Should not use *** for self-correlations R(x,x)=1

[friendly]

I'm not a fan of significance stars, but certainly in text output and plots, you should not use *** for the self-correlations r(x,x)=1

[IndrajeetPatil]

Thanks for your feedback!

Yes, I agree that this should not be the case.

For plots:

or for text outputs:

## # Correlation Matrix (pearson-method)
## 
## Parameter    | Sepal.Length | Sepal.Width | Petal.Length | Petal.Width
## ----------------------------------------------------------------------
## Sepal.Length |      1.00*** |       -0.12 |      0.87*** |     0.82***
## Sepal.Width  |        -0.12 |     1.00*** |     -0.43*** |    -0.37***
## Petal.Length |      0.87*** |    -0.43*** |      1.00*** |     0.96***
## Petal.Width  |      0.82*** |    -0.37*** |      0.96*** |     1.00***
## 
## p-value adjustment method: Holm (1979)

[bwiernik]

Agreed. Should also change coloring of the diagonal to not be solid and dark.

What do folks think about changing the defaults to omit stars?

[strengejacke]

Diagonal could be solid, or maybe grey? Agree on omitting stars.

[IndrajeetPatil]

+1 for omitting stars by default.

[friendly]

YES to (light) grey. Try this to see what a difference it makes to perception.

[strengejacke]

library(correlation)
x <- correlation(iris)
s <- summary(x, redundant = TRUE)
plot(s)

^{Created on 2022-08-19 by the reprex package (v2.0.1)}

Plot ok like this, or any suggestions for changes?

[strengejacke]

Same for print now:

library(correlation)
x <- correlation(iris)
summary(x, redundant = TRUE)
#> # Correlation Matrix (pearson-method)
#> 
#> Parameter    | Sepal.Length | Sepal.Width | Petal.Length | Petal.Width
#> ----------------------------------------------------------------------
#> Sepal.Length |              |       -0.12 |      0.87*** |     0.82***
#> Sepal.Width  |        -0.12 |             |     -0.43*** |    -0.37***
#> Petal.Length |      0.87*** |    -0.43*** |              |     0.96***
#> Petal.Width  |      0.82*** |    -0.37*** |      0.96*** |            
#> 
#> p-value adjustment method: Holm (1979)

^{Created on 2022-08-19 by the reprex package (v2.0.1)}

[bwiernik]

Looks good. Can you switch the default to not show stars?

[DominiqueMakowski]

isn't showing stars a feature that people are looking for? (and thus could help for the visibility of the package, then it' easy to turn them off)

[bwiernik]

IIRC correctly, plot.summary() and print.summary() don't have any argument to disable stars or to show intervals. I personally think they should be non-default