Beruflich Dokumente
Kultur Dokumente
2 (TFI)
2008-18861 Prof. Rafael Villar
Validity
A true measure
The ability of measurement procedures to produce empirical data consistent with the theoretical meaning of the
concept to be measured
An indicator may be valid for a particular purpose and definition, but less valid or invalid for others
We can never have absolute validity but some measures are more valid than others
2. Content Validity
- Captures the entire meaning
- Measures should sample or represent all ideas or areas in a conceptual space
- Comparing the items in a measure with items which could have been used permits assessing its content validity
- Question answered:
Is the full content of a definition represented in a measure?
- Three steps:
1) Specify the content in a construct’s definition
2) Sample from all areas of the definition
3) Develop an indicator that taps all of the various parts of the definition
3. Criterion Validity
- Agrees with an external source
- Uses some standard or criterion that is known to indicate a construct accurately
- The validity of an indicator is verified by comparing it with another measure of the same construct in which a
researcher has confidence
- Two sub-types:
a. Concurrent
- An indicator must be associated with a pre-existing behavior just that is to be judged to be valid
- Ability to distinguish between groups that it should theoretically be able to distinguish between
b. Predictive
- An indicator predicts future events that are logically related to a construct
- Measure can correctly predict something that theoretically it should be able to predict
4. Construct Validity
- Multiple indicators are consistent
- Requires a definition with clearly specified conceptual boundaries
- Question answered:
If the measurement is valid, do the various indicators operate in a consistent manner?
- Two sub-types:
a. Convergent
- “Alike ones are similar”
- Multiple measures of the same construct hang together or operate in similar ways
b. Discriminant (divergent)
- “Different ones differ”
- Indicators of one construct hang together or converge, but also diverge or are negatively associated with
opposing constructs
Reliability
It is a dependable measure
Response is similar and predictable in different times and conditions
The method of conducting a study or the results from it can be reproduced or replicated by other researchers
Stability and internal consistency
Reliability is inversely related to random error
Improving Reliability
- Item analysis – most effective way to increase reliability
- Consists of computation of item difficulties and item discrimination indices
- Items that are too difficult, too easy, and/or have near-zero or negative discrimination are replaced with better items