Skip to main content

Table 2 Statistical results of studies included in the analysis

From: Multiple testing in orthopedic literature: a common problem?

 

Journal A (N = 72)

Journal B (N = 55)

Total number of p-values presented

2555

1783

Number of p-values originating from multiple testing

1531

1046

Number of p-values ≤ = 0.05

563 (37%)

377 (36%)

Median number of p-values per publication#

15 (8–33)

15 (8–21)

Median risk of obtaining at least one significant result€

54% (34–81)

54% (34–66)

  1. #Values presented as median with interquartile range between brackets.
  2. €Median risk and interquartile range expressed in percentages. Calculated for uncorrected studies using the formula for Family Wise Error Rate mentioned in the text, assuming independence and conditional on that the null hypotheses are true.