What is internal consistency reliability?

What is internal consistency reliability?

• A. A measure of whether similar questions on the same test are answered in the same way
• B. A measure of whether the test correlates with external criteria
• C. The agreement between raters
• D. The stability of test scores over time

Answer: (A)

Internal consistency reliability refers to how well the items on a test hang together to measure the same underlying construct. When a test truly targets one trait or ability, the items should behave in a coordinated way—people who endorse or score high on one item tend to do so on the others as well. Cronbach’s alpha is a common statistic used to quantify this, with higher values indicating that the items are consistently related to the same construct. This concept is different from how well a test correlates with external criteria (criterion validity), how closely different raters agree (inter-rater reliability), or how stable scores are over time (test-retest reliability). For example, a well-constructed mood inventory should show items about mood, energy, and sleep correlating with each other, reflecting a common underlying construct, rather than measuring unrelated things.

Internal consistency reliability refers to how well the items on a test hang together to measure the same underlying construct. When a test truly targets one trait or ability, the items should behave in a coordinated way—people who endorse or score high on one item tend to do so on the others as well. Cronbach’s alpha is a common statistic used to quantify this, with higher values indicating that the items are consistently related to the same construct. This concept is different from how well a test correlates with external criteria (criterion validity), how closely different raters agree (inter-rater reliability), or how stable scores are over time (test-retest reliability). For example, a well-constructed mood inventory should show items about mood, energy, and sleep correlating with each other, reflecting a common underlying construct, rather than measuring unrelated things.