What are the three approaches to estimating reliability?
1) test-retest, 2) alternate or parallel forms, and 3) internal or inter-item consistency.
What factors determine the method or methods employed to estimate reliability?
The purpose of obtaining a measure of reliability.
How high should the coefficient of reliability be?
It depends on the purpose and importance of the decisions made based on the test scores.
What does a reliability coefficient in the .90s rate?
A grade of A, with a value of .95 or higher for the most important types of decisions.
What does a reliability coefficient in the .80s rate?
A grade of B, with below .85 being a clear B-.
What range of reliability coefficients rates a weak, 'barely passing' grade?
From .65 through the .70s.
What does a coefficient of inter-rater reliability provide information about?
Error as a result of test scoring and how consistently two scorers score the same test items.
What are the different kinds of error variance reflected in different reliability coefficients?
Error due to test construction, administration, unidentified error, and scorer error.
What is the purpose of determining reliability in terms of breaking down error variance?
To understand the sources of error in a test, such as test construction, administration, and scoring.
What are the considerations related to the nature of the test when estimating reliability?
Homogeneity versus heterogeneity of test items, dynamic versus static characteristics, restriction or inflation of range, and speed tests versus power tests.
What are some special problems in measuring reliability for tests designed for infants?
The fast-changing abilities of young children and the need to account for developmental changes between testings.
What is the difference between a homogeneous test and a heterogeneous test?
A homogeneous test is functionally uniform throughout, while a heterogeneous test contains diverse items.