Notes
We still prefer to use the term “OSCE” to refer to a test administration format in which examinees rotate through a series of stations, some of which may involve standardized patients, but it's clear that particular terminology battle was lost years ago.
Disconnected designs occur when subgroups of examinees take test forms composed of non-overlapping sets of cases, SPs, and/or raters. See Swanson et al. for more detailed discussion of disconnected designs and approaches to connecting them.Citation 7
This will be true whether assignment of examinees, SPs, and raters to sessions is random or systematic (e.g., raters observe at the clinical sites at which they normally work): The resulting variance components will not be readily interpretable if the underlying session structure is ignored in the analysis.