research methods - psychology

aim

general statement that the researcher intends to investigate.

hypothesis

clear, precise, testable statement that states the relationship between variables

directional hypothesis

states direction (more, less), conclusive previous research

non directional hypothesis

no direction (differ), no conclusive previous evidence

IV

experimenter manipulates + measures impact on dv

DV

experimenter measures

operationalisation

make measurable (units)

extraneous variables

variable that effects DV that isn't IV (annoying)

confounding variables

type of EV, can't tell if it is the CV or the IV, systematic

investigator effects

impact of investigators behaviour on an outcome

randomisation

chance methods to reduce unconscious bias + investigator effects

standardisation

everything exactly the same for all participants, reduce EVs

experimental design

ways participants can be organised in relationship to experimental conditions

independent group design

participants allocated to a group that represents an experimental condition

repeated measures

all participants in all conditions

matched pairs design

pairs matched on key traits, one in group A and one in B

random allocation

each participant has the same chance of being in each condition (dec investigator effects)

counterbalancing

half AB half BA (reduce order effects)

lab experiment

controlled enviroment (measure effect of manipulated IV on DV + EVs kept track of)

field experiment

natural setting (IV manipulated + effects on DV measured)

natural experiment

measure the effect of a natural change in IV on a decided DV

quasi experiment

variables simply exist (IV not determined by anyone)

experimental condition

group with IV

control condition

group with no treatment (compare to)

population

group the sample is drawn from, the group that are the focus of the research

sample

group who take part in the research, representative of the population

random sampling

all target populations have an equal chance of being selected (random no. generator)

systematic sampling

every nth no. of the target population (e.g. every 10th person on the register)

stratified sample

sample reflects proportions of subgroups in the population

opportunity sampling

anyone willing + available

volunteer sample

participants select themselves (-ve = particular personality)

sampling bias

certain groups not correctly represented in the sample, limiting generalisations

generalisation

extent findings can be applied to populations, possible if representative sample

ethical issues

conflict between participant rights and goals of research to produce worthwhile data

informed consent

participants are aware of the aims of the research, reduce natural behaviour (consent letter, presumptive, prior general, retrospective)

deception

deliberately misleading or withholding info (full debrief + right to withhold data + counselling)

protection from harm

no more physical+ psychological risk than in their daily lives (right to withdraw at any point + withhold data + counselling)

confidentiality

participants control info about themselves, right of privacy (anonymity + not shared)

pilot study

small scale investigation before real one (check stuff works and allows modifications to be made)

single-blind

participants not told aim or condition (reduce demand characteristics)

double-blind

participant + researcher don't know who is in which condition

naturalistic observation

watching/recording behaviour where it would normally occur

controlled observation

watching/recording behaviour in a structured environment

covert observation

participants unaware of observation

overt observation

participants are aware + have consented to observation

participant observation

researcher becomes a member of the observed group

non-participant observation

researcher remains outside of the group

behavioural categories

target behaviour broken down into operationalised components

unstructured observation

write down everything they see

event sampling

record every time target behaviour occurs (tally)

time sampling

records no. time behaviour occurs in a fixed time frame e.g. every minute

inter-observer reliability

agreements between observers, correlating observations total agreed / total observations, >80 = high reliability

questionnaire

set of written questions used to assess a persons thoughts/experiences (open / closed questions)

interview

live, questions asked to assess the others thoughts / experiences (structured / unstructured)

correlation

association between two co-variables (+ve or -ve or none - no variable manipulation, can't establish cause + effect)

qualitative data

words or non numerical

quantitative data

data that can be counted

primary data

info obtained first hand by researcher for the purpose of a project

secondary data

info that has already been collected and pre-dates the project

meta-analysis

combining findings from multiple studies on a topic to produce overall conclusions

distribution

normal = mean, mode + median all same -ve = tail left, mean -> mode -> median +ve = tail right, mode -> median -> mean

peer review

ensuring high research quality by being assessed by experts (research funding, validate relevance, suggest improvements)

economy

maximum amount of people at work (£15bn, women at work/childcare)

case study

in depth investigation into an individual, group or event

content analysis

indirect study of behaviour by examining communications of others

coding

quantitative data, identify each instance of chosen categories

thematic analysis

qualitative data, identifying implicit + explicit ideas within the data (often emerges after coding)

reliability

how consistent (clustered)

test-retest reliability

assess the same person on two separate occasions (similar answers = reliable)

validity

extent observed effect is genuine + measures what it was supposed to + generalise (centre)

face validity

basic - look at it

concurrent validity

psychological measure relates to existing measure

ecological validity

findings can be generalised (external)

temporal validity

generalised to other time periods (external)

stats choosing

experiment/correlation, data type, unrelated/related/association (Carrots Should Come Mashed With Swede Under Roast Potatoes)

nominal data

discrete categories (e.g. yes or no)

ordinal data

ordered (e.g. rate how great i am, subjective)

interval data

numerical scales (e.g. science with time)

chi - square

independent group or correlation, nominal

sign test

repeated measures, nominal

mann whitney u test

independent group, ordinal

wilcoxon

repeated measures, ordinal

spearmans test

correlation, ordinal

unrelated t test

independent group, interval

related t test

repeated measures, interval

pearsons test

correlation, interval

probability

likelihood that event will occur, 0 - 1

significance

how sure we are that a difference/correlation exists, allow to reject null hypothesis (0.05)

type I error

incorrect rejection of true null hypothesis (false positive)

Type II error (beta)

failure to reject false null hypothesis (false negative)

abstract

key details of the research report

intro

look at past research on similar topic, aim, hypothesis

method

description of what the researcher did (design, sample, apparatus, materials, procedure, ethics)

results

description of what the researcher found (descriptive + inferential stats)

discussion

consideration of what results tell us

references

list of sources in full detail

objectivity

all personal bias minimised + doesn't influence research

empirical method

scientific approach based on gathering of evidence through direct observation + experience

replicability

is research can be repeated by other researchers

falsifiability

Popper - theory can't be scientific unless it can be proved false