Please enable JavaScript.
Coggle requires JavaScript to display documents.
6: Validity and Reliability (Focus: Qualitative methods) (Reliability…
6:
Validity and Reliability
(Focus: Qualitative methods)
Reliability necessary for validity
invalid = worthless
Nothing is 100% valid
Triangulation = Different approaches/methods leading to same result
(Denzin, 1970)
adds one more:
within methods
: replica
Both quantitative and qualitative measures
Levels: from group to society: If possible, several levels are to be preferred
Several kinds of validity, e.g. content validity, cultural validity, face validity +++
External validity: Possibility for generalization
positivism:
Sine qua non
Naturalistic research:
Comparability
and
translatability
. Multi site studies
Threats:
selection effects
settings effects
history effects
construct effects
constructs
:
Must be a construct that others agree measures the problem in question
Must be meaningful to the participants themselves
Tests
(Campbell and Fiske, 1959)
:
Convergent techniques
: See if different methods with same construct produces similar results
Discriminant techniques
: Use same method on different constructs to see if results differ
Must different constructs necessarily produce different results :question:
Research must be true to paradigm, thus validating findings within traditions of that paradigm
In qualitative research, replace validity with authenticity/understanding
(Maxwell, 1992)
Meanings that subjects give to data is important. Honesty the most important virtue
(Agar, 1993)
: Qualitative research validated by the in depth involvement of researcher
(Hammersley)
: No! Researcher holds no such privileged position. Researcher only represents reality, reproduces not.
Data collection must be representative of the whole sample, field, i.e.
(Fielding and Fielding, 1986)
(Maxwell, 1992)
: Validities in qualitative research:
Descriptive
(not made up stories)
Interpretive
Theoretical
: (explanations)
Generalizability
(internal validity rather than external)
Evaluative
(judgmental framework instead of descriptive)
Internal validity:
low-inference descriptors
:star:, p. 135
multiple researchers
participant researchers
peer examination of data
mechanic recordings
(LeCompte and Preissle,1993)
Validities:
Fairness
(balanced representation) of the multiple realities in, and constructions of, a situation.
Ontological authenticity
: (e.g. making the familiar strange, reducing ‘cultural blindness)
Educative authenticity
(new appreciation of these understandings).
Catalytic authenticity
(gives rise to action)
Tactical authenticity
(benefit to all involved)
Other forms of validity
Ecological validity
: Trying to study subjects in natural settings
Important in education studies :!:
Challenge: To give a detailed description of a situation without identifying participants
Cultural validity
(Morgan, 1999)
: To appreciate and respect the culture within what you study takes place
Catalytic validity
: Research should lead to positive action
Related:
Consequential validity
Criterion-related validity
:
Predictive validity
: Results from first round corresponds to later research
Concurrent validity
: Different instruments give same result at same point in time
Oppskrift på hvordan øke validitet, s. 144
E.g. avoid leading questions
Halo-effect
: Researcher's knowledge of subject influence judgement