Skip to main content
. 2024 Oct 24;19(10):e0312387. doi: 10.1371/journal.pone.0312387

Table 2. The effect of Free Practice (FP) on OSCE performance (PSM).

FP No FP Difference S.E. T-Stat. Common support
Physical examination
Unmatched 90.85% 85.02% 5.83% 0.965 6.05
ATT 90.85% 85.98% 4.87% 1.030 4.72 243/245
Neurological examination
Unmatched 85.56% 83.36% 2.20% 1.509 1.46
ATT 85.56% 83.73% 1.83% 1.524 1.20 245/245
Anamnesis
Unmatched 83.96% 83.38% 0.58% 1.554 0.38
ATT 83.96% 83.60% 0.36% 1.515 0.24 244/245
Diagnosis
Unmatched 84.71 84.37 0.34% 1.217 0.28
ATT 84.71 84.26 0.45% 1.221 0.38 245/245
Medical Skills
Venipuncture
Unmatched 94.48% 96.35% -1.87% 2.212 -0.85
ATT 93.76% 96.53% -2.77% 2.542 -1.09 57/86
ECG
Unmatched 92.46% 92.91% -0.45% 2.872 -0.16
ATT 91.2% 93.12% -1.92% 3.499 -0.55 42/48
PVC
Unmatched 94.86% 87.58% 7.28% 2.129 3.42
ATT 94.53% 84.55% 9.98% 4.966 2.01 109/135
Feeding tube
Unmatched 93.26% 89.91% 3.35% 2.285 1.47
ATT 92.00% 89.28% 2.72% 2.893 0.94 52/61
Rectal examination
Unmatched 93.45% 88.52% 4.93% 1.834 2.69
ATT 93.85% 89.30% 4.55% 2.281 1.99 36/45
Injections
Unmatched 97.06% 91.77% 5.29% 1.562 3.39
ATT 97.06% 90.76% 6.30% 2.015 3.13 84/84

Average treatment effects on the treated (ATT) obtained from Propensity Score Matching with the parameters gender, vocational training, school-leaving grade, the average exam grade in the M1-phase and a binary indicator taking the value = 1 if all exams preceding the OSCE had been passed. Kernel matching (Epanechnikov kernel) with replacement and a strict condition of common support. For quality of matching tests, see Table 3 and Supplement 3 in S1 File. For propensity score estimates, see Supplement 2. A T-Statistic > |1.96| corresponds to p<0.05. Common support denotes how many of the respective total respondents could be used for matching. “Unmatched” is a sample test for the groups Free Practice (FP) and No FP. S.E.: standard error.