HANDOUT 5.1:
WHAT DIFFERENT RELIABILITY COEFFICIENTS ASSESS

RELIABILITY

COEFFICIENT

PRIMARY SOURCES OF UNRELIABILITY (RANDOM ERROR)
RELATIONSHIP OF RELIABILITY COEFFICIENT TO MEASURE'S TRUE RELIABILITY
Alternate Forms
Parallel Forms
Delayed Equivalence

  • Participant,
  • Testing Conditions,
  • Scoring,
  • Items

Fairly accurate estimate. However, because forms are not identical, it may slightly underestimate measure's reliability.
Test-Retest
  • Participant,
  • Testing Conditions,
  • Scoring

Fair estimate.However, if participants remember test items when taking retest, this coefficient may provide an inflated estimate of the measure's reliability.
Kuder Richardson,
Cronbach's Alpha

  • Items

 Inflated estimate because many sources of unreliability are not measured.
Split Half
  • Items

 Inflated estimate because many sources of unreliability are not measured.
Inter-Observer,
Inter-Rater,
Conspect,
Scorer

  • Scoring

 Highly inflated estimate because many sources of unreliability are not measured.

To Handout 5.2 (Indexes of Internal Consistency)

Back to Chapter 5 Main Menu