Reliability

Reliability

 

Reliability is used as a technical statistical term that refers to a measurement instrument’s ability to differentiate between subjects, patients, or participants in some underlying trait. Reliability goes up as the variability between subjects increases, and down as the variability within subjects (over time, or over raters) increases. Reliability is typically expressed as an intra-class correlation coefficient with between-subject variability in the numerator and total variability (between-subject and within-subject) in the denominator.

« Back to Glossary Index