Kin 380 - Chapter 6

logical validity

- the quality of researchers' arguments, their application of theory to support their study, and appropriate interpretation of results
- answers the question "does this logically make sense, and has it been sufficiently justified by the researchers?"
- looking for logical flow

construct validity

- do the measures used by researchers actually measure what they intend to measure
- measuring variables is a fundamental part of research
- a measure should have validity evidence
- validity is a matter of degree; not all or nothing
- content validity
- convergent validity
- discriminant validity

nomological network

- a web of evidence that researchers collect to justify their choices for using particular measures and support that the measures indeed measure what they intend to measure

content validity

- does the content make sense
- on the surface, the measure looks like it assesses what it's supposed to assess
- does not require reporting stats
- usual first approximation of validity

convergent validity

- does measure converge with things it's related to
- consists of moderate to high correlations with measures of similar constructs

discriminant validity

- how does it differ from constructs it should differ from
- evidence consists of zero or low correlations with measures of different constructs

measurement validity

- the extent to which a variable has been adequately measured

measurement reliability

- the consistency (or repeatability) of a measure
- reflects the degree to which a measure is free from error variance
- any measurement is comprised of two components:
1) a true score
2) measurement error

internal consistency

- the consistency of scores within a test, or the degree of interrelatedness among items
- assessed by Cronbach's coefficient alpha, ranging from 0 to 1

Cronbach's alpha

- based on inter-item correlation
- the higher the items covary the greater the internal consistency
- a limitation is that it is affected by the number of items

test-restest reliability

- coefficient between scores on repetitions of the same test
- not appropriate for measures affected by repetition or dynamic constructs (eg. one-leg balance test)

b

Which statement is true?
a) a measure can be valid, yet not reliable
b) a measure can be reliable, yet not valid

internal validity

- the merits of researchers' claims that any change in an outcome is a result of their intervention and not the result of other factors

internal validity (threats)

- factors that may alter the dependent variable separate from the effects of the IV are considered threats
- may be related to experimental procedures, treatment/manipulation of the IV, of participant characteristics

control group

- many of the threats to internal validity can be controlled for by having a control group as similar to the experimental group as possible

random assignment

- increases the likelihood of groups being similar prior to intervention
- treatment should be the only systematic effect

external validity

- the generalization of the research findings to other people, contexts, or time

external validity (threats)

arise when experimenters draw incorrect inferences from the sample data to other persons, other settings, and past or future situations

WEIRD research

- western
- educated
- industrialized
- rich
- democratic
- can limit generalizability of research findings