Calculates Cohen's d for two sample comparisons.

effect_size_t(data, response, group, absolute = FALSE, paired = FALSE,
na.rm = TRUE)

## Arguments

data |
A `data.frame` . |

response |
The response variable (dependent). |

group |
The group variable, usually a `factor` . |

absolute |
If set to `TRUE` , the absolute effect size is returned. |

paired |
Whether the effect should be calculated for a paired
t-test, default is `FALSE` . |

na.rm |
If `TRUE` (default), missing values are dropped. |

## Value

`numeric`

of length 1.

## Details

The effect size here is Cohen's d as calculated by
\(d = \frac{m_{diff}}{S_p}\), where \(m_{diff} = \bar{x}_1 - \bar{x}_2\) and
\(S_p =
\sqrt{
\frac{n_1 - 1 \cdot {s_{x_1}}^2 + n_2 - 1 \cdot {s_{x_2}}^2}
{n_1 + n_2 - 2}
}
\).

For `paired = TRUE`

, \(S_p\) is substituted by \(S_D = S_{x_1 - x_2}\)
via `sd(x1 - x2)`

.

## Examples

set.seed(42)
df <- data.frame(x = runif(100), group = sample(c("A", "B"), 100, TRUE))
effect_size_t(df, "x", "group")

#> [1] -0.16946915