pool_parameters df_error Infinity
ndsubison2178 opened this issue · 1 comments
ndsubison2178 commented
What is the reason for seeing values like "Inf" in column df_error ? when using pool_parameter function to pool results after fitting model on mice imputed dataset ? Thanks.
strengejacke commented
Thanks, this is actually a lack of documentation. See this page how parameters resp. model_parameters()
calculates DF. You can, however, pass down ci_method
to get same df's as for summary(pool())
, see example below. I have added this example to the docs.
library(parameters)
library(mice)
#>
#> Attaching package: 'mice'
#> The following object is masked from 'package:stats':
#>
#> filter
#> The following objects are masked from 'package:base':
#>
#> cbind, rbind
data("nhanes2")
nhanes2$hyp <- datawizard::slide(as.numeric(nhanes2$hyp))
imp <- mice(nhanes2, printFlag = FALSE)
# example for multiple imputed datasets
models <- lapply(1:5, function(i) {
glm(hyp ~ age + chl, family = binomial, data = complete(imp, action = i))
})
pool_parameters(models, ci_method = "residual")
#> # Fixed Effects
#>
#> Parameter | Log-Odds | SE | 95% CI | Statistic | df | p
#> ----------------------------------------------------------------------------------
#> (Intercept) | -24.25 | 3705.79 | [-7773.81, 7725.30] | -6.54e-03 | 19.25 | 0.995
#> age [40-59] | 19.15 | 3705.79 | [-7730.39, 7768.69] | 5.17e-03 | 19.25 | 0.996
#> age [60-99] | 19.20 | 3705.79 | [-7730.34, 7768.75] | 5.18e-03 | 19.25 | 0.996
#> chl | 0.02 | 0.03 | [ -0.06, 0.10] | 0.66 | 7.05 | 0.529
#>
#> Uncertainty intervals (equal-tailed) and p-values (two-tailed) computed
#> using a Wald distribution approximation.
m <- with(data = imp, exp = glm(hyp ~ age + chl, family = binomial))
summary(pool(m))
#> term estimate std.error statistic df p.value
#> 1 (Intercept) -24.25308549 3.705795e+03 -0.006544638 19.248074 0.9948455
#> 2 age40-59 19.14995832 3.705788e+03 0.005167580 19.248074 0.9959301
#> 3 age60-99 19.20461747 3.705789e+03 0.005182330 19.248074 0.9959185
#> 4 chl 0.02245509 3.393717e-02 0.661666585 7.047894 0.5292163
Created on 2024-01-04 with reprex v2.0.2