Validity & Reliability in Testing

Study Notes

Validity and reliability are essential for assessing exam quality and usefulness.
They pertain to the scores produced by tests rather than the tests themselves.
Example of application: “SAT scores are a reliable and valid indicator of college performance.”

Reliability indicates the stability and clarity of exam results.
It assesses if scores are consistent or if variations occur by chance.
Measured through a correlation ranging from 0 (unreliable) to 1 (perfectly reliable).
Even a difficult test can be reliable if it maintains consistent ranking of student performance.

Test-Retest Reliability: Measures consistency over time by administering the same test at different points (e.g., January 1 and February 1).
- High reliability indicated by consistent rankings despite score changes.
Internal Consistency Reliability: Assesses consistency among items in a test using two metrics:
- Split-Half Reliability: Compares scores from two halves of a test.
- Cronbach’s Alpha: Measures average consistency across all possible split-halves; high coefficients indicate strong internal consistency.

Content Validity: Ensures test items appropriately represent the intended subject area by aligning items with learning objectives.
Criterion-Related Validity: Assesses how test scores relate to actual performance outcomes, can be further divided:
- Predictive Validity: Evaluates how well a test can predict future performance (e.g., graduation rates).
- Concurrent Validity: Compares test scores with established tests conducted at the same time to gauge correlation.
Construct Validity: Determines if a test accurately measures the theoretical construct it's designed to evaluate.

Without evidence of reliability and validity, test results lack meaning and usefulness in educational contexts.
Reliable and valid assessments are crucial for effective teaching and accurate evaluation of student performance.