This vignette describes the mp_power() workflow
and how to summarize results with summary().
d <- mp_design(clusters = list(subject = 20), trials_per_cell = 4)
a <- mp_assumptions(
fixed_effects = list(`(Intercept)` = 0, condition = 0.4),
residual_sd = 1,
random_effects = list(subject = list(intercept_sd = 0.1))
)
scn <- mp_scenario_lme4(
y ~ condition + (1 | subject),
design = d,
assumptions = a,
test_method = "wald"
)
res <- mp_power(scn, nsim = 10, seed = 42)
#> boundary (singular) fit: see help('isSingular')
#> boundary (singular) fit: see help('isSingular')
#> boundary (singular) fit: see help('isSingular')
#> boundary (singular) fit: see help('isSingular')
#> boundary (singular) fit: see help('isSingular')
summary(res)
#> $power
#> [1] 0.3
#>
#> $mcse
#> [1] 0.1449138
#>
#> $ci
#> [1] 0.06673951 0.65245285
#>
#> $ci_method
#> [1] "clopper-pearson"
#>
#> $diagnostics
#> $diagnostics$fail_rate
#> [1] 0
#>
#> $diagnostics$singular_rate
#> [1] 0.5
#>
#> $diagnostics$type_s
#> [1] 0
#>
#> $diagnostics$type_m
#> [1] 1.562924
#>
#>
#> $nsim
#> [1] 10
#>
#> $alpha
#> [1] 0.05
#>
#> $failure_policy
#> [1] "count_as_nondetect"
#>
#> $conf_level
#> [1] 0.95