McDonald's omega (ω)

Like Cronbach's alpha, omega measures the reliability of a scale. It uses the estimated strength with which each question contributes to the common construct (the loading), whereas alpha assumes all questions contribute equally. Omega lies between 0 and 1; the same rules of thumb as for alpha apply (from .70 acceptable, from .80 good).

Example: Five questions about concentration: α = .76 and ω = .80. The questions contribute to different degrees, so the scale is more reliable than alpha suggests.

Caution: Omega assumes the questions measure a single thing. If they measure several, omega cannot be interpreted meaningfully; a factor analysis helps then. Reverse-worded questions must be reverse-coded first.