- Fixed issue is
esc_mean_sd()
when arguments contained missing values.
- Deprecated functions have been removed.
- Update website (https://strengejacke.github.io/esc).
- Reduce package dependencies.
- Function that convert from one effect size into another are now prefixed with
convert_*()
, instead ofesc_*()
(e.g.convert_d2r()
instead ofesc_d2r()
).
esc_mean_sd()
,esc_mean_se()
andesc_mean_gain()
can now calculate effect-sizes for within-subject-designs, when argumentr
is given.
- Fix issue with wrong computation of confidence intervals when converting effect sizes from d to f (
esc_d2f()
).
- Functions now also return effect sizes Cohen's f or eta squared.
- More functions to convert effect size into other effect size measures (
cohens_f()
,odds_ratio()
,log_odds()
,pearsons_r()
,eta_squared()
andcohens_d()
).
- In rare cases,
esc_mean_sd()
tried to calculate the square root of negative values when computing the pooled standard deviance. In such cases, an alternative formular for the pooled SD is used.
combine_esc()
did not work for effect sizes computed withesc_t()
, when the returned effect size wasr
.
effect_sizes()
did not always properly extract study names and used numeric indices instead character values.
effect_sizes()
to generate a data frame with results of multiple effect size computations, based on data from another data frame.
- Fix control statements with condition with greater than one, which currently give a warning, but in future R versions may throw an error.
write_esc()
as a convenient wrapper to write results to an Excel csv-file.