effects_table() returns the effect table of an outcome_model() fit:
one row per term, with its estimate, confidence interval and corrected
p-value, rounded for reading. Arguments keep the intercept or only the
significant terms.
Arguments
- x
An
outcome_model()result; anything else raises an error.- intercept
Whether to keep the intercept row. Default
FALSE.- significant
Whether to keep only the terms whose corrected p-value (
p_adj) is belowalpha. DefaultFALSE.- alpha
The threshold for
significant, between 0 and 1. Default0.05.- digits
The decimal places of the numeric columns. Default
3;p_valueandp_adjare never rounded.
Value
A data frame with the columns of the model's effect table (term,
estimate, std_error, statistic, ci_lower, ci_upper, p_value,
p_adj, and for a binomial fit odds_ratio, or_lower and
or_upper), one row per retained term. The intercept row is dropped
unless intercept = TRUE, and every term is kept unless
significant = TRUE restricts them to p_adj < alpha.
Examples
set.seed(1)
d <- data.frame(hint = rbinom(200, 1, 0.5))
d$success <- rbinom(200, 1, plogis(-0.3 + 0.9 * d$hint))
effects_table(outcome_model(d, outcome = "success", predictors = "hint"))
#> term estimate std_error statistic ci_lower ci_upper p_value p_adj
#> 1 hint 1.522 0.305 4.994 0.925 2.12 5.910034e-07 5.910034e-07
#> odds_ratio or_lower or_upper
#> 1 4.583 2.522 8.33