Research methods


Types of variables

Independent (IV): Variable that the experimenter is deliberately changing. There are usually two levels of the IV to enable comparisons.

Dependent (DV): What is measured by the researcher. The only thing that should affect the DV is the change in the IV.

Operationalism: It is important that variables are measurable, so we identify some clear operational instructions.

Extraneous (EV): Unwanted variables that could affect the DV is not controlled. For example, noise, temperature, lighting. Controlled best in a laboratory.

1 of 36


A theory is a suggested explanation for behaviour. Psychologists test theories using objective research methods.

An aim is a general statement that explains the purpose of a study.

Testable Hypothesis: A clear and precise testable statement. Relationship + DV + two levels of the IV. Alternative hypotheses - A statement of the relationship or difference between variables. Null hypotheses - A statement of no relationship or no difference between variables.

Research procedures: Instructions to ppts - Standardised procedures = giving each ppt exactly the same info about the study to ensure what is said to them not act as an EV. Standardised procedures - Researcher uses exactly the same methods and instructions for all ppts. Randomisation - Using chance to control for bias.

2 of 36

Methods and experiments

Quantitative - Use of data that can be counted

Qualitative - Data that can be expressed in words and are non-numerical, such as a description.

Experiments: Use quantitative data. They look at a measurable change in the DV which has been caused by different levels of IV. All experiments have an IV and DV.

3 of 36

Types of experiments - Lab

An experiment conducted in a controlled environment. Experimenter manipulates the IV. Evaluation:

  • EVs can be controlled. Therefore, the researcher can be sure the IV caused the DV. Cause and effect can be established.
  • Standardised procedures can be used. So the study can be replicated. Can test the validity of results.
  • The environment is not like everyday life. So ppts' behaviour is less 'typical'. So can't generalise results to the wider world.
  • Ppts may know they are being tested and change behaviour to help experimenter. Means data may not be valid.
4 of 36

Types of experiments - Field

An experiment conducted in a natural setting. Experimenter manipulates the IV.

  • Often more realistic than lab experiments, because ppts are not aware they are being studied. Enhances the validity of results.
  • Some control over EVs, because it is possible to use standardised procedures. This means cause-and-effect conclusions are more valid.
  • The researcher may lose control of some EVs because of the real-life setting. This makes it more difficult to show cause and effect.
  • There may be ethical issues. People may not know they are involved in an experiment. This means they cannot give their informed consent.
5 of 36

Types of experiments - Natural

An experiment conducted in a natural or lab setting. The experimenter does not manipulate the IV. It would have changed anyway.

  • Usually, high validity, because variables are naturally occurring and relate to everyday life. However, this is not always the case.
  • DV is often tested in a lab. Therefore, EVs can be well controlled because standardised procedures can be followed.
  • Few opportunities for this kind of research, because it sometimes relies on natural events that occur infrequently. Reduces the usefulness of the method.
  • Maybe EVs that could affect the results. E.g. uniques characteristics of the ppts cannot be controlled because they cannot randomly be allocated to groups.
6 of 36

Experimental designs: Independent groups

Separate groups of people for each level of the IV. There is usually a control and experimental group.

  • Order effects are not a problem. The ppts only do the task once. This means they won't benefit from practice.
  • There are different ppts in each group. This means that ppt variables may affect the results and act as an EV. Reduces the validity of the results.

Dealing with ppt variables: Allocation to conditions.

Use a random way of allocating the ppts to the conditions or using a systematic method.

7 of 36

Experimental designs: Repeated measures

All ppts take part in all the conditions. There is usually an experimental and a control condition.

  • There are no ppt variables. Each ppt is compared against rather than other people. Enhances the validity of results.
  • Fewer ppts needed. In an independent groups design, you need twice as many ppts to get the same number of data items. Makes repeated measures less expensive.
  • Order effects occur when ppts are tested twice. The order in which they do the tasks may make a difference. Affects the validity of the results.

Dealing with order effects: Counterbalancing.

Half the ppts complete the conditions in one order, and the other half in the opposite order.

8 of 36

Experimental designs: Matched pairs

Ppts paired on relevant variables. Ppts are then matched and one member of each pair goes in each group.

  • There are no order effects as ppts are tested only once.
  • There are fewer ppt variables. This is because those taking part are matched on a variable that is important for the experiment. Enhances the validity of the results.
  • Matching ppts takes time and effort. It doesn't control all ppt variables. This means it may not be worthwhile.
9 of 36

Sampling methods

Target population: The group of people the researcher is studying.

Samples: Sample of ppts from the target population.

Generalisation: The sample should be representative so we can generalise to the target population

Bias: It is difficult to select a group of ppts that perfectly reflects the target population.

10 of 36

Sampling methods: Random sampling

Putting names of all members of the target population into a hat/computer program so that every member of the target population has an equal chance of being selected.

  • There is no bias. Every person in the target population has an equal chance of being selected. So samples more representative.
  • Takes more time and effort than other methods. This is because you need to obtain a list of all the members of your target population and then randomly select them. The effort may not be worthwhile.
11 of 36

Sampling methods: Opportunity sampling

Selecting the most readily available group of people.

  • Easy, quick and cheap to carry out because you simply choose people who are nearby. This makes the method less expensive.
  • Sample less likely to be unrepresentative of the population. This is because the sample is drawn from one place. This reduces the generalisability of the results.
12 of 36

Sampling mthods: Systematic sampling

Selecting every nth person from a list of all the people in the target population.

  • Avoids researcher bias. The researcher has no say over who is selected. This makes it more representative.
  • May still be biased. The sample may consist of one particular group of people. This decreases the representativeness.
13 of 36

Sampling methods: Stratified sampling

Selecting ppts in proportion to their frequency in the target population.

  • Most representative method. All subgroups are represented in proportion to the numbers in the target population. This enhances representativeness.
  • Very time-consuming. It may take a while to identify proportions of subgroups and then recruit ppts. This discourages researchers from using this method.
14 of 36

Ethical issues

Ethical issues in psychology: A conflict between ppts' rights and well-being and the need for researchers to gain valuable findings.
Informed consent: At the start of a study ppts should be given info about the purpose of the study, so an informed decision can be made. Ppts told they can leave at any time. If the researcher doesn't reveal the aim of the study at the start, ppts must be informed at the end.
Deception: Ppts should not be lied to or misled about the aims of the study without justification. Mild deception is justifiable. Major deception is used but only when the benefits outweigh the ethical cost.
Protection from harm: Ppts' physical and psychological safety should be protected at all times. Stress and embarrassment are included in this. Ppts must be reminded they can leave at any time.
Privacy: Ppts have the right to control info about themselves. It is acceptable to observe in public places but a public place can be quite private.

Confidentiality: Personal data should be protected and respected.

15 of 36

Dealing with ethical issues

The BPS guidelines: A code of conduct that every professional psychologist in the UK has to follow in order to deal with ethical issues in their research.
Dealing with informed consent: Ppts sign a form that outlines the research procedures. If this is not possible at the start, they sign a form at the end of the study.
Dealing with deception and protection from harm: Ppts offered right to withdraw during the study and right to withdraw data at the end. Ppts are given a full debriefing at the end of an investigation to explain the true aims and/or reduce any distress. May be offered counselling.
Dealing with privacy and confidentiality: Keep data safe. All ppts should be anonymous. They can be referred to by a number or initials.

16 of 36


Interrviews: Face-to-face, real-time contact. Can take place over the phone or via text.

Structured: Interviewer reads out a list of prepared questions. Follow-up questions may also be prepared beforehand.

Unstructured: Interviewer has a general aim but does have a few questions prepared in advance. New questions based on previous answers.

Semi-structured: Some questions decided in advance. Follow-up questions emerge from the answers.

17 of 36

Evaluation for Interviews

  • Interviews produce a lot of info from each person. Especially true of unstructured. Means that unexpected results may occur.
  • Insight can be gained into thoughts and feelings. Observations only show what people do. Interviews provide a different perspective.
  • Data can be difficult to analyse. This is because of the breadth of info collected. Clear conclusions difficult.
  • People less comfortable giving personal info face-to-face, especially if questions are on a sensitive topic. Limits info collected.
18 of 36


A prepared list of written questions which can be completed face-to-face or in writing, over the phone or on the internet.

Open questions: Produces qualitative data as ppts can answer then how they want to.

Closed questions: Produces quantitative data - a fixed range of possible answers.

19 of 36

Evaluation for Questionnaires

  • Researcher gets info very quickly from lots of fo different people. This is because a questionnaire can be sent to many people. So generalisation is easier to make.
  • Data tends to be easier to analyse than interviews. This is when closed questions are used. Easier to draw conclusions.
  • Respondents may not give truthful answers. This social desirability bias affects the validity of responses. Reduces the validity of the data collected.
  • Questions may be unclear or leading. This makes it difficult to answer questions. Ppts' responses may lack validity.
20 of 36

Observation studies: Types

A researcher watches or listens to ppts and records data.

Naturalistic vs controlled: Naturalistic ob is recorded in a place where it would normally occur. Nothing is changed in the environment. If a level of control is needed, a controlled ob will be used. E.g. Zimbardo's study.

Covert vs overt: Covert ob = ppts are not aware their behaviour is being recorded. Overt ob = ppts are told in advance.

Ppt vs non-ppt: Ppt ob = researcher becomes part of the group s/he is studying. Non-ppt = researcher remains separate from the people s/he is studying

21 of 36

Observation studies: Behaviour

Categories of behaviour: The observer breaks the target behaviour into different types of behaviour. Each type/category should be observable and obvious so that it can be counted each time it occurs. E.g. flirting behaviours could be broken down into four categories: eye contact, smiling, laughing, touching.

22 of 36

Observation studies: Interobserver reliability

Two observers break the target behaviour into different types of behaviour.

To establish interobserver reliability the researcher creates categories of behaviour, observers record the same sequence of behaviour, then they compare their data and talk over differences and amend categories.

23 of 36

Evaluation for Observation studies

  • Observation data are based on what people do rather than what they say they do. Enhances the validity of the conclusions.
  • Observation studies look at real-life behaviour. People may not be aware they are being observed. Therefore, the data collected has greater validity.
  • There may be ethical issues. You cannot always gain people's consent when observing them in public places. Means that some observations should not be conducted.
  • May be some observer bias. Observer's expectations can influence what they see. Therefore, the observations may lack validity.
24 of 36


Correlations show how things are linked together. They tell us the strength and direction of the association or relationship between co-variables.

Co-variables: Correlations are quantitative so co-variable is reduces to a number. E.g. aggression could be reduced to how aggressive a person is on a scale of 1-10.

Scatter diagrams: Correlations are plotted on a scatter diagram. One co-variable goes on the x-axis and the other one goes on the y-axis. A dot is placed where they meet. Shows the direction and strength of the correlation.

25 of 36

Correlations: Types

Positive: As one co-variable increases the other co-variable increases. E.g. number of people in a room and noise.

Negative: As one co-variable increases the other one decreases. E.g. Number of people in a room and amount of space.

No correlation: No relationship between the co-variables. E.g. Time taken to complete a crossword and number of packets of crisps sold in a local shop each week.

26 of 36

Evaluation for Correlations

  • Correlations are a good starting point for research. If two variables are related this gives researchers ideas for future investigations.
  • Correlations can be used to investigate more complex relationships. E.g. the curvilinear relationship between people's alertness and time of day. This means correlation has many uses.
  • Correlations do not tell us whether one co-variable causes the other. This means it is not possible to show cause and effect. This limits the usefulness of the technique.
  • Intervening variables may affect the co-variables. This is because there is no control of EVs. This means it is possible to draw a conclusion that is wrong.
27 of 36

Case studies

An in-depth investigation of a single individual, group, event or institution. Often involves unusual or unexpected people/events but can also be used with everyday experiences.

Mostly qualitative data which expresses people's experiences in words. Case studies often involve interviews with the person's friends and relatives. May include quantitative data.

Longitudinal: Tends to take place over a long period of time. This may mean collecting data from the past or following the person for many years.

28 of 36

Evaluation for Case studies

  • Researchers tend not to have a specific aim. This means they are often open-minded and less blinkered by what they hope to discover. Increases the validity of the results.
  • A good method for studying rare behaviour that can't be investigated using experiments because there are only one or two people who could be studied. Gives a greater insight into topics that may not be studied in other research.
  • Often only focuses on one individual or event. This means it is difficult to generalise the results beyond the particular person or event being studied. Reduces the validity of the results.
  • The analysis may be subjective. The info collected may be biased by the researcher's own 'reading' of the case. Therefore, conclusions drawn may lack validity.
29 of 36


Reliability: Consistency. If you can repeat a measurement and get the same results, the measurement is reliable.

Quantitative methods:

  • Experiments - Can be controlled using standardised procedures, so each ppt has exactly the same experience.
  • Interviews & Questionnaires - If the same person answers the same questions in the same way then the interview/questionnaire is reliable. Closed questions achieve reliability more easily as choices are fixed.
  • Observations - Observer should produce the same observations if the same behaviour is watched/listened to twice. Interobserver reliability is when observers produce the same data.

Qualitative methods: Less reliable. Unstructured interviews and case studies are difficult to repeat in the same way.

30 of 36


Whether a result reflects 'real-world' behaviour.

Sampling methods: Do not always produce a sample that is reflective of the target population. Opportunity sampling is less likely to produce a representative sample, stratified sampling is more likely.

Experimental designs: Repeated measures design is influenced by order effects and can be overcome by counterbalancing. Independent groups design is affected by ppt variables and can be overcome by random allocation.

Quantitative methods: Lab experiments often involve artificial tasks or settings which are not used reflective of real life. Ppts are aware of being studied. Greater control enhances validity. Field experiments, Evs are not always controlled, tasks may be artificial and ppts may be aware of being studied - all reduce validity. Methods producing numerical data may lack validity as they reduce behaviour to a score. Qualitative methods: Greater validity as they provide unrestricted info. Data depends on subjective interpretation, reduces the validity of conclusions.

31 of 36

Qualitative and Quantitative data

Quantitative: Numbers - but can involve words and data about what people think and feel as long as the answers can be counted.

  • Data can be easy to be analysed. Can be converted to averages and then graphs and charts. Means groups of people can be easily compared.
  • Lacks depth and detail. Obtain little info about thoughts and abilities. It doesn't reflect how complex things are in the real world.

Qualitative: Words - but data can be turned into numbers. E.g. an interview about people about early childhood experiences, the number of times the words 'mother' or 'father' are said is counted.

  • Data is in more depth and detail. The researcher can gain more insight as the ppt is free to express their thoughts and feelings. Increases validity of data.
  • More difficult to analyse. May be hard to summarise material and draw conclusions. Means conclusions may be based on the researcher's opinion.
32 of 36

Primary and Secondary data

Primary: Data obtained first hand by the researcher for the purposes of a research project.

  • Suits the aim of the research. Is authentic because it comes first hand from the ppts themselves. Means the data may be more useful.
  • Takes more time and effort. The researcher must design and carry out a study rather than using readily available secondary data. Slows down the process and increases expense.

Secondary: Data from sources such as other studies or statistics. Has been collected by someone else for a different set of research aims.

  • Data is convenient to use. Is because it has already been checked and collected. Reduces expenses.
  • Data may not quite fit what the researcher wants and/or it may come from a poorly designed study. May reduce the validity of the research.
33 of 36

Measures of dispersion

Descriptive statistics: Express numbers in a way that gives an immediate impression of the overall pattern.

Range: Represents spread. Tells us whether a set of data are close together or spread out. Highest-lowest.

  • Easy to calculate
  • Distorted by extreme scores

Mean: + all scores and divide by the number of scores

  • Uses all data values when calculated.
  • Distorted by extreme scores, so less reflective.

Mode: Most common score

  • Very easy to calculate
  • Quite unrepresentative overall
34 of 36

Interpretation and display of Quan. data

Scatter diagrams: To display correlation. One co-variable on the x-axis, other on the y-axis. A dot is placed where these meet.

Frequency tables: A systematic way of presenting data, organised in rows and columns. Displays how often an event occurred, using tallies.

Frequency diagrams:

  • Histogram: Continuous data. No spaces between bars.
  • Bar chart: Data not continuous. Could be placed in any order.
  • Normal distribution: Symmetrical spread of frequency data that forms a bell-shaped curve. Mean, the median, mode at the same point.
35 of 36


Decimals: A way to represent fractions out of 10, 100, 1000 etc.

Fractions: A decimal is another way of writing a fraction. To reduce to lowest form, identify a number that divides evenly into the top and bottom numbers.

Ratios: Another way to express a fraction.

Percentages: Fractions out of 100. 12% means 12 out of 100

Arithmetic mean: Same as normal mean.

Standard form: Move the d.p. to get a value between 1 and 10. Work out how many times you move the d.p. to the left or right. e.g. 3,280,000 can be written as 3.28 x 10^6

Significant figures: Round large numbers to the nearest 1000th, 10000th etc or nearest 10th, 100th etc. 32462 becomes 32,000 to 2 significant figures.

Estimate results: A rough calculation.

36 of 36


No comments have yet been made

Similar Psychology resources:

See all Psychology resources »See all Research methods and ethics resources »