Psych- research methods

central tendency

mean median mode

mean

the average, eval- uses all data = accurate representation, influenced by extreme data/outliers

median

the middle score when numbers are in numerical order, eval- unaffected by outliers, all other data is ignored

mode

the most frequent number in the data set, eval- unaffected by outliers, ignores other data

measures of dispersion

range, standard deviation

range

subtract highest data from lowest, eval- quick and easy to calculate, vulnerable to extreme scores

standard deviation

how far each data deviates from the mean, eval- considers all data, can only be used with interval data

observations

controlled and naturalistic, overt and covert, participant and non-participant, behaviour categories, inter-rater reliability

controlled/lab

studying behaviour in a controlled, structured environment, easily replicated, data from observations can be easier to analyse, can lack validity- social desirability (Hawthorne Effect)

naturalistic

studying behaviour in its natural setting, eval- greater ecological validity, less reliable, harder to replicate

participant

where the researcher joins in, becomes part of the group, eval- greater understanding, have to rely on memory as cant openly take notes

non-participant

where the researcher does not take part in the group, eval- a more objective view of what is occurring, social desirability, (Hawthorne Effect)

covert

ppts are unaware they are being observed, eval- no social desirability, no consent

overt

ppts are aware they are being observed, eval- more ethical, social desirability

behavioural categories

researchers have to decide what behaviour to observe, breaking the behaviour eg aggression down into eg hitting, kicking which can be observed/measured, eval-

inter-rater reliability

measures the consistency and agreement of people evaluating the same thing

positive correlation

both variables increase or decrease

negative correlation

as the amount of one variable increases/decreases the other decreases/increases

no correlation

no relationship between the two variables

questionaire

are a self report technique where ppts are given set questions to answer, open and closed eval- easy/cheap, social desirability

ethics - British Psychological Society’s code of ethics

right to withdraw, consent, de-briefing, deception, confidentiality

type 1 error

false positive, accepts the alternative hypothesis and rejects the null

• too lenient with the p level/significance

type 2 error

false negative, accepts the null hypothesis and rejects the alternative hypothesis

• too strict with the p level and significant level

peer review

before a study is published a peer review is to check validity and quality of the study

-ensures only high quality research is published as it can become a part of mainstream thinking

EVAL- promotes high standards in research, prevents scientific fraud

-contributes to the “file drawer effect” only statistically significant findings are published

probability, p level

the likelihood of an event occurring as 0.5 (50%)

random sampling

every member of the target audience has an equal chance of being selected eval- fair, impractical group might be too larger

systematic sampling

a systematic methods is chosen for selecting from a target group, eval- unbiased, not given an equal chance to be selected

stratified sampling

divides the target group into sections, showing a key characteristic, same ratio of the characteristics in the groups as in the target population , eval- avoids misrepresentation, takes time

opportunity sampling

ppts willing and accesible to take part are targeted, eval- easy/cheap, not representative, bias

volunteer sampling

ppts who volunteer to be in study eval- a large sample size, unrepresentative data

null hypothesis

this predicts a relationship will not be found

research hypothesis

a significant relationship between variables, it will be directional (one tailed) if a specific direction is made but non-directional (two tailed) if it says there will be a difference but no direction

IV

is changed

DV

is measured (affected by IV)

EV

affects the IV, could affect the research

statistical testing

reliability

measure whether something stays the same,

internal reliability

internal consistency of a measure eg whether questionaire questions are measuring the same thing method- split half method, split data in half then compare

external reliabilty

assesses consistency from one use to another, method- standardising procedures eg. IQ one year to another year

validity

whether a measure actually measures what it claims to be measuring

face validity

whether it looks subjectively promising measuring what it’s supposed to IMPROVE- questions revised so they relate more to the topic

internal validity

whether study, conduct, analysis answer the research question without bias IMPROVE- better research design → double-blind

external validity

whether data can be generalised to other situations IMPROVE- better research design → double-blind

temporal validity

findings apply across time (is it still valid now)

ecological validity

whether data can be generalised to the real world

matched pairs

ppts recruited specifically to relevant characteristics, eval- no order effects, difficult and expensive

repeated measures design

same ppts take part in all conditions, eval- not subject to individual differences, order effects

independent groups design

where different ppts take part in different condition, eval- no order effects, less time consuming, difficult and expensive to recruit

nominal data

data organised into categories

ordinal data

data in rank order

interval data

can be assigned a numeric value

ratio

a continous scale with absolute zero

pilot study

a small scale study before engaging in a larger scale eval- identify mistakes and effectiveness- ignores the irrelevances of the study

quasi experiment

experimental approach but cant control the IV bc its not in the researchers control EVAL- high internal validity

primary data

data that is collected by the researcher- questionaires, surveys, interviews

secondary data

information that already exists that has been collected by someone else

content analysis

used to analyse qualitative data and turn it into quantitive data - identifying common categories and then re reading and tallying every time it reappears

counterbalancing

used to deal with order effects when using a repeated measures design

-each group does the experiments in different orders (group 1- condition A then B, group 2- condition B then A)

normal distribution

where mean, mode, median are equal

skewed distribution

where mean, median and mode are not equal

inferential stats

the stats table

descriptive stats

mean, median and mode

event sampling

where an observer records the number of times a certain behaviour occurs EVAL- better representation of the experiment at hand

time sampling

where an observer records behaviour at prescribed at intervals EVAL- can miss behaviour, hard to determine the intervals

psychology and the economy

more effective treatments can be developed for psychological health problems which means people will be able to return to work and this reduces the burden on the employers, NHS and taxpayer

concurrent validity

compares a new test with an old test to see if they produce similar results IMPROVE- removing seemingly irrelevant questions the re-checking the validity

thematic analysis

converts qua

random allocation

random allocation of participants to conditions in an experiment (names out of a hat, computer generater)

field

constructed in a more natural environment but not in a lab but variables are still well controlled EVAL- higher ecological validity because more natural behaviour occurs

lab

in a controlled environment variables are controlled EVAL- greater accuracy and replicability, lacks ecological validity

natural

nat environment no control of variables EVAL- high external validity (real life issues) hard to randomise

displays of quantitive data

bar charts, histograms, scatter graphs, tables

demand characteristics

participants will change their behaviour in line with what their interpretation of the study is looking for → social desirability effect, hawthorne effect

investigator effects

when the researcher unintentionally influences the outcome of their research (non-verbal → raised eyebrow = change in behaviour + physical characteristics → gender )

single blind procedure

where researchers do not tell the ppts that they are given a a test or control treatment

double blind procedure

where ppts nor researchers knows whos receiving a particular treatment

structured interview

set of questions

unstructured interview

general headings no specific questions

randomisation

random allocation to conditions

standardisation

everyone gets exactly the same condition EVAL- increases reliability and validity and replicability

hawthorne effect

observing smth which will change the behaviour of what is being researched → help you effect and screw effect