| Predictor* | Difficulty Index | Discrimination Index | Point biserial |
|---|
| Coefficient | p-value | Coefficient | p-value | Coefficient | p-value |
|---|
| Model 1$# | F(2,170) = .78 | .46 | F(2,170) = 6.84 | .001 | F(2,170) = 15.192 | <.001 |
| Case/Noncase based items | −.04 | .26 | .08 | .019 | .01 | .67 |
| Open-ended (essay)/MCQs items | −.01 | .89 | −.10 | .001 | −.14 | <.001 |
| Model 2@# | F(2,72) = .76 | .47 | F(2,72) = .40 | .67 | F(2,72) = .18 | .83 |
| Case/Noncase based items | −.02 | .77 | .04 | .39 | −.01 | .79 |
| Number of Choices | −.03 | .23 | .01 | .78 | .01 | .60 |
- *because the number of choices variable can’t be included in the same model as the item type variable, two separate models (one for item format and item type and then one for item format and number of choices among the MCQ questions) were analyzed
- $N = 173
- @N = 75
- #The interaction of the two factors in the model was tested and no significant effect was detected with also no effect on the significance on other factors
