2. criteria-related validity (criterion validity)
Criterion validity defines the degree of statistical agreement between the test result and the criterion for which the test was determined. (Example: 30-meter sprint is correlated with long jump performance. )Calculated correlation = criterion validity (validity coefficient) Criterion validity is considered particularly important in performance diagnostics.
The criterion validity is divided into Determination of the criterion validity: With correlation of r=1, the performance can be estimated without error. external criterion validity: The test performance is correlated with an external criterion. (e.g. 6-jump is correlated with long jump) inner criterion validity: The test performance is correlated with other measured values of the same validity range.
(e.g. CounterMovement-Jump is measured with the jump belt test)
- Compliance validity – test and criterion values are collected simultaneously. Compliance validity should be greater than 0. 80.
- Prediction validity – first test, then collection of the criterion values
- Determination of the test values (X) on a sample
- Collection of the criterion values (Y) on the same sample
- Calculation of the correlation (X,Y) = rxy
3. construct validity
Definition:Construct validity is the extent to which a measurement procedure actually captures the theoretical construct it is intended to measure. In most cases, construct validity is determined by means of confirmatory factor analysis. Construct validity is important for the theoretical clarification of what the test should measure.construct = theoretical/ideal structure (e.g. endurance, strength, speed, intelligence etc.) Constructs can be divided into: Furthermore the construct validity is divided into: The construct validation is done in 3 steps:
- Homogeneous constructs: Skills that are closely related (e.g. bounce – maximum force)
- Heterogeneous constructs: Skills that are not closely related (e.g. strength – fitness)
- Convergent validity (consistency with other tests that measure similar factors)
- Discriminant validity (with which variables of other constructs the test should not correlate)
- Derivation of testable hypotheses
- Verification of the hypotheses
- Drawing conclusions
