psychexchange.co.uk shared resource
TRANSCRIPT
UNIT FOURResearch methodsRevision booklet
Research Methods all you
need to know for the examThe application of scientific method in psychology The major features of science, including replicability, objectivity, theory construction,
hypothesis testing, the use of empirical methods
Validating new knowledge and the role of peer reviewDesigning psychological investigations
Aims Hypotheses, including directional and non-directional Operationalisation of variables, including independent and dependent variables Selection and application of appropriate research methods Design of naturalistic observations, including the development and use of behavioural
categories Design of questionnaires and interviews Studies using a correlational analysis Observational techniques Self-report techniques including questionnaire and interview Case studies Experimental method, including laboratory, field and natural experiments Experimental design (independent groups, repeated measures and matched pairs) Selection of participants and sampling techniques, including random, opportunity and
volunteer sampling and the implications of sampling strategies, for example, bias and generalising
Issues of reliability, including types of reliability, assessment of reliability, improving reliability
Assessing and improving validity, including internal and external Ethical considerations in design and conduct of psychological research, and ways in
which psychologists deal with them Control of extraneous variables Demand characteristics and investigator effects Pilot studies
Data analysis and reporting on investigations
Presentation and interpretation of quantitative data including graphs, scattergrams and tables, and Appropriate selection of graphical representations
Analysis and interpretation of quantitative data. Measures of central tendency, including median, mean, mode. Measures of dispersion including ranges and standard deviation
Analysis and interpretation of correlational data. Positive and negative correlations and the interpretation of correlation coefficients
Probability and significance, including the interpretation of significance and Type 1/Type 2 errors
Factors affecting choice of statistical test, including levels of measurement The use of inferential analysis, including Spearman’s Rho, Mann-Whitney, Wilcoxon, Chi-
Squared Processes involved in content analysis Presentation, analysis and interpretation of qualitative data Conventions of reporting on psychological investigations
The application of scientific methodology in Psychology
The Nature of Science:The scientific method requires theories to be objectively tested. The requirement for objectivity is based on the philosophical doctrine of empiricism. In order to be seen as objective, research must be rational and conducted and presented in such a way as to make it replicable.
Objectivity; taking an objective approach to an issue means having due regard for the known valid evidenceReplication; can it be repeated? Record methods carefully and use valid measures.Falsification; is the principle that a proposition or theory cannot be considered scientific if it does not admit the possibility of being shown false. Theory construction: one aim of science is to record facts, but an additional aim is to use these facts to construct theories to help us understand and predict the natural phenomena around us. A theory is a collection of general principles that explain observations and facts. Hypothesis testing: Karl Popper suggested that in order for a theory to be scientific it must be able to make predictions about behaviour and to do this it must be possible to show they are wrong (falsifiability).
Approach that support psychology as a science
Approach that does not support psychology as a science
Biological PsychodynamicBehavioural
Peer review essential (should control for….) Validity Correct analysis of data Accurate Quality
Peer review where psychologists read the article and judge whether the research is credible and valid. Critics suggest that impartial review is an unachievable ideal because research is conducted in a social work and social relationships affect objectivity.
Problems with peer review Fabrication (data is made up) Falsification (where data exists but has been altered) Plagiarism (where work has been copied from others)
Issues: Gender and institution bias, file drawer phenomenon, research could be rejected if not in keeping with current theory
Designing psychological investigations
An aim is a statement of the purpose of a study.Hypothesis is a statement that is testable. A research hypothesis is a general prediction made at the beginning of an investigation , about what the researcher expects to happen.
Types of hypotheses Experimental hypothesis/alternative hypothesis predicts that something other
than chance alone has played a part in producing the results obtained. Null hypothesis predicts that the results obtained from an investigation are due to
chance alone. A directional hypothesis predicts the direction in which results are expected to occur. E.g. of a directional hypothesis. More words are recalled from a list when using rehearsal as mnemonic technique than when no mnemonic technique is used.
A non-directional hypothesis does not predict the expected direction of outcome. E.g. of a non-directional hypothesis. There is a difference in the number of words recalled from word lists presented with or without the presence of background music.
Non-experimental methodsQuestionnaires: are used to obtain views of large numbers of people about specific issues or topics.
The questions on a questionnaire can be either ‘closed’ or ‘open’. Closed questions provide people with a fixed set of responses, whereas open questions allow people to express what they think in their own words.
The responses to closed questions are easy to summarise and quantify whereas the responses to open questions are much more difficult to analyse. But in some ways the rich detail provided by open questions can provide useful insights into the topics being researched.
One problem encountered when using questionnaires is ‘response set’. What can happen is that people stop reading the questions and just tick the same answer all the way down the page.
Advantages Disadvantages
Large amounts of data can be collected very easily and quickly
Rely on self-report data which may be biased
Questionnaires using closed questions are highly replicable.
Data may be biased by response rates
Open ended questions can provide rich detail
Social desirability effects may influence responses
Interviews: There are two types of interviews – structured and semi-structured.
In structured interviews each participant is asked the same questions in the same order. Many of the questions will have a fixed set of responses that the participant has to choose from.
Advantages Disadvantages
Data can be summarised and analysed quite quickly
Any interesting issues raised by participants cannot be followed up
It is possible to interview large numbers of people fairly efficiently
Semi-structured interviews - combination of fixed questions and open-ended questions where participants are given the opportunity to talk about particular topics.
Unstructured interview - no fixed questions at all.
Advantages Disadvantages
Generates huge amounts of data which enables interesting topics to be explored in great depth
Analysis can be very time consuming
Useful when dealing with complex issues such as bereavement, violence and parenting
Replicating the research can be difficult
Observational studies: covert observation and overt observation.
In participant observation the researcher becomes part of the group they are studying and takes part in the group’s activities as well as recording the behaviour of the group.
In non-participant observation the researcher is separate from those being observed and records the behaviour but does not take part in any of the activities.
Researcher must define the behaviour he/she observes.
Psychologists use behavioural categories to record particular instances of behaviour.
For example, observers may use check lists or tally chart for recording observations.
Sampling techniques used for observation can be time sampling and event sampling.
Advantages Disadvantages
Participants are behaving naturally in an everyday environment. This means that the studies have high ecological validity.
One of the main ethical issues is consent. Participants should normally be aware that they are taking part in a study and have the right to withdraw. However, there is the problem that if participants know they are being observed they may change their behaviour.
This means that the studies have high ecological validity.
Observational studies also tend to focus on ‘whole behaviours’ rather than breaking behaviour down into small units
Case studyThis is an in-depth study of one individual or a small group of people. Case studies are often carried out on those who are of special interest because they are outstanding or different in some way.
Advantages Disadvantages
The case can be viewed in depth and long sequences of behaviour can be followed
Results cannot be generalised as they may not be representative of the general population.
It is not artificial as it relates to an individual or small group’s experiences.
Sometimes, case studies are carried out retrospectively and the evidence may become distorted in people’s memories.
Rich qualitative data can be collected Replication is difficult
Experimental method Manipulates an independent variable (IV) in order to investigate the change in
dependent variable. All other variables which might influence the results (called extraneous
variables) are controlled. Participants are randomly allocated to the experimental and control conditions.
IV: Variable that is deliberately manipulated by the experimenter and is assumed to have a direct effect on the dependent variable.
DV: Variable that is measured
Laboratory experiments
In the experimental method, participants are put under different conditions and then their performance is measured. The laboratory is a convenient setting for the experimental method because the experimenter can then control what happens.
Advantages Disadvantages
The experimenter can control the experiences of the participants. It is possible to isolate cause and effect relationships more precisely.
The laboratory is an artificial situation. Participants may act differently from the way they would in real life.
Equipment can be ready and available for measuring the participants’ behaviour.
Tasks may have little relevance to real life.
Field experimentExperimental investigations carried out in natural settings (e.g. homes, school). Involve direct control of the IV and allocation of participants to groups.
Advantages Disadvantages
Reduced demand characteristics Degree of control over IV is less than lab
Higher in ecological validity than lab experiments though still has control over IV
More difficult to control extraneous variables
Natural experiments: Take advantage of a naturally-occurring change in behaviour. By measuring before and after the change, the experimenter hopes to find out what affect the change had. An example of a natural experiment might be a study that looks at the effects on aggressive behaviour of the introduction of TV for the first time into a previously isolated community. Do not confuse natural experiments with naturalistic observation.
Advantages Disadvantages
Behaviour is observed in natural surroundings.
There is no control over other changes that might happen.
Samples are usually large and representative.
Measure may be not as accurate as direct observation because of the presence of the experimenter.
Experimental designsRepeated measures – the same PP’s are tested in all conditions, this means there is no individual differences and uses less PP’s which means consistent results. However there can be order effects as PP’s know what is coming. – control used is counterbalancingIndependent groups design – different PP’s are used in each of the conditions, the PP’s are usually randomly allocated to each condition to try and balance any differences. There are no order effects in this instance, there are no demand characteristics but there are group differences.
Matched pairs design – different but similar PP’s are used in each of these conditions, the PP’s are matched to another PP in other groups, identical twins are the perfect matched pair. There are less group differences than in independent design, but matching is difficult and time consuming.
Sampling techniques Random sampling, Every member of the target population has an equal chance
of being selected, e.g. selected by computer from the electoral roll. Unbiased, though may not be representative / problems with generalisation
Opportunity sampling, Use whoever is immediately available e.g. go out and accost passers-by. Biased sample / not be representative / problems with generalisation
Volunteer sampling, use whoever puts themselves forward e.g. advertise in the paper. Biased sample / not representative. / problems with generalisation
ReliabilityInternal Reliability: The consistency of a measure within a test (i.e. all items need to be measuring the same thing).External Reliability: The ability to replicate the results of a study and get the same or highly similar results.In the context of an experimental research it refers to consistency of results.
In observation method, the extent of agreement between observers is called inter-rater or inter-observer reliability. Reliability can be improved through training.
In self report methods, the two types of reliabilities are
Inter-interviewer reliability- is the consistency of the outcome of interviews by different interviewers.
Two assessment methods of reliability: Split-half method and test-retest method.
Validity –whether the results obtained in a study are true/genuine. Internal validity: Whether the study did test, what it intended to measure. External validity: the extent to which the results of the study can be generalised to other situations ( ecological validity) and people.
Laboratory experiments are not necessarily low in external validity. If low in mundane realism, reduces generalisability of the findings.
In observation method, internal validity is affected by observer bias. In self-report techniques, two ways to assess internal validity are
Face validity: the extent to which the test looks as if it is measuring what it should measure. Concurrent validity: assessed by comparing the outcome of a test with an already established test on the same topic.
Assessing and Improving Reliability and Validity:Assessing Reliability: Assessing Validity:
Internal: Assessed using the split-half method. This is when one half of the test is compared with the other in order to check whether the scores are consistent.
Internal: Running a pilot study allows researcher to identify whether there are any issues with the study (e.g. confusing items on the questionnaire). The researcher would also operationalise all variables.
External: Assessed using the test-retest method. The test is carried out on the same of similar PPs several times and the similarity of results is recorded. Alternatively, a correlation coefficient can be worked out between two sets of scores. The higher the correlation coefficient, the higher the reliability!
External: The researcher could test the findings from one study in a number of different settings and with different PPs. This would allow the researcher to see whether or not their results can be applied to the entire population or to what extent they are limited.
Ethical considerations
Humans (BPS guidelines) Informed Consent, Debriefing, Confidentiality, Right to Withdraw,
Protection from harm Care to be taken with socially sensitive issues ie. Race, social tension,
homophobia etc. Animals
Moral justification Speciesm (just because we can talk does that mean we can control
everything?) Difficult to judge pain and emotion Can give greater control and objectivity Animal research is controlled and can be ethical
Variables can be controlled by • Counterbalancing• Random allocation
Extraneous variables can be controlled by keeping them constant or eliminating them altogether. All the variables other than the independent variable that might affect the dependent variable are called extraneous variables. If a variable other than the IV, produces a change in a DV, the results of the study are said to have been confounded.
Demand Characteristics: The features of an experiment that inform participants about the aim and influence their behaviour. Could be a confounding variable
Investigator effects:- anything the investigatior/experimenter does which has an effect on the participant’s performance or behaviour. Could be a confounding variable
Pilot Study – is a preliminary study or trial often carried out to predict snags and find out about the main features of a main research study that will follow it. A trial-run basically. Piloting is trying out the scale or research design on a small sample.
Data analysis and reporting on investigations
Analysis and interpretation of quantitative data.
Bar chart: used when data is in discrete categories. The bars should not touch. Can be used for mean scores.
Histogram: used for continuous data from one data set. Used for ordinal, interval ratio data
Scattergram: Used for correlation to show relationship between two variables.
Frequency polygon: Used for continuous data, to compare 2 data sets. Used for ordinal, interval ratio data
Tables of resultsTable to show the mean and SD for words and pictures
Words PicturesMean 6 9SD 2.3 1.9
Contingency table to show the differences between males and females in physical and verbal aggressive acts observed in TV soaps
Males FemalesPhysical 35 7Verbal 12 26
Measures of central tendency:Mode: A measure of central tendency that represents the most occurring value in a set of scores. Used for nominal data
Median: The measure of central tendency that represents the value of the middle score. Is not effected by extreme scores, however does not take into consideration all values. Used for ordinal data.
Mean: A measure of central tendency where you add up all of the scores and divide by the number of scores you have. Takes into consideration all values but is effected by extreme scores. Used for interval data
Measures of dispersion.Range: The spread of scores from the highest score to the lowest. Used with the median
Standard Deviation: The measure of dispersion that shows how far away the raw score is from the mean of it’s distribution. Used with the mean
Analysis and interpretation of correlational data.
For a correlational study, the data can be plotted as points on a scattergraph.
A line of best fit is then drawn through the points to show the trend of the data.
If both variables increase together, this is a positive correlation. When the correlation coefficient is calculated it is between 0 - +1, the nearer the value is to +1, the stronger the relationship.
If one variable increases as other decreases this is a negative correlation. When the correlation coefficient is calculated it is between 0 - -1, the nearer the value is to -1, the stronger the relationship.
If no line of best fit can be drawn, there is no correlation.
Probability and significance
Probability: The likelihood of an event occurring 100% would be very likley
Level of significance: The 5% probability level set for the results being due to chance factors, this is written as P=0.05
Type I error: This is when the null hypothesis is rejected when in fact it is true. Can occur when an insufficiently stringent significance level is adopted.
Type II error: the null hypothesis is accepted when in fact it is false. This error can occur when significance levels are too strict.
Levels of measurementNominal data: If you are placing data in a tally chart, this would be nominal data. If you are counting the no of times something has happened, this is nominal data e.g. no of males/females stopping for a pedestrian at a zebra crossingOrdinal data: Use of self report questionnaire with a rating scale would be ordinal data. The likert scale is the most commonly used scale, where Ps are given a statement and asked to state level of agreement, form ‘agree very much’ to ‘disagree very much’ . Interval data - This is where data is in the form of equal units e.g. score on an objective test. In such a case, if one P has scored 20 on a test and another has only scored 10, we would be correct in saying the first person has performed twice as well as the second. Ratio Data - Ratio data is interval data with an absolute value of zero, e.g. distance or speed. Inferential analysis, including Spearman’s Rho, Mann-Whitney, Wilcoxon, Chi-Squared
Chi –Square – Is used when looking for association or differences, where data is nominal with independent measures design. (value is X2)
Mann-Whitney – Used when looking for differences, where data is at least ordinal level and for independent measures design (value is U)
Wilcoxon T Test – Used when looking for differences, with data at least ordinal level and repeated measures/matched pairs design (value is T)
Spearman’s Rho - Used when Testing for a correlation, the data is at least ordinal level (value is r)
For chi-squaRe and speaRman’s in order for the results to be significant the observed value should be gReateR than or equal to the critical value at p=0.05
For Mann Whitney and Wilcoxon in order for the results to be significant the observed value should be less than or equal to the critical value at p=0.05
Mnemonic:- gReateR than for tests that include an r, no r then less than
Presentation, analysis and interpretation of qualitative dataQualitative data is non-numerical such as books, films, pictures.
Content Analysis - Search and analysis of qualitative materials (especially texts in books, textbooks, magazines but can also be images on TV adverts etc) E.g. reading a senior school textbook and analysing how sexist or stereotyped the text/or illustrations are. The content analyser is looking at how to code the words/themes/phrases into units. This can be then turned into quantitative data, e.g. there were four illustrations of a girl doing stereotypical girl things: cooking, sewing, looking after a baby. It can remain purely qualitative though. The method is quite subjective/ prone to bias and time consuming.
Top down approach - Codes which are used are taken from an existing theory. The codes are therefore decided before the observation/interview takes place.
Bottom up approach - The codes are not decided before the observation/interview, but after. This is often used when investigating less researched areas.Conventions of reporting on psychological investigations
Abstract:-Brief summary of the report covering the aims/hypothesis, method/procedure, results and conclusions. Good to look at to see if a report is worth reading.
Introduction:- Introduce the reader to the topic area and background to the study. Leading to the aim and research predictions (hypothesises).
Method:-This section it will describe how the study was conducted including the design, participants, apparatus/resources/materials, pilot study and procedure. It should have enough information to allow for replication.
Results:- Descriptive – allows for visual representation of differences between the groups and inferential – analysis of the data and null hypothesis to be accepted or rejected- statistics.
Discussion:- Summary of the findings of the results, explanations for anomalies, evaluation of the method and the implications of the research and make suggestions of further research.
References/Bibliography:- All research documents, journals, internet resources and books that was mentioned and used for additional background research.
Appendices:- Copies of all resources/materials used, raw data and statistical calculations.
Standardised Procedures: To ensure that all participants have the same experience, researchers should ensure that they are all tested:
Standardised instructions: these are given in the chosen location and there may be more than one set for different (IV) conditions of the investigation. The standardised instructions are important, as each participant must receive the same instructions so they understand what to do in the study. Also, it ensure that each participant is treated the same and should help eliminate experimenter effects. You can read the instructions to the participant(s), or let them read the instructions themselves.
Test your knowledge
For each of the methods of investigation explain each and provide two advantages and disadvantages Laboratory Experiments, Field Experiments and natural Experiments
Identify and explain the six of the ethics that all Psychologists need to abide by:
What does the acronym DIP stand for and how can you use it in your exam answers?
Identify the three types of graphs that you can use and explain under what conditions you would use them?
Explain how to calculate the following and describe their purpose within investigations Mean, Mode, Median and Range
What does a high standard deviation show?
What does a low standard deviation show?
What is the difference between these two Quantitative and Qualitative Data?
What is content analysis?
What is a hypothesis?
What are the two types of hypothesis and what is the difference between them?
What is the difference between a directional and a non-directional hypothesis?
What is an independent variable?
What is a dependant variable?
What is an extraneous variable?
Explain the four types of extraneous variable
What does it mean to operationalise a variable?
Define reliability
Define validity
What is the difference between internal and external validity?
Why would an experimenter use standardised proceudres?
How can validity be measured?
What is meant by the Target Population?
Explain Random Sampling
Explain Opportunity Sampling
Explain Volunteer Sampling
What is the advantage and disadvantage of Repeated Measures design
What is the advantage and disadvantage of Independent Measures design
What is the advantage and disadvantage of match Pairs Design
Why would an experimenter use standardised instructions?