A statistic for estimating a test's reliability that depends largely on how people vary on individual items relative to overall test variance; a low alpha indicates unreliability.

A statistic for estimating a test's reliability that depends largely on how people vary on individual items relative to overall test variance; a low alpha indicates unreliability.

• A. Cronbach's alpha
• B. External reliability
• C. Factor analysis
• D. Likert-type scale

Answer: (A)

Internal consistency reliability is about how well the items on a test hang together as a single measure. Cronbach's alpha is the standard statistic for this because it directly estimates how much of the total score variance comes from a common underlying factor versus item-specific noise, by looking at the average inter-item covariances relative to the total variance. When respondents’ answers vary on individual items in a way that doesn’t align with the overall pattern, those item correlations weaken, alpha drops, and the test appears unreliable. The strength of this measure is that it captures the idea that a coherent, single construct should produce consistent responses across its items; a low alpha signals that the items aren’t jointly measuring that construct reliably. External reliability refers to stability over time, factor analysis examines the underlying structure of the test, and a Likert-type scale is a type of response format rather than a reliability index.

Internal consistency reliability is about how well the items on a test hang together as a single measure. Cronbach's alpha is the standard statistic for this because it directly estimates how much of the total score variance comes from a common underlying factor versus item-specific noise, by looking at the average inter-item covariances relative to the total variance. When respondents’ answers vary on individual items in a way that doesn’t align with the overall pattern, those item correlations weaken, alpha drops, and the test appears unreliable. The strength of this measure is that it captures the idea that a coherent, single construct should produce consistent responses across its items; a low alpha signals that the items aren’t jointly measuring that construct reliably. External reliability refers to stability over time, factor analysis examines the underlying structure of the test, and a Likert-type scale is a type of response format rather than a reliability index.