1.0 through 0.0 to −1.0.

The closer the number is to +1 or −1, the stronger the link, and the sign gives the direction. So −0.9 is a stronger relationship than +0.3.

Pause and predict: which is the stronger relationship, +0.4 or −0.7?

Check your answer: −0.7. Strength depends on distance from zero. The minus sign only says that as one variable goes up, the other goes down.

Beyond the textbook. For the exam: correlation is used for prediction, and experiments are used to show cause and effect. In reality: two variables can move together because a third one drives both. Ice cream sales and drowning cases rise together in summer, but ice cream does not drown anyone. Why it matters: a high correlation by itself never proves that one variable causes the other.

Survey research: asking many people

Surveys began as a way to find existing facts, such as the literacy rate or religious affiliations, and attitudes to things like family planning. They are now sophisticated enough to help infer causal relationships. In a December 2004 survey of 817 people aged 25 to 55 in eight cities, respondents rated their happiness on a 5 point scale. About 47 per cent said they were extremely happy, and about 80 per cent said money cannot buy happiness.

Interviews can be Individual to Individual, Individual to Group (a variant is the focus group discussion), Group to Individual (a job interview) or Group to Group. Surveys gather information quickly from thousands of people. But people may misremember, hide their real beliefs or say what they think the researcher wants to hear.

Psychological tests

A psychological test is a standardised and objective instrument for assessing a person's standing in relation to others on some mental or behavioural characteristic. Objectivity means that two researchers testing the same group get more or less the same scores. A good test needs three further properties.

A test can be reliable without being valid, for example a bathroom scale that gives the same wrong weight every time.

Most tests combine speed and power. No one should rely on a single test, and test data should be combined with background, interests and past performance.

Case study

A case study is an in depth study of one case: an individual with distinctive characteristics, a small group with something in common (for example creative writers such as Rabindra Nath Tagore and Mahadevi Verma), an institution or an event. It uses many methods, such as interviews, observation and tests, from people connected to the case. Freud's psychoanalytic theory emerged from meticulous records of individual cases, and Piaget built his theory of cognitive development on observations of his three children. Minturn and Hitchcock studied socialisation of children among Rajputs of Khalapur, and S. Anandalakshmy studied childhood in a weavers' community in Varanasi.

The catch is generalisation. One case may not represent others, so researchers use multiple sources and several investigators, and keep a chain of evidence. Because every method has limits, a combination of methods is best. If they converge on the same result, you can be more confident.

Analysing data

The two are complementary, not contradictory. A full picture needs both.

Limits of psychological enquiry

Ethics in research

Three principles guide researchers: respect for persons' privacy and choice to participate, beneficence (protecting participants from harm) and justice (sharing the benefits of research). They turn into five practices.

Beyond the textbook. For the exam: in the smoke study, students were told they were being interviewed on a topic. This is deception, which is why consent and debriefing matter so much. In reality: modern ethics committees review any study using deception or distress before it begins, and many studies of that era could not be repeated in the same form today. Why it matters: the ethical rules are not paperwork; they decide which questions science is allowed to ask.

Exam traps

60-second revision

Where this chapter connects

• Previous: What is Psychology? introduced the natural science and social science streams that these methods serve.
• Next: Human Development puts these methods to work: Piaget's stages grew out of his observations of children, and Harlow's monkey study raises the ethics questions covered here.
• Class XII link: Variations in Psychological Attributes applies reliability, validity and norms to intelligence and aptitude tests.