Research Methods

Research Methods

HideShow resource information

Aims and Hypotheses

Aim: general purpose of an investigation, what you are trying to achieve in the investigation.

Hypothesis: a precise, testable statement or prediction about the expected outcome of an investigation.

Null Hypothesis: a prediction that states that results are due to chance and are not significant in terms of supporting the idea being investigated. E.g. there is no evidence that there is a difference between groups in the amount they remember.

Research Hypothesis: a prediction that states that results are not due to chance and that they are significant in terms of supporting the idea being investigated. E.g. there is evidence that there is a difference between groups in the amount they remember.

1 of 54

One-tailed Hypothesis: is a directional hypothesis. E.g. instead of saying there will be a difference between groups in the amount they remember, you predict which group will remember the most.

Two-tailed Hypothesis: is one in which the direction of results is not predicted. E.g. you may predict a difference between groups, but have no idea which ay the difference will fall.

2 of 54

Design Issues

Factors associated with good design.

-The following factors are important to consider when designing an investigation.

  • A pilot study is test run on a few participants; this enables you to check for design faults before carrying out an investigation on larger scale, this is a routine procedure.
  • Reliability of results is very important, so if a study is replicated the findings should be similar.
  • Validity, does a test measure what it was designed to measure, for example, do IQ tests really measure intelligence?
  • lnternal validity, extent to which study is free of design faults, which may affect results.
  • Ecological validity, a type of 'external validity'. This means the extent to which generalisation can be made from the test environment to other situations.
3 of 54

Design Issues

Repeated Measures

Testing the same group of people in different conditions, the same people are used repeatedly.


Avoids the problem of participant variables

  • Fewer people are needed.


  • Order effects are more likely to occur
  • Counterbalancing
  • Randomisation
4 of 54

Design Issues

Independent groups:

Testing separate groups of people, each group is tested in a different condition (one of them being controlled).


  • Avoids order effect. If a person is involved in several tests they may become bored, tired and fed up by the time they come to the second tests, or becoming wise to the requirements of the experiment.


  • More people are needed than with the repeated measures design
  • Difference between participants in the groups may affect results, for example, variations in age, sex or social background. These differences are known as participant variables.
5 of 54

Design Issues

Matched pairs

Testing separate groups of people-each member of one group is same sex, age, or social background as a member of the other group.


  • Reduces participant variables
  • Avoids order effects


  • Very time-consuming trying to find closely matched pairs
  • Impossible to match people exactly, unless identical twins

In each case, there are one or more experimental groups, where the independent variable has changed and a control group where the independent variable has not changed.

6 of 54

Design Issues

Counterbalancing: alternating the order in which participants perform in different conditions of an experiment. E.g. group 1 does 'a' then ‘b’; group 2 does 'b' then 'a' to eliminate order effects.

Randomisation: material for each condition in an experiment is presented in a random order; this is also to prevent order effects.

7 of 54

IV's, DV's and confounding variable

Independent variables (IV): variable the experimenter manipulates – assumed to have direct effect on the dependent variable

Dependent variable (DV): variable the experimenter measures, after making changes to the IV, which are assumed to affect the DV.

Extraneous variables (EXV's): other variables, apart from the IV that might affect the DV. They might be important enough to provide alternative explanations for the effects, for example, confounding variables.

8 of 54

IV's, DV's and confounding variable

Measuring the dependent variable

When planning their research, psychologists must decide how to measure the dependent variable. These measures might be, for example, the time spent looking, the number of words recalled, the time taken to complete a task, the number of positive reinforcements or the level of shock given.

Measuring behaviour is more complex: for example, measuring aggression. The psychologists’ needs to define what behaviours are to be considered aggressive. For example; pushing-how hard? Facial expressions- which ones? Verbal behaviour – what kind? How will the behaviour be measured – how often it occurs or how long it lasts, or both?

9 of 54

IV's, DV's and confounding variable

Controlling other variables

To be confident that the IV has indeed caused the DV, the researcher must control all other aspects of the experiment – must control other variables. Some of the other variables, which researchers need to consider, are:

· Situational variables – these are aspects of the environment which may affect the participants behaviour in the experiment, such as the variation in light. Other variables could be the time of day, whether others are present, or background noise. Situational variables should be controlled to ensure they are the same for all participants.

10 of 54

IV's, DV's and confounding variable

  • Participant variables – this refers to the ways in which each participant differs from the others, and how this could affect the results. For example, if participants doing a word memory task were tired, dyslexic or had poor eyesight, this could affect their performance and the results. Researchers try to ensure that such variables are evenly distributed between the research groups. The method of doing this is described under Experimental design.
  • Standardized procedure – this means that each participant is treated in exactly the same way, each doing exactly the same tasks, with the same materials, in exactly the same order. This reduces the variables in the procedure.
  • Standardized instructions – each participant must be given exactly the same instructions, ideally by the same person and in the same way. If some participants were given instructions, which included demonstrations of how to do a task, and others were not, this could affect the results. One way of ensuring standardization is to provide written instructions, which should be simple and clear.
11 of 54

Ethical Issues

Ethics are the standards of behaviour that we use in our dealings with others. To behave ethically when conducting research, we must treat others with respect and concern for their well being, we must not take advantage of their trust or their lack of knowledge. Unethical behaviour discredits psychology and the work of other psychologists. People my refuse to help with future research if they have been offended by unethical experiments. Ethical concerns apply to both humans and animals, though our focus here is on humans.


You must work within your limits an must be very cautious about giving advice, as people tend to think that anyone studying psychology is able to advise them on their problems.

12 of 54

Ethical Issues


Participants should be volunteers and be informed what the research is about no what they will be asked to do before they re asked to consent to take part – this is informed consent. People should not be deceived into taking part by being told the study is about something else, or by ignoring aspects of the study, which might affect their willingness to take part. Information should be withheld only if there is no other way of carrying out the research. This deception should not persuade the participant to give consent, which they would later regret. Full details should be given during the participants debriefing.

Participants should be told they have the rights to withdraw from the research study at any time, and be reminded of this right during a long study or if the participant appears to be distressed or uncomfortable. The researcher should stop the study if participants are uncomfortable or distressed, even if they have not asked to withdraw.

13 of 54

Ethical Issues

After the research, participants should be debriefed so they know what the study was about; their own results should be available to them and the researcher must answer any questions.

Some people may be unable to give informed consent, such as children or those with special needs. Nevertheless, they must be asked if they are willing to help you, but full consent must be gained from whoever is responsible for that person, such as a parent or career. These people must be given full information, just as if they were the participants, before being asked for their consent.

The responsible authority in the institution must also approve research, which is to be carried out in an institution such as a school, factory or supermarket.

Consent is not necessary when observing people in public, as they could be observed by anyone. Nevertheless, such observations could be intrusive or unnerving, so they should be carefully planned to avoid this.

14 of 54

Ethical Issues


Information about the identity of the participants and any data gained from them must remain confidential. It is unethical to give information about a participant to someone else (such as another researcher) without that persons consent. Data should not be accessible to others and a number or letter in the research report should identify participants, in order to protect their anonymity.

15 of 54

Ethical Issues


Researchers must ensure that any equipment is safe to use, and that participants are not asked to do anything that is illegal or might cause them physical harm, nor should participants experience psychological harm, such as distress, fear, anger, embarrassment, offence. Great care should be taken with children, as they are particularly vulnerable and may be unhappy or harmed by experiences that adults would find problem-free.

All participants should leave the investigation feeling as good about themselves as when they started it.

Researchers should be honest about their abilities and competence. They must never make u their own data or use someone else’s data and claim it is their own.

16 of 54

Sampling methods

Sampling techniques are very important. Several types of sampling are used according to the type of study and the subjects you want to target.

Random sampling:

Everyone in the entire target population ha an equal chance of being selected.


  • Representative
  • Least bias because it is an equal chance method
  • The researcher has no influence/control over who gets picked


  • Very time-consuming
  • Costly
  • Difficult to do if target population is big
17 of 54

Sampling methods

Opportunity sampling:

Uses people from target population available at the time


  • Doesn’t have to be planned
  • Simple
  • Cheap
  • Not time-consuming


  • Not representative
  • Only restricted to people who are available at that time
18 of 54

Sampling methods

Volunteer sampling:

Researcher places an advert, participants respond to advert and volunteer to take part by contacting the researcher.


  • Very little time and effort required from researcher


  • Likely you get a biased, unrepresentative sample because will only appeal to certain types of people.

Stratified sampling:

Divides target population into groups, people in sample from each group in same proportions as population.

19 of 54

Sampling methods

Systematic sampling:

Chooses subjects in a systematic way, for example, every 10th person from a list or register.


  • To being with, all population is involved; this is an improvement of opportunity and volunteer sampling.


  • Not everyone has an equal chance of taking part so unrepresentative
  • Time-consuming
  • Still possible to get a biased sample
20 of 54

Sampling methods

Quota sampling:

Sample must include same proportions


  • Guaranteed to get a mix of the population


  • Complicated
  • Time-consuming
  • More effort
  • Expensive
21 of 54

Reliability and validity

Reliability: if repeated procedures are replicated and produce similar results

Internal validity: if test measures what it set out to measure then internal validity is high. If there are confounding variables not being controlled then it is low.

External validity: two types:

Ecological – low if not like real life settings, if artificial/unnatural

Population – wide range of people would have high population validity. If group is biased or not varied then population validity is low.

22 of 54

Reliability and validity

Ways of ensuring reliability

1. Test re-test reliability: Participants complete the same test twice, with a gap (i.e. a week) in-between. If they score similar results on both occasions, then the method is reliable.

2. Split half reliability: This involves splitting the test into two halves. We then carry out a correlation on the two halves, in order to ensure that both halves of the test are of equal difficulty, if they are, then the test is reliable.

3. Observer (inter-rater) reliability: Used when carrying out an observation. To ensure observer reliability, you must have at least two observers watching and recording what they see in the same way, using the same forms. If all observers record the same things, then their observations are reliable.

23 of 54

Reliability and validity

Ways of ensuring validity

1. Content validity: This involves an independent expert examining the content of the test/research method to see if it looks like it is measuring what it is suppose to measure. If they agree that it is measuring what it is suppose to, then the test/method has good validity.

2. Concurrent validity: This involves comparing a new test with an already established test designed to measure the same thing. If scores are similar, then the new test is valid.

3. Predictive validity: This involves checking validity by seeing if future behaviour is consistent with what we could predict based on our test, for example, if someone is diagnosed a schizophrenic we would expect them to goon to show further symptoms of this diagnosis.

24 of 54

Levels of Data

When choosing which statistical test to use we must be able to identify the level of our data.

Nominal level of measurement:

This is where we have 'categories', e.g. students, teachers, non-teaching staff

Nominal data is where we count the numbers or frequencies within each category

The nominal level of measurement provides us with the least amount of, and most basic, quantitative information. It is purely 'how many times' something occurs. It is linked to the mode.

25 of 54

Levels of Data

Ordinal level of measurement:

  • Ordinal data tells us who came 1st, 2nd, 3rd and so on in a test, for example. However, it does not tell us how far ahead the winner was from second place; it tells us nothing about distances between positions, as the intervals may be unequal.
  • Ordinal measurement is the ranking of data, for highest to lowest or vice versa. It is linked to the median.
  • The ordinal level of measurement provides more information than nominal data, but less information than interval level data.
26 of 54

Levels of Data

Interval level of measurement

  • Interval measurement allows us to talk about the distances between points on a scale. It is assumed that the measurements are on a scale with approximately equal interval between them.
  • Temperature is an example of interval data.
  • However, interval level data does not have a true zero. E.g. whichever method of measuring temperature we use (Fahrenheit or centigrade), 0 degrees does not mean no temperature. Minus values are possible.

Ratio level of measurement

  • Ratio data provides us with the strongest and most precise method of measurement. Ratio scales have a true zero, e.g. time in seconds and distance in cm. negative numbers have no meaning.
27 of 54

Descriptive statistics

Quantitative research: gathers data in numerical form and is concerned with making 'scientific' measurements. Quantitative data analysis uses a barrage of inferential statistical tests.

Qualitative research: gathers information that is not in numerical form, for example, diary accounts, open-ended questionnaires, unstructured interviews and unstructured observations. Is useful for studies at the individual level, and to find out, in depth, the ways in which people think or feel.

28 of 54

Descriptive statistics

Measures of central tendency

Mean: arithmetic average, all values in a set of data are added together and divided by the number of values.


  • Most representative
  • Takes into account of all scores


  • Affected by outliners
  • Interval or ratio data only
29 of 54

Descriptive Statistics

Mode: most frequent score in a set of data can be bi-modal (more than one).


  • Easy to calculate
  • Unaffected by extreme scores


  • Doesn’t take account of every score
  • Not useful for small data sets
  • Use with nominal data only
30 of 54

Descriptive Statistics

Median: score in the middle when items are ranked smallest to largest. Two middle scores are added together then divided by two.


  • Used with ordinal, interval or ratio level data
  • More representative then the mode
  • Not affected by very large or small scores


  • Does not represent all scores
31 of 54

Descriptive Statistics

Measures of dispersion (spread)

Range: difference between largest and smallest scores, quote biggest and smallest scores, or take smallest from biggest score and quote this figure.


  • Easy to calculate
  • Use with ordinal, interval or ratio level data


  • Does not indicate how tightly/widely spread scores are.
32 of 54

Descriptive Statistics

Standard deviation: average variation around the mean, larger value indicates wider spread, relies on data with equal interval between points.


  • Very sensitive, s it uses all scores


  • Difficult to calculate
  • Interval or ratio level data only
33 of 54

Descriptive Statistics

Graphs and charts

Graphs and charts give a quick visual impression of any patterns or trends in your results. They should be used to help summarise your results.

Bar charts

  • Used for nominal data (data in categories)
  • The x-axis (frequencies are usually on the y-axis) does not need to show a complete scale (if showing categories)
  • lThere should be gaps between the bars.
34 of 54

Descriptive Statistics


  • Used for interval or ordinal data
  • No intervals (if data is grouped) are missed, even if they are empty. Class intervals are represented by their mid-point at the centre of each column.There are no gaps between columns

Frequency polygons.

  • Used for interval or ordinal data
  • All class intervals are represented
  • Instead of columns, a line is used to join the mid-point of each class interval.
35 of 54

Inferential statistics

Level of measurement MOCT Testing for a difference Testing for a correlation Related designs (Repeated & matched) Unrelated designs (Independent) Nominal Mode Sign Test Chi-Squared Ordinal Median Wilcoxon Matched Pairs Signed Ranks test Mann-Whitney U Test Spearman's Rank Order Correlation Coefficient Interval/Ratio Median / Mean Wilcoxon Matched Pairs Signed Ranks test Mann-Whitney U Test Spearman's Rank Order Correlation Coefficient

36 of 54

Levels of Significance

We use statistical tests in order to decide whether the results we have found are due to the variable we have manipulated or whether the results are due to chance alone.

If we accept that the results we have found are due to the variable we have manipulated, then we would reject the null hypothesis. Our results are significant.

If we accept that results we have found are due to chance alone then we would accept the null hypothesis. Our results are not significant.

37 of 54

Levels of Significance

Knowing whether or not to accept/reject a null hypothesis lies in the level of significance chosen for the statistical test.

A significant result is one, which has a low probability of occurring by chance.

The level of significance selected depends upon the level of confidence required. Often, a 95% probability that the results are not due to chance is accepted. This significance is shown as p0.05.

Where high certainty is required, such as in drug trials, then a stricter level of significance needs to be selected, e.g. p0.001 (99.9%).

38 of 54

Levels of Significance

However: we can never be 100% certain that results are not due to chance. There are two errors that may occur.

1. Type 1 error: the null hypothesis is wrongly rejected- the results are, in fact, due to chance factors and not due to the manipulation of our variable. The reason for a type 1 error is that a too lenient level of significance has been used (e.g. p0.01).

2. Type 2 errors: the null hypothesis is wrongly accepted – the results are in fact, due to the manipulation of our variable and not due to chance factors. The reason for a type 2 error is that a too strict level of significance has been used (e.g. p0.01 or p0.001).

It is often considered preferable to run a higher risk of making a type 2 errors rather than a type 1 error.

39 of 54

Levels of Significance

P< than 0.05 /0.01 (<5%/1%)

Refers to achieving a significance level of at least 5%, i.e. the probability of results such as these occurring through chance is no greater then 0.05.

P> than 0.05 /0.01 (>5%/1%)

Refers to achieving a significance level of more than 5%, i.e. the probability of results such as these occurring through chance is greater than 0.05.

40 of 54

Types of experiments


Controlled and scientific experiments. Researcher is able to control all variables, except the IV, which is manipulated.


  • Replication: can run study again to check finding
  • Control: those that might have an affect can be minimised.
  • Scientific and test cause and effect


  • Standardisation may create demand characteristics – participants may respond to what they think is expected of them
  • Artificial – experiments might not measure real life behaviour
41 of 54

Types of Experiments


Researcher manipulates the IV, but the experiment takes place in real life natural settings.


  • Ecological validity – less artificial
  • Demand characteristics – avoided if participants don’t know if they are in a study


  • Less control – other variables might be more likely in a study
  • Ethics – participants didn’t agree to take part
42 of 54

Types of Experiments


Is when the IV occurs naturally in real life, the researcher cannot ‘create’ different for the purpose of the experiment


  • Researcher can see people behave naturally as a result of little manipulation
  • Less demand characteristics
  • Passive researcher


  • Cannot be sure if results are due to manipulation or other uncontrolled variables
  • No control
43 of 54


Shows a relationship between two variables, both may already be occurring, but in order to find out if they are related the researcher must measure the variable and then calculate a correlation. The relationship can be plotted on a scatter gram.


  • Replicable – data can come from questionnaires
  • Shows relationship – positive (occurs when variable increases as the other increases) and negative (occurs when variable increase as the other decreases)
  • Objective


  • It is not possible to find out whether one thing causes another, only whether the two variables are related
  • Quantitative
  • No cause and effect
44 of 54


Written answers that depend on respondents being able to read and understand correctly. May be distributed by hand, by post or a distribution point. Once completed they can be returned by post or collected by hand. Questions can be closed or open-ended.


  • Quick and easy
  • Easily replicated
  • Very large sample can be used
  • Loads of information can be gathered
  • People who are geographically distant can be studied


  • Sample will be biased because it relies on people returning them
  • Demand characteristics
  • Lack population validity
  • People may not give honest answers
45 of 54


Structured (planned)

Consists of fixed questions with a limited range of possible answers, much like a questionnaire


  • Fastest to complete and provides data which is easy to quantify and analyse


  • Drawbacks of closed questions
  • High demand characteristics
46 of 54


Semi-structured (partially planned)

Comprise open-ended questions, which cover the information the researcher wants to gain.


  • Researcher can be flexible in what questions are asked
  • Can gain in-depth and accurate information from respondents


  • More difficult to compare answers
  • Demand characteristics
  • Socially acceptable answers
47 of 54


Unstructured (not planned)

Are the most informal and in-depth technique, they enable the interviewer to re-phrase questions if necessary, to ask follow-up questions or clarify answers that are ambiguous or contradictory.


  • Provides detailed information
  • Researcher can clarify questions


  • Results cannot be generalised to the population as a whole, and it is possible that the interviewer may bias the response or misinterpret the answers that are given.
48 of 54


When psychologists observe, they watch and analyse participant’s behaviour. It is usual to have more than one observer because behaviour is complex and the observer may be biased. Behavior is noted on an observation schedule. Researchers may run a pilot study.

49 of 54



Here the researchers have no control; they are not participants but look at behaviour, which occurs naturally, as it would in a school playground, for example. Before starting the study, the observers try to become familiar to those they are observing, in order to minimise the effect that their presence may have.


  • Behaviour occurs in its natural setting
  • Observation provides very detailed information
  • It can be used as a starting point for further, more controlled, research
  • It can be used when other methods might be unethical


  • The presence of observers could influence the behaviour of those being observed
  • It is difficult for observers to be completely objective
  • Many variables could affect behaviour so that it is not possible to draw any conclusions
50 of 54



This type of observation is frequently used with children or in social psychology research. Controlled observations are essentially laboratory experiments because an IV is being tested and control is possible. In order that control can be exerted and so that the behaviour is easier to observe. The observations may take place in an artificial setting such as research laboratory.


  • By controlling some variables, it is possible for the researchers to draw conclusions from their observation


  • The unfamiliar setting may affect participant’s behaviour, making it less natural.
51 of 54



Here the observer becomes one of the groups of participant that he or she wishes to observe. The observer may tell the others they will be observed (an overt observation), or may pretend to be one of the group and not inform them that they are being observed (a covert observation).


  • Allows researchers and observe people in a natural setting and gain some understanding of the causes of their behaviour
  • It is particularly useful for studying the way people behave when they are in groups


  • Remembering accurately may be difficult as unable to take notes
  • Observer may interpret or record information in a biased way
  • Ethical guidelines such as deception, consent and confidentiality may not be maintained.
52 of 54

Case study

Is an in-depth study of one person or a small number of people? It may include interviews of the person being studied as well as others who can provide information about the persons past or present experiences and behaviours.


It gives a detailed picture of the individual

  • It can be useful in treating individual problems,
  • It helps in discovering how a persons past ma be related to the present
  • It can form the basis for future research, by studying those who re unusual, psychologists can discover more about what is usual.


  • It relies on memory, which may be poor or distorted
  • The information gained about one person cannot be generalised to apply to other people
  • It relies on participants telling the truth
  • The interviewer may be biased if they are looking for certain information.
53 of 54


Deterministic – links to cause and effect (core stimulus always reduces same response). Leading questions give a determined response because of only giving a limited choice of answers.

Objective / subjective – objective: scientific unbiased opinion / Subjective: bias, involves interpretation.

Reductionism – the simplest explanation is the best

Internal/ external – internal: body, personally / external: environment

Scientific - controlled

54 of 54


No comments have yet been made

Similar Psychology resources:

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