Results from the Generalisability analysis with relative contribution of variance
| Results from the G-study showing the contribution of each source of variance | |||
| Source of variance V | Description | Relative contribution | Interpretation of results |
| Participant, Vpa | Variation among participants | 77.0% | Most of the variance comes from the various skill levels among the participants |
| The procedures, Vp | Variability among the two procedures | 0.5% | Almost none of the variance comes from the procedures due to the standardised testing setup |
| Rater, Vr | Variability among the raters | 3.2% | A small contribution to the variance indicates a high level of agreement among the raters |
| Interaction between the participant and the procedures, Vpa*p | If the participants consistently show a difference in the two procedures | 2.8% | The small contribution to the variance from this interaction indicates a small learning by testing effect and a successful blinding |
| Interaction between the rater and the participant, Vpa*r | If a rater assesses a particular participant differently | 2.3% | A small variance contribution indicating a successful blinding |
| Interaction between the rater and the procedure, Vp*r | If a rater shows a consistent difference in rating the two procedures | 0.0% | No variance contribution indicates a successful blinding |
| Interaction between participant, rater and procedures, Vpa*r*p | The remaining variability | 14.2% | Unavoidable, unexplained error |