Table 5 Reasons for inappropriate application of statistical tests found in 54 articles published in the Journal of the Egyptian National Cancer Institute, in the period from 2011 to 2016
Error | Number (%) |
|---|---|
Single misapplied test | 52 (96.3) |
 Small expected number for Chi-square test | 30 (55.6) |
 Univariate logistic or cox regression | 4 (7.4) |
 Compare survival between respondent and non-respondents | 5 (9.3) |
 Calculate diagnostic accuracy with no or inaccurate gold standard | 3 (5.6) |
 Combine different groups in analysis | 3 (5.6) |
 Include unknown as a category in analysis | 2 (3.7) |
 T test, ANOVA, or correlation for categorical or non-parametric data | 5 (9.3) |
Multiple misapplied tests | 2 (3.7) |