Converts the statistics reported in a paper (t, F, chi-square, r, Z) into effect sizes without the raw data: Cohen's d, Hedges' g (J correction), r, partial eta squared, an approximate odds ratio (log OR = d times pi over the square root of 3) and the 95% confidence interval of d, with the conversion chain visible and the APA 7 or Vancouver sentence ready to paste (formulas from Lakens, 2013, and Rosenthal, 1991). The parser tolerates spacing, capitalization, lost italics and European decimal commas.
Worked example: t(48) = 2.31 equals d = 0.67, 95% CI [0.10, 1.24], g = 0.66, r = .32 and an approximate OR of 3.35, with an N of 50 implied by the degrees of freedom; F(1, 120) = 9.4 equals d = 0.56 and partial eta squared = .07; chi-square(1, N = 200) = 8.66 gives phi = .21 and d = 0.43; r = .41 equals d = 0.90.
Does it work if the paper reports no means or SDs? Yes: the test statistic and its degrees of freedom are enough, and they also imply the sample size. What about an F with df1 greater than 1? It does not map onto a single d: it is an omnibus test, so the tool returns partial eta squared and explains why.
Paste t(48) = 2.31 or F(1, 120) = 9.4 and get Cohen's d, Hedges' g, r, partial eta², approximate OR and 95% CIs, with the APA 7 sentence ready to paste. Free.