Validity & Reliability Flashcards
1
Q
Psychological Influence
A
- psychology dominated by white, middle class men; “elite” at the time; research justified their position
- if they set out w/belief that women are naturally less intelligent, it’ll reflect in work; IQ tests seem to support them; is it now proven?
2
Q
Context
A
- context/time of research is vital when analysing articles; writers are inevitably context bound
- researcher is in charge of question development, etc.; can never be unbiased; naïve to think otherwise
3
Q
Validity
A
- the extent to which the test/method truly measures what its intending to; think operationalisation (ie. a maths test shouldn’t be testing verbal skills)
- 1+1 = measure of addition; but is it IQ?
- DART-BOARD CHART: reliable/invalid = side cluster; unreliable/invalid = mess; reliable/valid = bullseye
4
Q
Invalidity in Media
A
- headlines often rooted in truth but media often dangerously misrepresents results/manipulates psychological terms (ie. “breastfeeding increases likelihood of good health” could become “breastfeeding causes good health”)
- careful that sources have a healthy air of uncertainty
5
Q
Validity in Research
A
- various validity types; the translation of psych concept to tangible test/study (ie. Beck Anxiety Scale); aka. operationalisation
- sampling is similar; pop is construct and the sample operationalises
- valid finding = no alternative, logical explanation
6
Q
VR: Face
A
- is the concept well translated (your opinion doesn’t count)
- measured via sending test to experts; reply must describe whether measure analyses what it was supposed to
7
Q
VR: Content
A
- test’s ability to represent all aspects of the concept you’re trying to measure
- what criteria will be used to consider variables that may affect the concept (ie. what constitutes as self-esteem?)
8
Q
VR: Predictive
A
- assessing operationalisation’s ability to predict something it theoretically shouldn’t (ie. measure of math ability theoretically predicting engineering based profession value)
- used as screening device (ie. SATs)
9
Q
VR: Concurrent
A
- assessing operationalisation’s ability to distinguish between groups it is expected to (ie. in assessing manic depression, the measure should tell the difference between said diagnosis and a paranoid schizophrenic diagnosis)
- discriminating test; results more powerful if the test can discriminate between 2 similar groups
10
Q
VR: Internal
A
- to what extent does the study say what is expected
- threatened by:
HISTORY EFFECT
MATURATION EFFECT
REGRESSION TO THE MEAN
TESTING EFFECT (1)
TESTING EFFECT (2)
TESTING EFFECT (3)
INSTRUMENTATION EFFECT
SELECTION EFFECTS
MORTALITY
11
Q
VR-I: History Effects
A
- event (not manipulation) between pre-test/post-test; affects outcome (ie. study about attitudes to terrorism; publicised terrorist attack during study)
- control group allows accommodation as experimental effect is isolated; effect should be equal
12
Q
VR-I: Maturation Effect
A
- pps naturally changing between pre-test/post-test
- ie. study investigating effects on classroom intervention on moral reasoning in kids; kids hormones changing
13
Q
VR-I: Regression to the Mean
A
- if the pop performs poorly in pre-test, it’s likely to improve at post-test independently
- comparison w/control allows measurement
14
Q
VR-I: Testing Effect I
A
- pps perform BETTER at post-test due to PRACTICE
- control has similar practice opportunity
15
Q
VR-I: Testing Effect II
A
- pps perform WORSE at post-test due to tiredness
- control would also do this
