Fig 1.
Distribution of directly reported p values (restricted to .001–.15).
Fig 2.
Distribution of directly reported p values (restricted to .001–.15, equal weight of each paper).
Table 1.
Distribution of number of “just significant” (.045–.050] and “almost significant” (.050–.055) directly reported p values per paper.
Fig 3.
Distribution of re-calculated actual p values (restricted to .001–.15).
Fig 4.
Distribution of re-calculated actual p values (restricted to .001–.15, equal weight of each paper).
Fig 5.
Frequency of reporting directly over intervals of actual p value (rounded up).
Fig 6.
Frequency of use of one-sided tests given actual one-sided p values (printed above each bar rather than on the axis and with percent sign omitted for greater readability, rounded up).
Table 2.
Actual rounding in roundable downs and roundable ups.
Table 3.
Rounded and not rounded p values in the (.02-.03) interval.
Fig 7.
Fraction of mistakes leading to the reported p value being lower rather than higher than the actual p value, over the calculated p value (rounded up, percent sign omitted).
Table 4.
Distribution of number of mistakes per paper.