Caperos and Pardo reviewed the results presented in 186 articles published in four Spanish psychology journals. Of these 102 contained at least one of Fisher’s F-test, Student’s t-test or Pearson’s chi-squared tests.
Of 1212 reviewed statistical results, 12.2% had consistency errors, i.e. the p-value was inconsistent with the reported value of the statistic and the degrees of freedom. For about 2.3% of the tests, the inconsistency had consequences for the conclusion; a correct p-value would have lead to another conclusion than the reported one.
Consistency errors were common. One or more such errors were included in 48% of the articles, and for 17.6% of them the correct p-value changes at least one conclusion.
The authors conclude that the quality and precision of presenting statistical results need to be improved in Spanish psychology journals.
1. Caperos JM, Pardo A. Consistency errors in p-values reported in Spanish psychology journals. Psicothema. 2013;25:408-414.