Research methods

Experimental method

Aim: a general statement of what the researcher intends to investigate (the purpose of the study)

Hypothesis: a clear, precise, testable statement which states the relationship between the variables to be investigated.

  • Directional: states the direction (e.g. increase, decrease etc.)
  • Non-directional: unspecific (e.g. there will be a difference...)

Independent variable (IV): manipulated by the researcher so the effect on the DV can be measured.

Dependent variable (DV): measured by the researcher, any change should be caused by the IV.

Operationalisation: clearly defining variables in terms of how they can be measured.

1 of 37

Control of variables

Extraneous variable (EV): a nuisance variable that does not vary systematically with the IV but may have an effect on the DV.

Confounding variable: a variable that varies systematically with the IV and may have an effect on the DV but we cannot be sure of the true source of the change.

Demand characteristics: when the participants think that they know the purpose of the investigation and therefore change their behaviour within the research situation.

Investigator effects: any effect of the researcher's behaviour on the research outcome. This may include everything from the design of the study to the selection of and interaction with the participants.

Randomisation: the use of chance in order to control for the effects of bias when designing materials and deciding the order of conditions.

Standardisation: using the same formalised procedures and instructions for all participants.

2 of 37

Experimental design

Independent groups design: two separate groups experience two different conditions and the results of the groups are compared.

  • participant variables (-) researchers use random allocation to deal with this (+)
  • less economical than other designs because more participants are needed (-)
  • order effects aren't a problem and participants are less likely to guess the aim (+)

Repeated measures design: all participants experience all the conditions in the experiment.

  • order effects arise because of boredom or fatigue or the participants may improve at a skill-based task with each trial (-) researchers use counterbalancing to deal with this (+)
  • demand characterisitics (-)
  • participant variables are controlled and fewer participants are needed (+)

Matched pairs design: pairs of participants are first matched on some variable that may affect the DV, then one member of the pair is assigned to condition A and the other to condition B.

  • order effects and demand characteristics are less of a problem (+) participants can never be matched on every variable, time consuming and expensive (-)
3 of 37

Laboratory experiments

Conducted in highly controlled environments within which the researcher manipulates the IV and records the effect on the DV.


  • High control over extraneous variables = high internal validity.
  • Replication is possible because of the high level of control = increases validity.


  • Lacks generalisability to everyday life = low external validity.
  • Demand characteristics
  • Low mundane realism.
4 of 37

Field experiments

An experiment that takes place in a natural setting within which the researcher manipulates the IV and records the effect on the DV.


  • Higher mundane realism becasue the environment is more natural.
  • Produces behaviour that is more valid and authentic = high external validity.


  • Lack of control over extraneous variables = cause and effect between the IV and the DV in field stuides may be much more difficult to establish.
  • Replication is not alwasy possible.
  • Ethical issues = lack of consent and invasion of privacy.
5 of 37

Natural experiments

The change in the IV is not brought about by the researcher but would have happened even if the researcher had not been there. The researcher records the effects on the DV.


  • Provides opportunies for research that may not otherwise be undertaken for practical or ethical issues.
  • High external validity.


  • May only happen very rarely, cannot always generalise as they are usually extreme events.
  • Participants may not be randomly allocated to experimental conditions = particpant variables, bias.
6 of 37

Quasi experiments

The IV has not been determined by anyone - the variables simply exist, such as being old or young.


  • High control over extraneous variables = high internal validity.
  • Replication is possible = validity.


  • Cannot randomly allocate participants to conditions and therefore there may be confounding variables.
7 of 37

Random sample

All members of the target population have an equal chance of being picked. 

A complete list of all the members of the target population is obtained and each person is assigned a number. The sample is them generated through a lottery technique.


  • No researcher bias = cannot choose people they think will support their hypothesis.
  • More likely to produce a representative sample.


  • Time consuming.
  • May end up with a sample that is still unrepresentative.
  • Participants may refuse to take part.
8 of 37

Systematic sample

When every nth member of a target population is selected.

A sampling frame is produced which is a list of the target population in, for example, alphabetical order. A sampling system is nominated or produced randomly.


  • Avoids researcher bias.
  • Representative.


  • Time-consuming
9 of 37

Stratified sample

The composition of the sample reflects the proportion of people in certain strata within the target population.

The researcher first identifies the strata that make up the population and works out the representative proportions. Finally the participants are selected using random sampling.


  • Avoids researcher bias.
  • Representative sample = generalisation of findings is possible.


  • The strata cannot represent every participant variable so there's always going to be confounding variables.
10 of 37

Opportunity sample

Researchers decide to slect anyone that is willing and avaliable. They ask whoever is around at the time of their study.


  • Saves time and money.


  • Unrepresentative of the target population = findings cannot be generalised.
  • Researcher bias.
11 of 37

Volunteer sample

Participants select themseves to be part of a sample. Advertised through a newspaper etc.


  • Less likely to withdraw
  • Less time-consuming and requires minimal input from the researcher.
  • Participants are more likely to be helpful and willing.


  • Volunteer bias = attracts a certain type of person who is helpful and curious whihc affects how far the findings can be generalised.
12 of 37

Ethical issues

When a conflict exists between the rights of the participants and the goals of the research to produce authentic and valid data.

Informed consent: Making participants aware of the aims, the procedures, their rights (including the right to withdraw from the study at any time) and what their data will be used for.

Deception: Deliberately misleading or withholding information from participants at any stage of the investigation. Participants who have not recieved adequate information at the start cannot be said to have given fully informed consent. However, occasions when it can be considered justified.

Protection from harm: Participants should not be placed at any more risk than they would be in their daily lives. They should be protected from both physical and psychological harm, including being made to feel embarrassed, inadequate or being placed under unndue stress or pressure. Participants must be reminded of the fact that they have the right to withdraw.

Privacy and Confidentiality: Participants have the right to privacy and if this is invaded then confidentiality should be protected. This refers to people's right to have any personal data protected and is law under the Data Protection Act.

13 of 37

Ways of dealing with ethical issues

BPS code of conduct: British Psychological Society. Instructs psychologists in the UK about what behaviour is acceptable and not acceptable when dealing with participants. Built around 4 major principles: respect, competence, responsibility and integrity. The ethics committees use a cost-benefit approach to determine whether research proposals are ethically acceptable.

Informed consent: Participants should first be issued a consent form detailing all the relevant information that may affect their decision to participate, this is then signed if the participant agrees. If they are under 16, parental consent is required.

Presumptive consent: A similar group of people are asked if the study is acceptable.  

Prior general consent: Participants give their consent to take part in a number of different studies.

Retrospective consent: Participants are asked for their consent having already taken part in the study.

14 of 37

Ways of dealing with ethical issues continued.

Deception and protection from harm: At the end of the study the participants should be given a full debrief within which they should be made aware of the true aims of the investigation and details they were not supplied with during the study.

They must also be told what their data will be used for and offered the right to withdraw.

Participants must be assured that their behaviour was normal and the researcher must provide counselling in extreme cases if it is needed.

Confidentiality: If personal details are held these must remian protected however, it is more simple to maintain anonymity and refer to the participants as numbers or their initials etc.

In case studies, psychologists often use initials when describing the person involved (e.g. HM).

Participants must be reassured in the debrief form that their data will be protected.

15 of 37

Pilot studies and other procedures

Pilot Study: A small-scale version of an investigation that takes place before the real investigation is coonducted. The aim is to check that procedures, materials and measuring scales etc. work. It also allows the researcher to make changes if necessary.

Single-blind procedure: When the participant is not told the aim of the research or other details at the beginning of the study. This is an attempt to control for demand characteristics.

Double-blind procedures: Neither the researcher nor the participant is aware of the aims of the study (often a third party conducts it without knowing the purpose). This is often an important feature of drug trials as treatment may be administered to patients by someone who doesnt know the difference between the real one and the placebo.

Control groups and conditions: In a drug trial, the group that receives the drug is the experimental condition and the group that receives the placebo is the control group.

Control groups are used for the purpose of comparison, there is a significant difference between the two groups than the researcher can conclude it was due to the IV (assuming all of the confounding variables have remained constant).

16 of 37

Naturalistic and controlled observations

Naturalistic: Watching and recording behaviour in the setting it would normally occur. All aspects of the experiment are free to vary.


  • high external validity = findings can be generalised to everyday life.
  • Behaviour is more authenic and valid.


  • Replication is difficult = uncontrolled extraneous variables.

Controlled: Watching and recording behaviour within a structured environment. Control over extraneous variables.


  • Extraneous variables are less of a factor so replication is easier.


  • May produce findings that cannot be readily applied to everyday life.
17 of 37

Covert and overt observations

Covert: Participants' behaviour is watched and recorded without their knowledge or consent. Such behaviour must be public and happening anyway for it to be ethical.


  • No participant reactivity = behaviour is natural and data is more valid.


  • Ethical issues = lack of privacy and confidentiality, decpetion and lack of informed consent.

Overt: Participants' behaviour is watched and recorded with their knowledge and consent.


  • More ethically acceptable as informed consent is given.


  • Hawthorne effect = participants know they are being observed and change their behaviour.
18 of 37

Participant and non-participant observations

Participant: The researcher becomes a member of the group whose behaviour they are watching and recording.


  • Increased insight into the lives of the participants = increased validity.


  • Researcher may come to identify with the group and lose objectivity.

Non-Participant: The researcher remains outside of the group whose behaviour they are watching and recording.


  • Researcher maintains an objective psychological distance.


  • May lose valuable insight into the participants' behaviour as they are too far removed.
19 of 37

Structured and unstructured observations

Structured: Allows the researcher to quantify their observations using a pre-determined list of behaviours and sampling methods.

Unstructured: The researcher writes down everything they see and produces accounts that are rich in detail.


  • Structured: Recording data is easier and more systematic. Data produced is likely to be quantitative which means that analysing it is easier.
  • Unstructured: Benefit from more richness and depth of detail in the data collected.


  • Structured: Data isn't as in depth and detailed as unstructured observations and the researcher can only observe the behaviour categories.
  • Unstructured: Qualitative data is much more difficult to record and analyse. Greater risk of observer bias as behavioural categories aren't used.
20 of 37

Behavioural categories and sampling methods

Behavioural categories: When a target behaviour is broken up into components that are observable and measurable. Before the observation begins the researcher should ensure that they have included all of the ways in which the target behaviour may occur within their behavioural checklist.

  • Behaviours must be observable, measurable and self-evident.
  • Makes data colection more structured and objective.
  • Categories should be exclusive and not overlap.

Sampling methods: continuous recording of behaviour is a key feature however, this might not be practical or feasible.  

Event sampling: counting the number of times a particular behaviour is occurs in a target individual or group. Useful when behaviours happen infrequently and could be missed. However, if the event is complex they may overlook certain behaviours.

Time sampling: recording behaviour within a pre-established time frame. Effective in reducing the number of observations needed however, it may be unrepresentative as behaviours may be missed.

21 of 37

Inter-observer reliability

Single observers may miss important details or may only notice events that confirm their opinions or hypothesis. This introduces bias.

To make data recording more objective and unbiased, observations should be carried out by at least 2 researchers. It is vital that they are consistent in their judgements.

  • Observers should familiarise themselves with the behavioural categories to be used.
  • They then observe the same behaviour at the same time, perhaps using a pilot study.
  • Observers should compare the data they have recorded and discuss any differences in interpretation.
  • Observers should analyse the data from the study.
  • Inter-observer reliability is calculated by correlating each pair of observations made and an overall figure is produced.
22 of 37


A pre-set list of written questions that the participant responds to. Can be used as part of an experiment to assess the DV.

Open questions: No fixed range of answers and respondents are free to answer in any way they wish. Produce qualitative data = increased depth and detail but difficult to analyse.

Closed questions: A fixed number of responses. Produces quantitative data = easy to analyse but lacks detail. 


  • Cost-effective and can gather large amounts of data quickly.
  • Can be completed without the researcher being present.
  • Data is usually easy to anaylse and comparisons can be made.


  • Responses may not be truthful = social-desirablility bias (demand characteristics).
  • Response bias = Respondents tend to reply in the same way. (Acquiescence bias)
23 of 37


Structured: A pre-determined set of questions that are asked in a fixed order. Straightforward to replicate and reduces differences between interviewers. However, it isn't possible for interviewers to deviate from the questions or expand on any.

Unstructured: No set questions, the general aim is that a certain topic will be discussed, interaction tends to be free-flowing. The interviewee is encouraged to expand and elaborate their answers. Much more flexibility, the interviewer can follow up points that arise and are more likely to gain insight into the world of the interviewee. However, analysis of data is not straightforward and there is a risk that social-desirability bias may be an issue.

Semi-structured: There is a list of pre-set questions but the interviewer is also free to ask follow-up questions when they feel it is appropriate.

24 of 37

Self-report design


Likert scales: The respondent indicates their agreement with a statement using a scale of usually 5 points.

Rating scales: Respondents identify a value that represents the strength of their feeling about a particular topic.

Fixed choice option: Includes a list of possible options and respondents are required to indicate those that apply to them.


  • Standardised interiew schedule that reduces interviewer bias.
  • Conducted in a quiet room without others to increase the chance of the interviewee being honest.
  • Interviewees should be reminded on several occasions that their answers will be treated in confidence.
25 of 37

Writing good questions

Errors that should be avoided:

Overuse of jargon: Questions should use technical terms that participants may not understand as they may not answer them if they're difficult to interpret.

Emotive language and leading questions: Ensure that the author's attitude towards a particular subject doesn't come across in the question and particpants aren't being steered towards a cerain answer.

Double-barrelled questions and double negatives: Don't have a question that contains two questions in one as participants may become confused or only answer one of them etc. Questions that contain double negatives can often be difficult to decipher.

26 of 37


Correlations: Illustrates the strength and direction of an association between 2 or more co-variables. Plotted on a scattergram. 

Co-variables: The variables investigated within a correlation. They are not referred to as IV and DV because a correlation investigates the association between them, not a cause and effect relationship. 

Positive correlation: As one co-variable increases, so does the other. 

Negative correlation: As one co-variable increases, the other decreases. 

Zero correlation: No relationship between the co-variables. 

Difference between correlations and experiments: In a correlation there are no variables being manipulated and a cause and effect cannot be established. There are intervening variables that mean that correlations cannot be used to infer a reason for something happening. 

27 of 37

Correlations - Strengths and limitations


  • Useful preliminary tool for research - provide a precise and quantifiable measure of how 2 variables are related. May suggest ideas for future research and are often used as a starting point. 
  • Quick and economical to carry out - no need for a controlled envrionment and no manipulation of variables is required. Also, data that has been collected by others can be used. 


  • Doesn't provide an explanation for why variables are related, can't tell which variables are causing the others to change. 
  • Third-variable problem - there may be an intervening variable that isnt tested. 
  • Can be occasionally misused or misinterpreted. 
28 of 37

Qualitative and quantitative data

Qualitative: Words and non-numerical forms. Offers more detail and gives the respondent more licence to develop their thoughts, feelings and opinions = greater external validity. However, often difficult to analyse which means that it can rely on the subjective opionions of the researcher which may be biased

Quantitative: Data that can be counted, usually as numbers. Simple to analyse which means comparisons can be easily drawn. Data also tends to more objective and less biased. However, it is much narrower in scope and meaning which means that it has lower external validity and may not be able to be generalised to everyday life. 

29 of 37

Primary and Secondary data

Primary: Information that has been obtained first hand by the researcher for the purpose of the research project. This means that the data is specific to the aim of the research. However, it requires time and effort to produce the data. 

Secondary: Information that has already been collected by someone else and so pre-dates the current research project. This means that it is inexpensive and easily accessible. However, the quality and accuracy of the data may not be valuable to the experiment or match the researcher's needs. It may also be out-dated or incomplete and the researcher cannot be sure that it was carried out following the correct procedures. 

Meta-analysis: 'Research about research'. The process of combining results from a number of studies on a particular topic to provide an overall view. This may involve a qualitative review of conclusions and/or quantatitive analysis of results. 

30 of 37

Measures of central tendency

Mean: The arithmetic average calculated by adding up all the values in a set of data and dividing by the number of values there are.

It is more representative of the data as it takes all the values into account however, it is easily distorted by extreme values.

Median: The central value in a set of data when values are arranged from lowest to highest. 

It is not affected by extreme scores and is easy to calculate however, it may be unrepresentative. 

Mode: The most frequently occuring value in a set of data.

It is very easy to calculate however, it is a very crude measure and is not representative of the data as a whole. 

31 of 37

Measures of dispersion

Range: A simple calculation of the dispersion in a set of scores which is worked out by subracting the lowest score from the highest score and adding 1 to account for any margin of error.

It is easy to calculate however, it only takes to two most extreme values into account which may be unrepresentative of the data set.

Standard deviation: Shows how much the scores deviate from the mean by calculating the difference between the mean and each score. All the differences are added up and divided by the number of scores which gives the variance. The standard deviation is the square root of the variance. The larger the standard deviation, teh grater the dispersion of the data.

It is more complicated and time consuming to carry out however, it provides a more sophisticated and precise measure of dispersion. 

32 of 37


Scattergram: A type of grapgh that represents the strength and direction of a relationship between co-variables in a correlational analysis. Do not depict cause-effect relationships, merely associations.

Bar chart: The frequency of each variable is represented by the height of the bars. Used when there is discrete data (divided into categories). 

Histogram: The bars touch each other which shows continous data. The y-axis represents the frequency. 

Line graphs: Represent continuous data and use points connected to lines to show how something changes in value. IV is plotted on the x-axis and the DV on the y-axis. 

33 of 37


Normal: A symetrical spread of frequency data that forms a bell-shaped pattern. The mean, median and mode are all located at the highest peak. 

Skewed: A spread of frequency data that is not symetrical, where the data clusters to one end.

Positive skew: When the long tail is on the positive (right) side of the peak and most of the distribution is concentrated on the left. 

Negative skew: When the long tail is on the negative (left) side of the peak and most of the distribution is concentrated on the right. 

34 of 37

Statistical testing

The Sign test: Determines whether the difference between the 2 variables is significant. To use it: 

  • The researcher needs to be looking for a difference rather than an association.
  • They need to have used a repeated measures design. 
  • They need data that is organised into categories (nominal data). 


  • Convert the data into nominal categories by subracting the scores for one variable from the other. If the answer is postive then a plus sign (+) is recorded and if it's negative then a minus sign (-) is recorded. 
  • The + and - categories are added up. 
  • The less frequent sign is labelled 'S', this is the calculated value. 
  • The calculated value is compared to the critical value. A directional hypothesis is one-tailed and a non-directional hypothesis is two-tailed. The significance needs to be 0.05. 
  • 'N' on the table of critcal values is the number of participants that took part. 
  • The calculated value of S must be equal to or less than the critical value of 0.05. 
35 of 37

Peer review

Peer review: The assessment of scientific work by others who are specialists in the same field to ensure that any research intended for publication is of high quality. The main aims are: 

  • To allocate research funding.
  • To evaluate the quality and relevance of research. 
  • To suggest amendments or improvements. 
  • It establishes the validity and accuracy of research. 

Anonymity: A minority of reviewers use their anonymity to criticise rival researchers as many researchers are in direct competition for research funding. This means that many people prefer open reviewing where the names of the reviewers are made public.

Publication bias: Journalists want to publish significant findings to increase the credibility and circulation of the publication. They also prefer to publish positive results. This creates a false impression of psychology if journalists are being selective about what they publish. 

Burying ground-breaking research: Reviewers tend to be critical of research that contradicts their own and their beliefs. Established scientsts are likely to be chosen as reviewers which means that results that agree with the public opinion are more likely to be published than that that is new and innovative. 

36 of 37

Implications of research on the economy

Attachment research into the role of the father: Psychological research has shown that both parents are equally capable of providing the emotional support necessary for healthy psychological development. This may promote more flexible working arrangements within the family, for example, it is now the norm within many households that the mother is the higher earner and works longer hours and many couples share childcare responsibilities. This meabns that modern parents are better equipped to maximise their income and contribute more effectively to the economy. 

The development of treatments for mental illness: Absence from work costs the government approx. £15 billion a year and a third of all absences are cause by mental health issues. Due to research, sufferers are able to be diagnosed quickly and manage their illness more effectively and return to work. Thus the economic benefit of psychological research is considerable. 

37 of 37


No comments have yet been made

Similar Psychology resources:

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