Skip to main content

Table 1 Specificity, sensitivity, positive and negative predictive value in each scenario of the simulation study using the cut-off value of 90%

From: Item pre-knowledge true prevalence in clinical anatomy - application of gated item response theory model

Proportion (True prevalence)5%5%10%10%35%35%70%70%
Cheating efficacyHighLowHighLowHighLowHighLow
 Mean (SD)Mean (SD)Mean (SD)Mean (SD)Mean (SD)Mean (SD)Mean (SD)Mean (SD)
Specificity (%)77.82 (2.74)78.16 (3.76)81.65 (3.28)81.46 (3.41)94.41 (3.02)92.33 (2.83)98.00 (1.80)97.65 (2.14)
Sensitivity (%)69.78 (14.52)60.30 (15.73)81.35 (7.88)69.35 (12.16)83.75 (15.78)83.35 (7.28)90.75 (2.48)68.99 (11.40)
Positive predictive value (%)14.84 (6.37)12.71 (3.48)34.07 (7.92)29.54 (5.57)88.42 (10.31)85.30 (4.50)99.07 (0.83)98.66 (1.20)
Negative predictive value (%)96.90 (8.59)96.66 (7.02)97.18 (2.83)96.00 (1.55)92.02 (6.77)89.88 (3.13)82.30 (4.13)58.31 (7.65)
Model absolute agreement (%)77.44 (2.81)76.97 (3.37)81.16 (5.56)80.16 (3.36)90.54 (4.17)88.21 (2.71)92.93 (1.75)77.59 (7.71)
Apparent Prevalence24.52 (2.7)24.07 (3.2)24.7 (2.9)23.7 (3.1)32.8 (7.2)33.2 (3.5)64.1 (1.9)49.0 (8.3)
Cohen’s Kappa (%)17.22 (6.86)15.07 (9.56)38.32 (6.96)31.64 (8.46)78.43 (10.68)73.76 (6.29)84.21 (3.71)56.26 (11.41)
Best Cut-off pointa99.9%100%98.6%98.8%86.8%91.4%79.3%65.9%
  1. aBest Cut-off point estimate by a classification tree