Internal-Consistency Reliability Quiz

Factors Affecting Internal Consistency Reliability

• The length of a test and its internal consistency reliability are positively related.
• Internal consistency is not affected by the time of administration, respondents' characteristics, or the test's content.

Estimating Reliability

• The most common method of estimating reliability is internal consistency reliability.
• Internal consistency is a better alternative to alternative forms and test-retest procedures.

Internal Consistency Reliability Procedure

• The internal consistency reliability procedure requires respondents to complete one form of a test on one occasion.

Spearman-Brown Formula

• The Spearman-Brown formula adjusts for the length of a test in the internal consistency reliability estimate.
• The formula is used to estimate the reliability of a test based on its current length and the desired length.
• The formula is not useful for data that is standardized (t-scores & z-scores).

Split-Half Reliability Approach

• The major advantage of the split-half reliability approach is that it can be used when there is no alternative form of the test.
• A limitation of the split-half reliability approach is that it may not be suitable for tests with a small number of items.

Cronbach's Alpha

• Cronbach's alpha is a measure of internal consistency reliability.
• It is important for an instrument to have a high level of covariance among its items to ensure reliability.
• Cronbach's alpha formula represents the average correlation among all possible split-half combinations of items.
• The numerator in Cronbach's alpha formula represents the sum of the covariances between items.
• The denominator in Cronbach's alpha formula represents the variance of the total score.
• The standardised coefficient alpha is used for raw scores and should have similar variances and be similar in magnitude to the coefficient alpha.
• At the theoretical level, Cronbach's alpha formula represents the proportion of the total variance that is due to the true score.

Limitations of Cronbach's Alpha

• The main limitation of Cronbach's alpha is that it can be affected by positively correlated error terms between two items within a test.
• If the assumption of essential tau-equivalence is not satisfied, then Cronbach’s alpha will tend to underestimate reliability.
• With a reasonable amount of items (10 or more), Cronbach’s alpha will underestimate reliability by 5-10%.
• A smaller number of items (say, 3 to 6 items) will cause the degree of the Cronbach's alpha underestimation to lessen.
• In practice, violating the assumption of non-correlated error terms can result in Cronbach’s alpha to overestimate reliability.

Kuder-Richardson Formula

• The Kuder-Richardson formula is of historical interest for estimating internal consistency reliability associated with composite scores based on dichotomously scored items.
• The formula is used for estimating internal consistency reliability for dichotomously scored items.

Adding Items to a Test

• The main factor to consider when deciding whether to add an item to a test is the item's correlation with the total score.
• The textbook's recommendation for adding items to a test is to add items that are highly correlated with the total score.
• The purpose of adding an item to a test is to increase its reliability.

Test your knowledge on internal-consistency reliability, the most common method of estimating reliability. Learn about its practical alternative to the alternative forms procedure and the test-retest procedure. Explore how the internal consistency reliability procedure treats different items within the same test as different forms.