EffectSizes.jl is a Julia package for effect size measures. Confidence intervals are assigned to effect sizes using either the normal distribution or by bootstrap resampling. The package implements types for the following measures:
| Measure | Type |
|---|---|
| Cohen's d | CohenD |
| Hedge's g | HedgeG |
| Glass's Δ | GlassΔ |
] add https://github.com/harryscholes/EffectSizes.jljulia> using Random, EffectSizes; Random.seed!(1);
julia> xs = randn(10^3);
julia> ys = randn(10^3) .+ 0.5;
julia> es = CohenD(xs, ys) # normal CI
-0.507, 0.95CI (-0.924, -0.089)
julia> effectsize(es)
-0.506674937960848
julia> quantile(es)
0.95
julia> CohenD(xs, ys, quantile=0.99)
-0.507, 0.99CI (-1.056, 0.042)
julia> CohenD(xs, ys, 10^4) # bootstrap CI
-0.507, 0.95CI (-0.594, -0.417)
julia> ci = confint(es)
0.95CI (-0.924, -0.089)
julia> confint(ci)
(-0.9244427501651218, -0.08890712575657417)
julia> lower(ci)
-0.9244427501651218
julia> upper(ci)
-0.08890712575657417
julia> quantile(ci)
0.95Ideas and PRs are very welcome.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent