Test Validity Overview

Test Validity

• Validity is an important aspect of testing, focusing on whether a test accurately measures what it intends to measure.

• Messick (1995) defined validity as the degree to which evidence and rationale support interpretations and actions based on test scores.

• Content Validity is evaluated during test development, ensuring the test measures both the subject matter and instructional objectives.

  • This type of validity is crucial for achievement tests.
  • Experts assess the fit of each test item to its respective domain using a Likert-type rating scale.

• Criterion-Related Validity focuses on the correlation between test scores and an external criterion.

  • This criterion can be another test or a standard measure.
  • The correlation coefficient (validity coefficient) indicates the strength of the relationship.

Types of Criterion-Related Validity

• Concurrent Validity measures the correlation between scores on a test and an already established, valid test measuring the same construct.

  • Tests are administered simultaneously.

• Predictive Validity assesses a test's ability to predict future performance on a similar criterion.

  • Entrance examinations aim to predict academic success are an example of this type of validity.

Face Validity

• A test exhibits Face Validity if it appears valid on its surface.
  • It's primarily based on the expertise of the assessor concerning the subject matter.
  • Face validity is not statistically computed but is often evaluated through observation of test items.

Importance of Test Validity

• Test validity ensures that tests measure what they are designed to measure.
• It contributes to the attainment of test goals.
• It supports the accurate evaluation of individuals' abilities and potential based on test results.

Factors Affecting Test Validity

• Poor Sentence Structuring and Inappropriate Vocabulary can hinder comprehension.
• Poor Item Construction, Ambiguous Statements, Misleading Questions, and Unclear Instructions all contribute to reduced validity.
• Inclusion of Items Measuring Different Behaviours from those specifically targeted by the test diminishes validity.

Reliability

• A reliable test consistently measures a person's performance across multiple administrations.

Measuring Reliability

• Equivalent Form Method: Two versions of a test are administered to the same group, and the scores are correlated to assess consistency across the versions.
  • If the correlation is high, it indicates reliability between the two forms.
  • Limitations include the time and cost of developing two similar tests, and the availability of an equivalent test.

Split-half Method

• The test is divided into two halves, and scores on each half are correlated.
  • This correlation coefficient reflects the split-half reliability.
  • The Spearman-Brown formula is used to estimate the reliability of the full test.
  • Limitations arise if the halves are not properly split, making the correlation unreliable, or if the test measures different constructs.

Kuder-Richardson (K-R) Method

• The K-R method uses the full test and is suitable for frequency scores.
  • It avoids the issue of splitting items into halves and comprises two techniques: KR-20 and KR-21.