Reliability: Consistency or Differentiating Among Subjects?

Because reliability is a prerequisite for validity, it is of interest to physical therapists. But what is reliability? Mitchell states, “A reliable instrument is one with small errors of measurement, one that shows stability, consistency, and dependability of scores for individuals on the trait, characteristic or behavior being assessed.”1(p377) This view is concordant wih the opinion offered by Rothstein: “Reliability is basically the consistency of a measurement.”2(p5) There is, however, a second definition of reliability, that offered by classical test theory, that is somewhat different from the statements expressed above. This theory depicts reliability as being a ratio of variances; specifically, it is the variance attributed to differences among subjects divided by the total variance.3 A recurring dilemma for readers of physical therapy journals is that authors and editors have frequently contributed to publications that have not differentiated clearly between consistency and the ability to differentiate among subjects—both known as reliability.