broom::tidy.lm() -- how to set number of digits? - tidyverse

Question

I'm trying to use broom::tidy() to extract the ANOVA summary table for linear models, but for a better display, particularly for multivariate linear models.
I can't find a way to control the number of decimal digits that appear in the result for the sums of squares and statistics.
Here is an example, with a simple lm(): Note that I can get more digits in the output with print(aov1, digits=) [Sorry: reprex isn't working on my machine.]
> data("mtcars")
>
> mtcars$cyl <- factor(mtcars$cyl) # make factors
> mtcars$am <- factor(mtcars$am)
>
> aov1 <- anova(lm(mpg ~ cyl + am, data=mtcars))
> aov1
Analysis of Variance Table
Response: mpg
Df Sum Sq Mean Sq F value Pr(>F)
cyl 2 824.78 412.39 43.6566 2.477e-09 ***
am 1 36.77 36.77 3.8922 0.05846 .
Residuals 28 264.50 9.45
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>
> print(aov1, digits=10) # print with more digits
Analysis of Variance Table
Response: mpg
Df Sum Sq Mean Sq F value Pr(>F)
cyl 2 824.7845901 412.3922950 43.65661 2.4769e-09 ***
am 1 36.7669195 36.7669195 3.89221 0.058457 .
Residuals 28 264.4956779 9.4462742
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
But tidy() doesn't appear to have any way to control the number of digits that I can find...
> tidy(aov1)
# A tibble: 3 x 6
term df sumsq meansq statistic p.value
<chr> <int> <dbl> <dbl> <dbl> <dbl>
1 cyl 2 825. 412. 43.7 2.48e-9
2 am 1 36.8 36.8 3.89 5.85e-2
3 Residuals 28 264. 9.45 NA NA
>
> tidy(aov1, digits=7)
# A tibble: 3 x 6
term df sumsq meansq statistic p.value
<chr> <int> <dbl> <dbl> <dbl> <dbl>
1 cyl 2 825. 412. 43.7 2.48e-9
2 am 1 36.8 36.8 3.89 5.85e-2
3 Residuals 28 264. 9.45 NA NA
>
> print(tidy(aov1), digits=7)
# A tibble: 3 x 6
term df sumsq meansq statistic p.value
<chr> <int> <dbl> <dbl> <dbl> <dbl>
1 cyl 2 825. 412. 43.7 2.48e-9
2 am 1 36.8 36.8 3.89 5.85e-2
3 Residuals 28 264. 9.45 NA NA
My goal is actually more general: to extract the univariate tests for each response in a multivariate linear model.
> data(NeuroCog, package="heplots")
> NC.mlm <- lm(cbind( Speed, Attention, Memory, Verbal, Visual, ProbSolv) ~ Dx,
+ data=NeuroCog)
> car::Anova(NC.mlm)
Type II MANOVA Tests: Pillai test statistic
Df test stat approx F num Df den Df Pr(>F)
Dx 2 0.2992 6.8902 12 470 1.562e-11 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
I can do this by reshaping the data to long format and group_by(response) followed by some tidy processing, but I want more digits in the values for the SS and F statistics.
> #' Reshape from wide to long
> NC_long <- NeuroCog |>
+ select(-SocialCog, -Age, -Sex) |>
+ tidyr::gather(key = response, value = "value", Speed:ProbSolv)
>
> NC_long |>
+ mutate(response = factor(response, levels=unique(response))) |> # keep variable order
+ group_by(response) |>
+ do(tidy(anova(lm(value ~ Dx, .)))) |>
+ filter(term != "Residuals") |>
+ select(-term) |>
+ rename(F = statistic, df1 = df,
+ SS = sumsq, MS =meansq) |>
+ mutate(df2 = 239) |> # kludge: extract dfe from object?
+ relocate(df2, .after = df1) |>
+ mutate(signif = noquote(gtools::stars.pval(p.value))) |>
+ mutate(p.value = noquote(scales::pvalue(p.value)))
# A tibble: 6 x 8
# Groups: response [6]
response df1 df2 SS MS F p.value signif
<fct> <int> <dbl> <dbl> <dbl> <dbl> <noquote> <noquote>
1 Speed 2 239 8360. 4180. 37.1 <0.001 ***
2 Attention 2 239 5579. 2790. 17.4 <0.001 ***
3 Memory 2 239 3764. 1882. 13.9 <0.001 ***
4 Verbal 2 239 4672. 2336. 27.3 <0.001 ***
5 Visual 2 239 3692. 1846. 16.6 <0.001 ***
6 ProbSolv 2 239 4165. 2083. 25.1 <0.001 ***