Research Methods

Science

Objectivity – scientists strive to be objective in their observations and measurements so their expectations should NOT affect they way they record findings (this would lead to SUBJECTIVE reporting).

Theory construction – one aim of science is to record facts, but an additional aim is to use these facts to construct theories, which help us understand and predict the natural phenomenon around us. A theory is a collection of general principles that explain observations and facts.

Hypothesis testing- forming a null and an alternative [experimental] hypothesis at the outset, which is fully operationalisable – In other words it contains reference to the IV and the DV [if it is an experiment]; OR both key variables [if it is a correlation study]

Empiricism – information is gained through direct observation or experiments rather than by a reasoned argument or unfounded beliefs (i.e. a “hunch”)

Replicability – one way to demonstrate the validity of any observation is to repeat it. If the outcome is the same, this confirms the truth of the original results, especially if two or more people discover the same end results. To allow for such precise replication, researchers must record their method carefully, so that the same procedures can be followed in the future.

Peer Review

-The assessment of scientific work by others who are experts in the same field -The intention of is to ensure that any research conducted and published is of high quality - generally unpaid - usually several reviewers for each article/assessment - Their task is to report on the quality of the research - their views are considered by a peer review panel - If peers agree, then an article may be published. Articles CANNOT be published prior to peer review.

- Allocation of research funding – research is paid for by government and charitable bodies, and we have a duty to spend their money responsibly - public bodies require reviewers to enable them to decide which research is likely to be worthwhile.

- Publication of research in scientific journals and books – scientific journal provide scientist with an opportunity to share the results of their research with others – work is peer reviewed as a mean of preventing incorrect or faulty data entering the public domain.

Unachievable ideal – sometimes you can’t find another ‘expert’ to review your data

Publication bias – peer review tends to favour the publication of positive results - editors want research that has important implications in order to increase the standing of their journal

Conventions for reporting psychological investigations

abstract A summary of the study covering the aims, hypotheses, methods, results

and conclusions

Introduction /aim

What the researcher intends to investigate. This often includes a review of previous research (theories and or studies) explaining why the researcher intends to conduct this particular study. The researchers may state his/her research predictions and a hypothesis/ hypotheses

method A detailed description of what the researcher did, providing enough information for replication of the study. Included in this section is information about pps (the sample), the testing environment, the procedures used to collect data, and any instructions given to pps before (the brief) and afterwards (the debrief)

results This section contains what the researcher found, often called statistical data, which includes descriptive statistics, (tables, averages, graphs) and inferential statistics (the use of statistical tests to determine how statistically significant the results are)

Raw data (individual pp scores ) DO NOT appear here – they go in the appendix section

discussion The researcher offers explanations of the behaviours observed in the study, and might also consider the implications of the results and make suggestions for future research

references The full details of any journal articles or books that are mentioned in the report

appendices Apparatus, scripts used, raw data, consent forms, pp responses, calculations for mean, standard deviation, statistical test, etc

Experimental methods

lab field natural

Description -An experiment conducted in a

-Investigating casual relationships in more

-IV is no directly manipulated or

controlled environment natural surroundings controlled by the experimenter

Strengths -high in terms of internal validity because many EVs can be controlled.

-Control also increases Replicability, which is desirable

-It may be possible to control EVs

-Experimenter effects reduced as pps are often not aware of being in a study

-Can draw casual conclusions

-Higher ecological validity

-Avoid some participant effects

-Allows research where IV cant be manipulated

-Enables psychologists to study ‘real’ problems

weaknesses -Some EVs, such as experimenter effects and demand characteristics may reduce internal validity.

-Control reduces external validity as a highly controlled situation may be less like everyday life

-Less control

-More time consuming

-Cause and effect cannot be inferred as the IV has not been directly manipulated

-P’s are not randomly allocated to conditions, which may reduce validity

Experimental designs

Repeated measures Independent measures Matched pairs

Description -Each pp is tested in all conditions of the study (i.e. Loftus & Palmer – car crash study)

-Same P’s in each condition

-Two (or more) groups of P’s – one for each condition

-P’s matched on key variables

Strengths -Good control for participant variables

-Fewer p’s needed

-No chance for order effects as pps are only used once , so no opportunity to ‘practice’

-No order effects

-Participants variables partly controlled

weaknesses -Order effects (boredom/practice)

-P’s guess the purpose

-Needs more P’s

-Lacks control of participant variables

-VERY difficult and time consuming to accurately match up all pps – even if using identical twins, environmental factors may lead to differences

Non-experimental methods

Self report methods- Easily repeated in exactly the same way than unstructured interviews - problem for self report methods is honesty - social desirability bias means that respondents may provide answers to put themselves in a good light – Open questions provide rich insight but are difficult to analyse

Questionnaires and interviews to find out what people think and feel - Interviews are essentially real time, face-to-face (or over the phone) questionnaires – option to conduct a fairly unstructured interview where the questions are developed by the interviewer as a response to the earlier answers given by the interviewee

Observational studies- Provide a rich picture of what people actually do (rather than what they say they do) - observers may be biased (observer bias) their observations can be affected by their expectations.

–Watch what people do – A lot of information to collect – Psychologists use behavioural categories to record particular instances of behaviour and sampling methods such as recording behaviour every 30 seconds (time sampling) or every time a certain behaviour occurs (event sampling)

Correlation analysis- does not demonstrate a cause - is useful in identifying where relationships between co-variables exist - Can be done with large data sets and can be easily replicated - there may be unknown variables that can explain why the co-variables being studies are linked together

– some studies are concerned with the relationship between two variables – Use correlation analysis

Case studies– Difficult to generalise from individual cases as each one has unique characteristics.

– detailed study of a single individual, institution or event - uses information from a range of sources, such as from the person concerned, and also from their family and

friends – Use interviews, psychological tests and observations – Generally longitudinal

Pilot studies - A small scale trial run of a research study to test any aspects of the design

Reliability

Inter-rater reliability Where two psychologists produce the same outcomeTest-retest reliability Where a person is given a questionnaire/interview/test on

one occasion and then this is repeated again after a reasonable interval – if the measure is reliable the outcome should be the same every time.

Validity

Ecological/external validity Concerns outside a study – the extent to which the results of the study can be generalised to other situations and people.

Predictive validity Your SATS results at age 11 are used to predict your GCSE success at age 16. Your GCSE average scores are used to predict your ‘A’ level success, and so on.

Content/internal validity Concerns things going on inside a study – whether the researcher did test what they intended to test.

Sampling technigues

Description Strength WeaknessRandom -every member of

the population has an equal chance of being selected e.g. taking names out of a hat/random name selector

-Potentially unbiased

-Needs to be drawn from a large population to be unbiased

Opportunity -Consists of those people available to the researcher e.g. approaching people in the street

-Easy and quick -Very biased – may be certain type of person around (students/workers)

Volunteer Participants self-select – volunteer to take past e.g. by advertising

-Access to a variety of p’s

-Volunteer bias – certain types of people may volunteer

Ethical issues

Informed consent -Aims of research must be made clear to the P’s and they consent to it – anything that may influence their participation must be disclosed

Confidentiality -If confidentiality cannot be assured this must be disclosed – Data protection act

Deception -Information must not be withheld from p’s – p’s should not be mislead

Debriefing -Following an investigation p’s should be fully informed in the nature of the research – their experiences should be discussed

Protection of p’s -P’s must be protected from mental an psychological harmWithdrawal -P’s have the right to withdraw at any time

Probability and significance

-Experimental (alternative) hypothesis – There will be a significant difference between the conditions or a significant relationship between variables.

-Null hypothesis – States there will be no difference or relationship.

-P = probability – The smaller the value of P the more likely we can accept experimental hypothesis and reject the null hypothesis.

-If P = 0.9 = 90% chance something is down to chance

-In psychology we need to be 95% certain our results were not by chance so to accept experimental hypothesis and reject the null hypothesis p ≤ 0.05 = 5% chance results were down to chance.

Type 1 error -Mistake of rejecting the null hypothesis when it is actually true – made mistake by accepting experimental hypothesis – A very strict P value makes this kind of mistake less likely.

Type 2 error -Mistake of accepting null hypothesis when it is false- should have rejected it – Strict P value makes this mistake more likely.

Statistical tests

Nominal data Data belongs to one group OR another

E.g. smoker/non-smoker, Male/female

Ordinal data Data is ranked in order E.g. 1st 2nd 3rd 4th

Nominal data Ordinal data Looking forIndependent

measuresCHI SQUARED MANN WHITNEY

Difference between 2 conditionsRepeat measures /

matched pairs- WILCOXON

Correlational - SPEARMANS RHO Relationship between 2 variables

How is it calculated Strengths WeaknessesMean Add up all values

and divide by number of values

Makes use of all the data

Can be unrepresentative of all data as it is affected by extreme valuesCan be misinterpreted if there are extreme

valuesNot appropriate for nominal data

Median Middle value from an ordered list

Not effected by extreme scores

Not as ‘sensitive’ as the mean – not all values are reflected in the medianNot appropriate for nominal data

Mode Most common value in a data set

Useful when data in categories

Not useful when there are several modesOnly measure suitable for nominal data

Range Calculated by finding the difference between the highest and lowest scores in a data set

Easy to calculate May be effected by extreme values

Standard deviation The spread of data around the mean

More precise as all values of data are taken into account

Some characteristics of the data are not expressed e.g. the influence of extreme values

Graphs

Bar charts – Height of bar chart represents frequency – suitable for words and numbers

Scattergram/scattergraph – Suitable for all correlational data – dot or cross is shown for each pair of values – shows positive/negative correlation – if there is no detectable pattern there is zero correlation.

Critical / Observed values

Observed value = result of statistical test

Critical value = found from table

Directional hypothesis = one tailed

Non-directional hypothesis = two tailed

Statistical test Observed value must be:SPEARMANS RHO Equal to or HIGHER than critical valueCHI SQUARED Equal to or HIGHER than critical valueMANN WHITNEY Equal to or LOWER than critical valueWILCOXON Equal to or LOWER than critical value