This page is about:
GME can provide help and support for these projects.
One of the cornerstones of evaluation in medical education is the use of direct observation as a formative and summative measure of complex performances and competencies. To make direct observations an effective and consistent tool for providing feedback it is important to use well developed rating forms. Forms that function well are said to have a higher degree of validity and reliability. Let’s explore these two important concepts and how they play out in the world of observations.
Reliability is the degree of stability, repeatability or consistency a measure exhibits. For a rating scale, one of the key functions is the degree of agreement between multiple raters observing the same thing, known as Inter-rated reliability. This agreement is expressed as a correlation coefficient. For longer rating instruments, the effect if length can be measured using an internal consistency indicator like the split-half method.
Validity is an indicator to the extent to which an instrument is actually measuring what it is designed to measure. Face validity (aka content validity) employs a panel of experts to look over an instrument and is used as an initial indicator during development. For ratings scales, one of the standard reliability indicators is a correlation between ratings and an established external performance indicator, such as grades or board examinations. This type of “criterion referenced” validity is also expressed as decimal coefficient, where bigger is better.
Finally, consider the Differential Validity, which is the potential for members of certain groups to be measured differently. Internal consistency measures aggregated by race and ethnicity can be indicators of this threat in addtiion to careful analysis of item response patterns.
Graduate Medical Education
Program Director’s Forum
John Roden, PhD