Model_parameter function, the standardization of coefficients

[hiswy]

Hello,

While I was reading the vignette on the standardization of model coefficients,
I was wondering if the method = 'refit' standardize the outcome variable as well. So the results would be 'total standardization' rather than 'partial' where predictors would be standardized only...

[mattansb]

That depends on the type of model and the values of the include_response= argument. GLM(M)s do not have their responses standardized. You can find more info in standardize(<default>) and standardize_parameters().

When the response of unstandardized, you will get a footnote in the printout of standardize_parameters().

For example:

model <- glm(am ~ cyl * mpg, data = mtcars, family = "binomial")

parameters::standardize_parameters(model, method = "refit")
#> # Standardization method: refit
#> 
#> Parameter   | Std. Coef. |        95% CI
#> ----------------------------------------
#> (Intercept) |      -0.58 | [-1.98, 0.70]
#> cyl         |       0.25 | [-1.54, 2.10]
#> mpg         |       2.10 | [-0.19, 5.28]
#> cyl × mpg   |      -0.36 | [-2.57, 1.54]
#> 
#> - Response is unstandardized.

[strengejacke]

In model_parameters(), the response is standardized by default, if applicable. You can set include_response = FALSE to avoid this. I added a bullet point to the docs mentioning this, and a link to this docs for further details.