Please enable JavaScript.
Coggle requires JavaScript to display documents.
Quant W11- reliability and validity - Coggle Diagram
Quant W11- reliability and validity
validity
the extend that an effect demonstrated in research is genuine, not produced by bogus variables, not limited to a specific context
ways to improve validity
standarised procedures- for both ppts and researchers
counterbalancing- in repeated measures designs to avoid order effects
blinding- eliminate possible variance due to ppt and researcher ecpectations
types of validity
internal validity
the extent to which an effect found in a study can be taken to genuinely cause by manipulation of the IV
external validity
the degree to which results generalise beyond the experimental context
the extent to which it is possible to make inferences from your sample to the larger population
ecological validity
extent to which research effect generalises across setting
construct validity
how should we measure psychological constructs that are abstract
happiness-intelligence, stress, self-esteem, motivation, emotion
population validity
extent to which research effect can be generalised across people
historical validity
whether an effect would stand the test of time and work today as it did some years ago
causal relationship
manipulation of the IV directly cuases changes in the DV
to be internally valid research needs to avoid any confounding varibales
threat to internal validity
attrition/mortality
ppt dropout rate
in some types of research ppts might be lost through death
common in longitudinal studies
history
evetns that occur between measurements
eg political issues, natural disasters
sampling
relates to sampling methods used
often a result of non-random assignment of ppts to conditions
pre-screening in the presence of confounding variable
maturation
more present in longitudinal studies
ppts may change over the course of the study
children vs adults maturation rate
testing and instrument issues
used to operationalise IVs and DVs
repeated testing using the same measures
eg order effects, practice effects
expectancy effects
ppts expectancy
eg Hawthorn effect (individual modify their behaviour in response to their awareness of being observed
the effect of the ppts expectancy about what they think is supposed to happen in a study
observer/experimenter
researchers expectation for the study affect their behaviour and how ppts respond to this
interactions between experimenter and participant
subtle differences in participant treatment
demand characteristics
cues in the study which help ppts to work out what is expected of them
pleasing the experimenter
social desirability bias
enlightenment- an issue when using psych students for research
blinding
single blind- either ppt or research assistant unaware of condition
double blind- both ppts and research assistant unaware of condition, standard for clinical trials (RCTs)
reliability
the extent to which a measurement is reproducible or consistent over time
reliability over time
a test/measure should yield consistent results when repeated measures are taken under the same conditions
eg- questionnaires, tests, a computer that records responses etc
types of reliability
external validity
test-retest reliability- correlations of peoples scores at one time, and at a later time, not relevant to all constructs
inter-observer or inter-rater reliability- extent to which researchers agree in thier rating or coding
internal reliability
internal consistency of the test
ppts tend to score similarly across multiple items of a construct
eg- a personality questionnaire, routinely score highly on neuroticism items
why does validity and reliability matter?
credibility of research and scientific rigour- like replication, important to the discipline
;to ensure findings dont conflict with eachother
clinical importance: mental health and treatment may be inaccurate
reliability and validity of quant data collection methods
experiments and RCTs
lab based and fully controlled- experimental manipulation of the IV, standardised procedures, random allocation of ppts
internal= high, external=low
quasi experiments
same as experiments but lacking either/both- random assignment into treatment and control groups, full control over the IV
internal= low, external= high
correlational studies
used to determine if one factor is related to another factor- non manipulated variable, non causal hypothesis
internal=low, external= high
questionnaires
frequently used data collection method
often multiple questionnaire measures used within a study
using multiple questionnaires to measure similar traits
validity of Qs- face= do i know what its trying to test?, content= do they capture the relevant elements of material?, criterion= testing the questionnaire with different groups, construct= how do we actually measure the construct