Table 1
Overall characteristics and comparison according to distractor format (n = 198).
| VARIABLE | OVERALL (n = 198) | 2 DISTRACTORS (n = 98) | 3 DISTRACTORS (n = 100) | p-VALUE |
|---|---|---|---|---|
| Nature of question | 0.383 | |||
| Factual, n (%) | 89 (44.9) | 41 (41.8) | 48 (48) | |
| Clinical case, n (%) | 109 (55.1) | 57 (58.2) | 52 (52) | |
| Low-functioning distractors (LFD) | ||||
| 0 LFD, n (%) | 56 (28.3) | 38 (38.8) | 18 (18) | 0.001 |
| ≥1 LFD, n (%) | 142 (71.7) | 60 (61.2) | 82 (82) | |
| Left unanswered >10%, n (%) | 113 (57) | 62 (63.3) | 51 (51) | 0.067 |
| Technical flaws | 0.045 | |||
| No flaws, n (%) | 105 (53) | 59 (60.2) | 46 (46) | |
| With flaws, n (%) | 93 (47) | 39 (39.8) | 54 (54) | |
| Difficulty index* | 0.68 (0.48–0.84) | 0.68 (0.35) | 0.68 (0.37) | 0.931 |
| Point-biserial** | 0.28 (0.12) | 0.29 (0.16) | 0.29 (0.15) | 0.708 |
| Preferred questions***, n (%) | 72 (36.4) | 38 (38.8) | 34 (34) | 0.485 |
[i] *Median (interquartile range).
**Mean (standard deviation).
***Difficulty index between 0.45 and 0.75, and a point-biserial > 0.20.
Table 2
Analysis of factors potentially associated with item difficulty (ID).
| VARIABLE | EASY ITEMS DI > 0.7 (n = 95) | NON-EASY ITEMS DI ≤ 0.7 (n = 103) | UNIV. p-VALUE | ADJUSTED OR (95% CI) | ADJ. p-VALUE |
|---|---|---|---|---|---|
| Number of distractors | 0.780 | — | — | ||
| 2 distractors, n (%) | 48 (50.5%) | 50 (48.5%) | |||
| 3 distractors, n (%) | 47 (49.5%) | 53 (51.5%) | |||
| Nature of the question | <0.001 | ||||
| Factual, n (%) | 56 (58.9%) | 33 (32.1%) | 2.62 (1.38–4.99) | 0.003 | |
| Clinical case, n (%) | 39 (41.1%) | 70 (67.9%) | Reference | ||
| Technical flaws | 0.109 | ||||
| No flaws, n (%) | 56 (58.9%) | 49 (47.6%) | Reference | ||
| With flaws, n (%) | 39 (41.1%) | 54 (52.4%) | 1.34 (0.71–2.56) | 0.369 | |
| Low-functioning distractors (LFD) | <0.001 | ||||
| ≥ 1 LFD, n (%) | 88 (92.6%) | 54 (52.4%) | Reference | ||
| 0 LFD, n (%) | 7 (7.4%) | 49 (47.6%) | 0.10 (0.04–0.24) | <0.001 | |
| Point-biserial (>0.30) | 0.212 | — | — | ||
| > 0.30, n (%) | 49 (51.6%) | 44 (42.7%) | |||
| ≤ 0.30, n (%) | 46 (48.4%) | 59 (57.3%) |
Table 3
Analysis of factors potentially associated with preferred questions.
| VARIABLE | COMPARISON | % PREFERRED QUESTIONS IN EACH CATEGORY | p (CHI2/FISHER) | ODDS RATIO (95% CI) | p (REGRESSION) |
|---|---|---|---|---|---|
| Presence of technical flaws | No flaws vs. ≥1 flaws | 31/105 (29.5%) vs. 41/93 (44.1%) | 0.034 | 0.57 (0.31–1.03) | 0.061 |
| Low-functioning distractors | 0 LFD vs. ≥1 LFD | 27/56 (48.2%) vs. 45/142 (31.7%) | 0.029 | 1.87 (0.99–3.56) | 0.055 |
| Number of distractors | 2 vs. 3 distractors | 38/98 (38.8%) vs. 34/100 (34%) | 0.485 | Not included | — |
| Nature of the question | Clinical case vs. Factual | 32/89 (35.9%) vs. 40/109 (36.7%) | 0.914 | Not included | — |
[i] Preferred questions: questions with difficulty index between 0.45 and 0.75 and point-biserial > 0.20. Binary logistic regression was performed using the Enter method. Variables with p < 0.20 in the univariate analysis were included in the model.