Previous Chapter Chapter 54: Measurement Reliability Next Chapter
Chapter 54: Measurement Reliability
Most researchers consider a measure's reliability to be its freedom from random error, the extent to which a measure is indicative of some particular variable rather than factors unrelated to that variable. Stated differently, a more reliable measure is a purer indicator of the variable in question, less contaminated by irrelevant information. A more reliable measure should yield more consistent results across situations in which the attribute being measured has not changed. More formally, reliability is defined as the proportion of variation in an indicator that can be attributed to an underlying variable.
In the case of multi-item indicators developed under classical test theory (CTT), reliability is the proportion of variation in the composite score that is attributable to a ...