Research Methods

HideShow resource information

Types of research

Quantitative research: uses methods that measure amounts of behaviour, usually by assigning a numeric value to what is being measured (the quantity). Quantative data is collected as numbers.

Qualitative research: measures what behaviour is like (the quality) and usually results in descriptive data. Qualitative data is collected as descriptions.

1 of 37

Experimental design- Laboratory experiments

Laboratory experiments: Researchers try to control all the variables except the one that is changed between the experimental conditions. Variable that is changed is called the independent variable (IV) and the effect it may have is called the dependent variable (DV). IV is manipulated and its effect (the DV)  is measured. Controlled and often in artificial settings.


  • Allow extraneous variables that might affect the IV or the DV can be minimised, the researcher can be sure that any changes in the DV are the result of the changes to the IV.
  • High levels of control make it possible to measure the effect of one variable on another; Statements about cause and effect can be made.
  • Replicateable to check the findings with either the same or a different group of participants.


  • May not measure how people behave outside the laboratory in their everyday lives. Experimental settings and tasks are contrived; findings may have low internal validity.
  • Aspects of the experiment may act as cues to behaviour that cause the participants to change the ay they behave (demand characteristics) because of what they think is being investigated or how they think they are expected to behave; this can lead to invalid results.
2 of 37

Experimental design- Field experiments

Field experiments: Is a way of conducting research in an everyday enviroment (e.g. school or hospital) where one or more IVs are manipulated by the experiementer and the effect on the DV is meaured. Increase in the naturalness of the setting and a decrease in the level of control that the experimenter is able  to achieve. Key difference is the extent to which participants know they are being studied. Participants behaviour is more natural.


  • Allow psychologists to measure how people behave in their everyday lives; high external validity.
  • Manipulation of the IV and some level of control make it possible to measure the effect of one variable on another, Statements about cause and effect can be made.
  • Reduces the probability that their behaviour results from demand characteristics depending on the experimental settings.


  • Not always possible to control for extraneous variables that might affect the IV or the DV, cannot always be sure that any changes in the DV are the result of changes in the IV.
  • Field experiments can be difficult to replicate and thus it may not be possible to check the reliability of the findings.
  • May not be possible to ask participants to their informed consent, maybe decieved and may not be debriefed which breaches of British Psychological Society ethical guidelines.
3 of 37

Experimental design- Natural experiments

Natural experiments:

  • Is one in which, rather than being manipulated by the researcher, the IV to be srudied is naturally occurring.
  • Naturally occurring variables are gender, age, ethnicity, occupation and smoker or non-smoker.
  • When the IV is naturally occurring, participants cannot be randomly allocated between conditions.
  • A natural experiment may take place in a laboratory or in a field experimental setting.


  • Natural experiments allow psychologists to study the effects of the IV's that could be unethical to manipulate.
  • Participants are unaware of the experiment, and the task is not contrived, reearch may have high internal validity.


  • Since participants cannot be allocated randomly between conditions, it is possible that random variables (individual differences other than the IV) can also affect the DV; this could lead to low internal validity.
  • Natural experiments can be difficult with a different group of participants, it may not be possible to check the reliability of the findings.
4 of 37

Correlations and correlational coefficients

Correlation is a statistical technique used to calculate the correlation coefficient in order to quantify the strength of relationship between two variables. E.g. there is a relationshp between aggressive behaviour and playing violent video games. Studies that use correlational analysis cannot draw conclusions about cause and effect. Just because two events occur together does not mean that one necessarily cause the other.  

Correlation coefficient is a mathematical measure of the degree of relatedness between sets of data. A perfect positive correlation, indicated by +1 where variable X increases and variable Y increases. A perfect negative corelation, indicated by -1, is where variable X decreases and variable Y decreases.


  • Correlational analysis allows researchers to calculate the strength of a relationship between variables as a quantitative measure.
  • Where a correlation is found, it is possible to make predictions about one variable from the other.


  • Researchers cannot assume that one variable causes the other.
  • Correlation between variables may be misleading and can be misinterpreted.
  • A lack of correlation may not mean there is no relationship because the relationship could be non-linear.
5 of 37

Analysing correlational data

  • If both variables increase together this is a positive correlation. (0 to +1)
  • If one variable increases as the other decreases, this is a negative correlation. (0 to -1)
  • If no line of best fit can be drawn, there is no correlation. (0) 
6 of 37

Naturalistic observations

Naturalistic observations they watch people's behaviour but remain inconspicuos and do nothing to change or interfere with it.


  • Behaviour can be observed in its usual setting and there are usually no problems with demand characteristics unless the situation in which the partcipants are being observed has been specially contrived.
  • Useful when researching children or animals.
  • Can be a useful way to gather data for a pilot study.


  • No explanation for the observed behaviour is gained because the observer counts instances of behaviour but does not ask participants to explain why they acted as they did.
  • Obserers may 'see what they expect to see' (observer bias) or may miss, or misinterpret, behaviour.
  • Studies are difficult to replicate.
7 of 37

Interviews and questionaire surveys

Asking questions about behaviour is that we all like others to think well of us. As a result, what we say about our behaviour and how we actually behave may be different. (Social Desirability Bias)

Structured interviews: All participants are asked the same questions in the same order. Structured interviews can be replicated and can be used to compare people's responses. Can be time consuming and require skilled researchers. People's responses can be affected by social desirability bias.

Unstructured interviews: Participants can discuss anything freely and the interviewer devises new questions on the basis of answers given previously. They can provide rich and detailed information, but they are not replicated and people's responses cannot be compared. Require trained interviewers.

Questionaires: Questionaires are usually written, but they can be conducted face to face or completed over the telephone or on the internet. Printed questionaires are completed by participants. They are similar to structured interviews in that all participants are asked the same questions in the same order. Usually restrict participants to a narrow range of answers. Questionaires are pratical way to collect a large amount of information quickly and they can be replicated. Problems can arise if the questions are unclear or if they suggest a 'desirable' response, as response can be affected by social desirability bias. Closed questions are used, participants cannot explain their answers.


8 of 37

Case Studies

A case study is a very detailed into the life and background of one parson ( or a small group of people). Case studies involve looking at past records asking other people about the participant's past and present behaviour.


  • Give a detailed picture of an individual and help to discover how a person's past may be related to his/her present behaviour.
  • Can form a basis for future research.
  • Studying the unusual we can learn more about the usual.


  • Can only tell you about one person so findings can never be generalised.
  • Interviewer may be biased an/or the interviewee may not tell the truth.
  • Retrospective studies may rely on memory, which may be inaccurate or distorted and past records may be incomplete.
9 of 37

Investigation design- Specification content

  • Aims
  • Hypotheses, including directional and non-directional
  • Experimental design ( independent groups, repeated measures and matched pairs)
  • Design of naturalistic observations, including development and use of behavioural categories
  • Design of questionaires and interviews
  • Operationalisation of variables, including independent and dependent variables
  • Pilot studies
  • Control of extraneous variables
  • Reliability and validity
  • Awareness of the British Psychological Society Code of Ethics
  • Ethical issues and the ways in which psychologists deal with them
  • Selection of participants and sampling techniques, including random, opportunity and volunteer sampling
  • Demand characteristics and investigator effects
10 of 37

Ethical research

Psychological research seeks to improve our understanding of human nature and ethics are standards regarding what is right or wrong.

An ethical issue occurs when there is conflict , for example, between what the researcher wants in orderto conduct valid or meaningful research and the rights of participants.

11 of 37

Ethical guidelines- BPS

  • Consent: Should inform all participants the objective of the investigation, the investigator should inform the participants of all aspects of the research or intervention that might reasonably be expected to influence willingness to participate.
  • Deception: Misleading of participants is unacceptable if the participants are typically likely to object or show unease once debriefed. Appropriate consultation must precede the investigation.
  • Withdrawal from the investigation: the paticipants should have the right to withdraw at anytime.
  • Confidentiality: information obtained about a participant during the investigation is confidential unless otherwise agreed in advance.
  • Protection of participants: investigators have a responsibility to protect participants from physical and mental harm during the investigation.
  • Obervational research: Studies based upon observation must respect the privacy and psychological wellbeing of the individuals studied. Observational research is only acceptable in situations where those observed would expect to be observed by strangers.
  • Giving advice: A participant solicts advice concerning educational, personality, behavioural or health issues, caution should be exercised.
12 of 37

The dilemma of deception

Can be argued that if participants are not decieved about the true aims of a study, their behaviour is affected and thus does not reflect how they would really behave in their everyday lives.

Dilemma for researchers is to design and conduct research that accurately portrays human behaviour while at the same time ensuring that the do not breach the ethical guidelines.

Researchers may not solve this dilemma by undertaking a cost-benefit analysis of the research before they commence however this can cause problems:

  • Impossible to calculate the costs and benefits before a study as the researchers cannot predict events accurately.
  • After a study it is difficult to calculate the costs and benefits as this may depend on when and who makes the judgement. The value of some researchers may judge the benefits and costs differently.
  • Approach may encourage researchers to ignore the rights of the individual participants on the grounds that many more people may benefit.
13 of 37

The dilemma of informed consent

All participants should be asked to give informed consent prior to taking part in research, however in some situations where deception may be used it is not possible to obtain fully informed consent from the participants of the study.

Presumptive consent: presumptive consent is gained , people who are members of the population who are to be studied are informed of the details of the study and asked whether is they were to participate they would consider the research acceptable.

Prior general consent: involves asking questions to people who have volunteered to participate, before they are selected to take part e.g. 'Would you mind being involved in a study in which you were decieved?' Participants who say they 'would not mind' may later have been selected to participate.


14 of 37

Research methods and ethical issues

Laboratory experiment: Participants feel reluctant to withdraw and may feel they should do things they would not normally do, participants ay be decieved.

Field experiment: May be difficult to obtain informed consent and participants may not be able to withdraw also to debrief the participants.

Natural experiment: Confidentiality may be a problem as the sample studied may be identifiable where naturally occurring social variables are studied ethical issues may arise when drawing conclusions and publishing the findings.

Correlational studies: ethical issues can arise when researching relationships between socially sensitive variables because published results can be misinterpreted as suggesting 'cause and effect.'

Naturalistic observations: informed consent is not being gained, people should only be observed in public places and where they would not be distressed to find they were being observed, an ethical issue may arise in terms of protecting confidentiality.

Iinterviews and questionaires:  Should not be asked embarrassing questions and should be reminded that they do not have to answerany questions if they do not wish to, protecting confidentiality is important.

15 of 37

Aims and Hypotheses

Research aim is a general statement of the purpose of the study and should make clear what the study intends to investigate, the aim states the purpose of the study.

Hypothesis states precisely what the researcher believes to be true about the target population, often generated from a theory and is testable statement.

Experimental hypothesis is used when experimental research is being conducted otherwise the term alternative hypothesis is used. The experimental hypothesis  states that some difference will occur; that the IV will have a significant effect on the DV.

Null Hypothesis is a statement of no difference or of no correlation- the IV does not affect the DV- and is tested by the inferential statistical test.

Directional hypothesis is termed a 'one-tailed hypothesis' because it predicts the direction in which the results in which the results are expected to go.

Non- directional hypothesis is termed a 'two-tailed hypothesis' because the researchers expect that the IV will affect the DV they are not sure how though.

16 of 37

Experimental design

The independent groups design: Different participants are used in each of the conditions.


  • No participants are 'lost' between trials.
  • Participants can be randomly allocated between the conditions to distribute indivdual differences evenly.
  • No practise effects.


  • Needs to be more participants.
  • May important differences between the groups.
17 of 37

The repeated measures design

The repeated measures design: the same group of participants are used in each conditions.


  • Requires few participants
  • Controls for individual differences between participants


  • It cannot be used in studies in which participants in one condition will affect responses in another
  • It cannot be used in studies where an order effect would create a problem.

Order effects: if a participants did one task and then did it again they may be better because they have had practise or worse as they may have lost interest or are tired.

Counterbalancing technique: the group of participants is split and half the group complete condition A followed by condition B the other half completes condition B followed by condition A; any order effects are balanced out.

18 of 37

The matched pairs design

Seperate groups of participants are used who are matched on a one-to-one basis on characteristics such as age to control for the possible effect of individual differences.


  • Controls some indivdual differences
  • Can be used when a repeated measures design is not appropriate.


  • A large number of participants are needed
  • Hard to match on some characteristics e.g. personality.
19 of 37

Operationalisation of variables

Operationalisation means being able to define variables in order to manipulate the IV and measure the DV . For example performance on a memory test may be operationalised as 'the number of words remembered.' Both the IV and the DV need to be precisely operationalised otherwise the results will not be valid and cannot be replicated because a researcher would not be able to set up a sudy to epeat the same measurements. 

20 of 37

Standardised instructions and procedures

All participants should be told what to do in exactly the same way and all participants should be treated the same way. 

21 of 37

Control of extraneous variables

Any variables that change between the conditions, other then the IV, are difficult to control i.e. how tired the participant gets. 

Enviromental variables that may affect participants' performance such as the time of day or locations should be controlled.

22 of 37

Pilot studies

To establish whether the design works and that participants can understand the instructions a pilot study must be carried out which is a trial run with a small number of participants.. 

This allows the researchers to make necessary adjustments and to avoid wasting valuable resources. 

23 of 37

Design of naturalistic observations

  • How different behaviour should be categorised in order to measure causes and effects. 
  • Preliminary observations: With an observational study the formulation of the hypothesis and decisions about how best to categorise the behaviour should be undertaken in conjuction with preliminary observations of a small sample of participants. 
  • Sampling behaviour: You need to decide who you are going to observe, how and when.
    • Focal sampling: records behaviour at one individual time. One disadvantage of this method is that your 'focal' person may not engage in any of the behaviour categories of interest and the person being watched may become aware that you are watching them. 
    • Scan sampling: consits of rapidly scanning a group at a specific time interval to record which behaviour(s) is occuring. This allows observation of a large number of individuals but the disadvantage that certain individuals or behaviour may be more conspicuous than others, leading to bias recording. 
    • Time sampling: divides the observation period up into sample intervals e.g. every 2 minutes. A watch can be used can indicate the time interval. The observer makes a note of behaviour occurring at each time interval on a pre-prepared tally chart. 
24 of 37


Reliability: reliability of results means consistency. 

Internal reliability: refers to how consistently a method measure within itself. To test for internal relliability, researchers may use the split-half technique in which half of the scores are compared with the other half to see how similar they are. 

External reliability: refers to the consistency of measures over time. External reliability can be tested by the test-retest method. The same participants can be tested on more than one occassion to see whether their results remain similar. 

Inter-observer reliability: assess whether, in n observational study, if several observers are coding behaviour, their codings or ratings agree with each other. To improve reliability, all observers must have clear and operationalised categories of behaviour and must be trained in how to use the system. Inter-observer reliability can be measured using correlational analysis, in which  high positive correlation among ratings indicates that high inter-observer reliability has been established. 

25 of 37


Internal validity: refers to the extent to which a measurement technique measures what it is supposed to measure, whether the IV really caused the effect on the DV or whether some other factor was responsible. Experiments may lack internal validity because of demand characteristics or participant reactivity or extraneous variables have not been controlled. 

Mandane realism: Do the measures used generalise to real life? Munddane realism is an aspect of internal validity that contributes to external validity. 

External Validity: refers to the validity of  study outside the research situation and provides some idea of the extent to which findings can be generalised. Three factors should be considered: 

  • How represtative is the sample of the population to which the result is to be generalised. (population validity)
  • Do the research setting and situation generalise to a realistic real life setting or situation? (ecological validity)
  • Do the findings generalise to the past and to the future (ecological and historical validity
26 of 37

Random sampling

Involves having all the names of the target population and giving everyone an equal chance of being selected, this can be selected from a computer or names from a hat (for a small population)


  • Avoids bias as every member of the target population has an equal chance of being selected.


  •  Impossible to obtain a truly random sample because not all the names of the target population may be known. 
27 of 37

Opportunity sampling

Involves asking whoever is available nd willing to participate. An opportunity sample is not likely to be representative of any target population because it will probably comprise of friends and family of the researcher. The people approachd will be those who are local and available.


  • The researchers can quickly and inexpensively acquire a sample and face to face ethical briefings and debriefings can be undertaken.


  • Opportunity samples are almost always biused samples, as whose who participates is dependent on who is asked and who happens to be available at the time. 
28 of 37

Volunteer sampling

People who volunteer to participate. 


  • The participants should have their informed consent, will be interested in the research and may be less likely to withdraw. 


  • A volunteer sample may be a biased sample who are not representative of the trget population because volunteers may be different in some way from non-volunteers. 
29 of 37

Sample Representativeness

  • The sample of the participants should be a true representation of diversity in the target population. 
  • If it not representativeness then the findings cannot be generalised to the population. 
  • Therefore the researchers need to think about the unmber of people needed:
    • Is the sample large enough to be representative of the target population. 
    • The sample needs to be manageable size, as too many participants lead to expense and time consuming. 
    • If the target population is small is may be more sensible to use the whole population as the sample. 
30 of 37

Participant effects

The Hawthorne effect: if people are aware they are being watched they may try harder on tasks and their behaviour may change. This may mean that any findings are artifically high which may lead to an invalid conclusion.

Demand characteristics: The research or the researcher may give cues to participants as to what the experiemtn is about and what the researcher is looking for the participants then may change their actions and behaviour. This may lead to a bias response  in which participants try to please the experimenter this will lead to the findings being invalid. Demand characteristics may be reduced if a single-blind procedure is used this is were paticipants do not know which condition they are participating in, or are given a false account of the experiment .

Social desirability bias: people usually try to show themselves in the best possible way. So when answering questions in interviews or questionaires, they may give answers that are socially acceptable but not truthful.

31 of 37

Investigator and/or experimenter effects

Investigator expectancy: the expectations of the researcher can affect how they design their research and bias how and what they decide to measure, and how the findings are analysed.

Experimenter bias: the experimenter can affect the way participants behave. One way to reduce experimenter effects is to use a double-blind procedure in which neither the experimenter nor the participants know what the research hypothesis is.

Interviewer effects: the expectations of the interviewer may lead them to ask only those questions in which they are interested, or to ask leading questions, or they may only focus on answers that match their expectations.

Observer bias: when observing behaviour, observers may make biased interpretations of the meaning of behaviour.

32 of 37

Qualitative and Quantitative data

Strengths of quantitative data:

  • Precise measures are used.
  • Data is highly reliable.
  • It is possible to see patterns and trends in the data.

Weaknesses of quantitative data:

  • May lack or lose detail.
  • Often collected in contrived settings.

Strengths of Qualitative data:

  • Rich and detailed.
  • Can provide information on people's attitudes, opinions and beliefs.

Weaknesses of Qualitative data:

  • May be subjective.
  • Can be an precise measure.
33 of 37

Measures of central tendency

The mean: To calculate the mean, all the scores are added up and the total is idvided by the number of the scores.

STRENGTH: It takes all the values from the raw scores into account.

WEAKNESS: Often the mean has a meaningless decimal point that was not in the original data.

The Median: the median is the central number in a list of rank-ordered scores.

STRENGTH: It is useful when scores are ordered data.

WEAKNESS: Can be misleading if used in small sets of data.

The mode: is the score that occurs most frequently in a set of scores.

STRENGTH: The mode is not effected by extreme scores.

WEAKNESS: The mode tells us nothing about other scores.

34 of 37

Measures of dispersion

Measure of dispersion tells us about how far spread out they are. The main measures of dispersion are the range and the standard deviation.


To calculate the range of scores subtract the lowest score from the hghest score.

STRENGTH: Easy and quick to work out, Includes extreme values.

WEAKNESS: May be misleading when there are extremely high or low scores in a set.

Standard deviation:

Is used to measure how the scores are distributed around the central point. The higher the deviation point the larger the spread of the scores.

STRENGTH: It is a sensitive measure of dispersion because of all the scores are used in its calculation.

WEAKNESS: Standard deviation is not useful when data are not normally distributed.

35 of 37

Graphs and Charts

Bar charts: Bar charts are used when scores are in categories, when there is no fixed order for the items on the x-axis, or can be used to show a comparison of means for continuos data. The bars in the bar chart should all be the same width but should not touch.

Histograms: Histograms show frequencies of scores (how the scores are distribruted) using columns. They should be used to display frequency distributions of continous data and there should be no gaps between the bars.

Scattergrams: Scattergrams are used to depict the result of correlational analysis. A scattergram shows at a glance whether there appears to be a posisitive or negative correlation or no correlation.

Table: Psychologists put results in data makes it easier for others to see and interpret results at a glance.

36 of 37

Content analysis

Analysing and presenting qualitative data:

  • Important that researchers avoid subjective or biased misinterpretations. Misinterpretation can be avoided by:
    • Using accurate language to operationalise the variables to be measured.
    • Using a team of observers who have verified that they have achieved inter-observer reliability.
    • Converting qualitative data; one way to do this by coding the data.

Coding qualitative data: A sample of qualitative data is collected.

Coding units: are identified in order to categorise the data. A coding unit could be a specific word/s or phrase/s that are looked for. They then may be counted to see how frequently the occur. The result frequency of occurance is a form of quantitative data.

37 of 37


No comments have yet been made

Similar Psychology resources:

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