|
- Reliability:
Do indicators produce consistent results in measuring a
concept?
- Just
as there are different forms of validity, there are
different forms of reliability.
- Are
you seeking consistency over time, over
observers, or across measures?
- Test-retest
reliability refers to consistency over time:
- To
what extent will a test at time t yield the
same results at time t+1?
- Your
research determines whether you value this type of
reliability.
- Inter-observer
reliability refers to consistency across
researchers.
- To
what extent do different researchers produce
identical scores on indicators?
- This
is especially important for research projects
dependent on team research.
- It's
also important when one researcher uses a
subjective method, such as content
analysis.
- Internal
consistency applies in using multiple indicators in an
"effects" model of measurement.
- We
will measure scale reliability only in the sense of
internal consistency.
|