TABLE 2

Results from the G-Study indicating the contribution of each source of variance

Source of varianceDescriptionEstimated VCRelative contribution %Interpretation of results
PhysiciansSystematic variation among physicians57.852.1Most of the variance derives  from different competence  levels between the physicians
AssessorsSystematic variability among assessors4.774.3The assessors had a high  degree of agreement
Interaction between physician  and assessorConsistent trend for a assessor to access a particular physician differently2.081.9There was no bias between  assessor and physician due to  successful blinding
Interaction between cases  and physiciansSystematic variability among cases12.3011.1Some difference in the  difficulty of the  EBUS-TBNA cases
Interaction between case,  assessor, and physicianAll remaining variability34.1130.7Expected unexplained error
  • VC: variance component; EBUS-TBNA: endobronchial ultrasound-guided transbronchial needle aspiration.