Criterion-Oriented Validation: Crocker and Algina Method for Test Validity

Criterion-Oriented Validation According to Crocker and Algina

TMA 10: Crocker and Algina: Study D. Criterion-oriented validation is carried out through representative samples.

Steps for Criterion-Oriented Validation

1. Identify the appropriate criterion.
2. Identify the test and the criterion to be scored.
3. Administer the test.
4. Obtain the measure of the external criterion.
5. Determine the intensity of the relationship between the test scores and the criterion.

Typical Estimation Failure

Typical failure NO ESTIMATE: Allows major scores in the prognosis of statistics to be used through our knowledge of statistical inference.

External Criterion

External criterion: Interest

1. A variable or characteristic of real interest that the test aims to predict or infer.
2. The need to establish appropriate criteria.
3. Decision making: clinical diagnosis and assignment of personnel selection to treatment programs.

Some external criteria: Intelligence and aptitude tests (academic performance), clinical diagnostic tests, specific aptitudes, and tests on contrasted groups for validation.

Steps in Detail

1. Identify the relevant criterion and one method, sufficiently representative.
2. Identify the population sample to be used in the test.
3. Administer the criterion.
4. Score the test.
5. Obtain the measure of the relationship between the test and the criterion: correlation coefficient (Pearson, Spearman, etc.), called the validity coefficient.

Validity Coefficient

1. It varies between -1 and +1 (a value of 0 indicates absence of relationship).
2. It does not imply a causal relationship.
3. A change in the length of the test or criterion causes a change in the validity coefficient.
4. A change in the length of the test or criterion causes a change in validity.
5. A change in the variability of the test or criterion scores causes a change in the validity coefficient.

Factors Affecting Validity

Factors affecting validity: Reliability of the test and criterion, and the length of the test.

Attenuation Formula (Spearman)

Case I: Validity coefficient of the test when reliabilities of the criterion and test are perfect.

Case II: Validity coefficient when the reliability of the test is perfect.

Case III: Validity coefficient when the criterion is perfectly reliable.

Case IV: Validity coefficient when the reliability of the criterion is greater.

Case V: Validity coefficient of the test where the reliability is greater.

General Case IV: Validity coefficient when the reliability of the test and criterion is greater.

K: Number of times the original length of the test is increased.

Choice of Statistical Technique