RM

randomisation

use of chance to reduce bias from investigator effects (using a random number generator/ selection of pp)

Replicability

the extent to which an experiment can be repeated and produce the same results

Falsibility

a hypothesis should be falsifiable- able to prove them false

empirical method

acquiring knowledge through investigation and data

objectivity

studies show that the more quantitative a study is, the more objective- scientists should remove their bias

paradigm

A shared assumption/ belief
A paradigm shift \= a shift in what's believed

quantitative

data expressed numerically and statistically eg. closed q's and experiments

evaluate

s\= - easy to evaluate
-more objective (no bias)

w\= - less external validity
- lacks detail

qualilative

data expressed through thoughts, feelings and interpretation eg. diaries and discussions

evaluate

s\= - broader detail
- greater external validity

w\= -hard to analyse
- relies on interpretation

primary data

data collected first-hand specifically for the investigation

test-retest

administering the same test to the same person on different occasions
it measures the same variable each time; is a good term-42check of consistency

strengths (p)

-fits the job/ targets info
-authentic and validity

weakness (p)

-time and effort
-expensive
-planning and prep needed

secondary data

information that's already been collected and is not specifically for the aim of the investigation

strengths (s)

-already exists
-cheap and accessible

weakness (s)

-hard to validate
-variation

meta-analysis

particular form of secondary data
data from a large range of studies with the same question and method are combined

strength (ma)

-large and varied sample
-combination of studies

weakness (ma)

-publication bias
-may pick the wrong studies

reliability

how consistent a measuring device is
they should measure a variable in the same way with similar results

measuring device

a tool used in an investigation to measure the dependant variable

test retest process

-participants scores should be CORRELATED with their 2nd set of scores
-these are plotted on a scatter graph with both scores on x and y axis

test-retest waiting period?

-long enough for pp to forget their responses
-but not long enough for the pp to change significantly

Inter-observer reliability

having 2 or more observers compare results

Inter-observer reliability process

-both researchers agree on behavioural categories and observe independently
-their results are put on a scatter graph and correlated

Validity

whether results are found to be true

internal validity

extent at which a test measures what it set out to measure

external validity

extent at which findings can be generalised beyond a research setting (race, gender)

temporal validity

extent to which findings can be applied outside of the time period they were found in

population validity

extent to which findings from a small sample can be generalised to a whole population

face validity

asking an expert to look at a measure and seeing if on the face of it it measures what its supposed to

concurrent validity

having pp do your measure and an established measure then correlating the results on a scatter graph

lab experiment

iv\= directly manipulated by researcher
env\= artificial

evaluate

s\= -more reliable (controlled variables)
-high internal validity; variables are controlled

w\=lacks external validity

field experiment

iv\= directly manipulated by researcher
env\= real world conditions

evaluate

s\= -high external validity due to real environment

w\= -hard to control extraneous variables and standardise procedures so
low reliability + internal validity

natural experiment

iv\= not manipulated by researcher, a naturally occuring change in situation
env\= artificial or real life

evaluate

s\= -can investigate natural situations that would be unethical to manipulate

w\=- no control over iv \= confounding variables

quasi

iv\= not manipulated by researcher, pre-exisitng difference in pp
env\= artificial or real life

evaluate

s\= -typically an artificial environment \= internal validity and reliability

w\= -no control over iv \= confounding variables

Questionare

a pre set list of written items
can have
open q (no fixed range of answers)
closed q (fixed responses )

evaluate

s\= -low cost
-distributed to large groups
-doesnt need a present researcher

w\= -untruthful responses
-social desirability
-response bias

questionare scales

likert scale\= agreement with a statement
rating scale\= strength of feeling about a statement
fixed choice\= possible options they need to tick

interviews

interaction between interviewer and interviewee

structural \= pre determined set of q so high replicability

semi structured \= mix, pre set q but elaboration and follow ups are allowed

unstructured \= despite the topic, it's conversational with no set q
more detail but hard to analyse

what are
leading q
double barrelled questions
double negatives

-prompts a specific answer
-2 questions in 1
-difficult to answer

observation

watching and recording pp behaviour

controlled vs natural

c\= environment set up by a researcher
n\= environment with no interference

covert vs overt

c\= pp don't know they're being observed
o\= pp know they're being observed

participant vs non

p\= researcher becomes involved
n\= researcher is on the sidelines

unstructured observation

-researcher writes down everything they see
-small observations with few pp
-qualitative data\= observer bias

structured observation

-ideal when lots is happening
-hard to record \= observe target behaviour
- quantitative data -\> lacks validity

Behavioural categories

must
-be precise
-be observable
-no overlap
-include all behaviours

event sampling

tally each time behaviour in a category is shown
s\= record at anytime, don't miss anything
w\= doesn't explain outside of given time

time sampling

record behaviour at specific time intervals
s\= how time effects behaviour
w\= miss key points outside of time points

what's content analysis?

1. familiarise yourself with data and get grasp of potential themes

2. identify codes.
go through data and pick specific themes

3. begin coding the data
comb through looking for references

4. after, count the frequency of references to see which code is most present/ present
also, carry out statistical analysis

evaluate

s\= -easy to analyse
-not an invasive method of investigation (no contact with people)

w\= -subjective
-data becomes only quantitative and lacks detail

pilot study

small scale trial run of an actual investigation

case study

-in depth description, investigation and analysis of an individual/ event
-usually qualitative eg questions and observations
-take a long time plus can include family + friends

evaluate

s\= -allow detailed insight
-understanding of normal functioning
-help produce future hypothesis

w\= - small sample so specific findings -subjective findings -personal experience
LOW VALIDITY

informed consent (ethical issue)

-making pp aware of the research aims, procedure, rights (eg to withdraw) and what their data is used for
-not always used, demand characteristics

how is it dealt with?

-pp are given a consent letter
-presumptive consent\= asks a similar group whether they'd give consent, depending on answer, consent is assumed
-prior general consent\= pp asked if they'd take part in a study where they might be deceived
-retrospective consent\= consent is asked after the study (eg if they've been deceived)

deception
protection from harm

-hiding information on purpose as long as no distress is caused -not placing pp in any more risk than daily life

how is it dealt with?

-pp are given a full debrief with the true aim of the study at the end
-right to withdraw if they're distressed
-reassure pp they're normal, offer counselling to extreme cases

privacy and confidentiality

-pp have a right to control information about themselves

how is it dealt with?

-don't record personal pp details, maintain anonymity

what to include in a consent form

-explanation of investigation
-free to ask questions, allowed to withdraw, no pressure
-space to sign
-step by step process
-data is confidential and anonymous
-how long is study gonna last
-full debrief

aims

what a researcher is aiming to find

hypothesis

testable statement regarding an outcome

Test of Correlation

looks for a relationship between 2 co-variables

Test of difference

identifies the differences between variables it can be grouped, iv and dv

directional hypothesis

a specific outcome can be anticipated between co-variables

Non-directional hypothesis

can only anticipate a change but don't know the effect

extraneous variable

a nuisance variable that doesn't consistently effect an experiment/ population eg traffic in one direction

confounding variable

a nuisance variable that systematically effects an experiment eg invalid test

investigator effects

their own expectations may influence pp-body language-pp selectionbias to themselves

Participant variables

the varying individual characteristics of pp which effect an experiment (extraneous)

demand characteristics

pp are influenced by cues regarding the experiment-please u effect to please the experimenter-screw u effect to sabotage hypothesis-social desirability

order effects

the difference in pp performance due to order in which they complete the experiment conditions

Counterbalancing

reduces order effects by splitting pp in half and each completing an experiment in alternating turns

standardisation

using the same instructions and method in an experiment for each pp

randomisation

The process of making something random to avoid systematic error

random allocation

giving all pp an equal chance to take part in an investigation to avoid systematic bias (number generator)

population

people being studied that a researcher wants to make conclusions about

sample

random portion that is representative of the population

independent groups

pp divided in 2 groups- 1 group does 1 condition and a second does another

Evaluation

s\= -no order effects-no demand characteristics
w\= -differences in pp are extraneous variables

repeated measure

all pp do the 1st condition followed by the 2nd condition

evaluation

s\= -no pp variables
w\= - order effects-demand characteristics

matched pairs

pp are in 2 groups that do different condition however, individuals from each group are matched in pairs based off key characteristics

Evaluation

s\= reduces pp variables-no demand characteristics-no order effects
w\=- time consuming + expensive-pp differences are extraneous variables

what makes an experiment

-test of difference
-iv and dv
-cause and effect

Random samples

-members of a target population have equal chance of being selected
-how? \= randomly number pp in a list and assign them a condition

Evaluate

S\= - no researcher bias- fairly representative
W\= - difficult and time consuming to compile list- no obligation to do so it becomes a volunteer sample

Systematic sample

- every nth member of a population is chosen-how? compile a list and nominate a sampling system

Evaluation

S\= no research bias fairly representative
w\= time consuming and difficult ends up becoming a volunteer sample

Stratified sampling

-proportions of sub groups
-how? research identifies the strata and proportions are calculated to represent the population

Evaluation

S\= no research bias highly representative
w\= not fully reflective, stratification is only at one level
Ends up becoming a volunteer sample

Volunteer sample

- self selection method
-how? Researcher puts a notice out

Evaluate

S\= minimal effort
w\= only one type of person (eg friendly )