Dominique Makowski

> it is more important to have consistency within each of our methods across models, than to be consistent with methods from other packages absolutely Do you suggest going for...

``` r report::report_performance(lm(mpg ~ ., data = mtcars)) #> The model explains a significant and substantial proportion of variance (R2 = 0.87, F(10, 21) = 13.93, p < .001, adj....

that information is in the attributes returned by `model_performance` if I recall

> one idea might be to encourage the notion that these sorts of test statistics are model comparisons against a null model That's very true, it often gets understood as...

Makes sense. Here's the (big) list of checks ☺️ - [ ] check_autocorrelation - [x] check_clusterstructure - [ ] check_collinearity - [ ] check_convergence - [ ] check_distribution - [x]...

I think this will be a semi-breaking change (the printing will be ugly for people until they updated insight. The safest alternative would be to keep the forceful printing inside...

yup

Maybe my thought was based on the checks in parameters... 😬

> How is that information used/what is it interpreted to convey? Never used that, but from [here](https://www.investopedia.com/terms/c/coefficientofvariation.asp): > The coefficient of variation (CV) is a statistical measure of the relative...

> My preference is strongly not (3). why? having an explicit alias for `performance_cv` doesn't sound like a bad idea