statistical test which returns a extensible tibble
Usage
stat_test(
df,
y,
x,
.by = NULL,
trans = "identity",
paired = FALSE,
paired_by = NULL,
alternative = "two.sided",
exclude_func = NULL,
method = "wilcoxon",
ns_symbol = "NS",
digits = 2
)Arguments
- df
tibble
- y
value
- x
sample test group
- .by
super-group
- trans
scale transformation
- paired
paired samples or not
- paired_by
a column for pair
- alternative
one of "two.sided" (default), "greater" or "less"
- exclude_func
a function has two arguments and return bool value, used if paired=TRUE and will keep the comparation pairs which return TRUE by this function.
- method
test method, 'wilcoxon' as default, one of
t|wilcoxon- ns_symbol
symbol of nonsignificant, 'NS' as default
- digits
significant figure digits of p value If the data pair of a single test returns TRUE, then exclude this pair
Examples
stat_test(mini_diamond, y = price, x = cut, .by = clarity)
#> # A tibble: 24 × 9
#> y clarity group1 group2 n1 n2 p plim psymbol
#> <chr> <chr> <chr> <chr> <int> <int> <chr> <dbl> <chr>
#> 1 price I1 Fair Good 5 5 0.31 1.01 NS
#> 2 price I1 Fair Ideal 5 4 0.90 1.01 NS
#> 3 price I1 Good Ideal 5 4 0.19 1.01 NS
#> 4 price IF Fair Good 4 5 0.063 1.01 NS
#> 5 price IF Fair Ideal 4 4 0.059 1.01 NS
#> 6 price IF Good Ideal 5 4 1.0 1.01 NS
#> 7 price SI1 Fair Good 5 4 1.0 1.01 NS
#> 8 price SI1 Fair Ideal 5 5 1.0 1.01 NS
#> 9 price SI1 Good Ideal 4 5 0.41 1.01 NS
#> 10 price SI2 Fair Good 4 4 0.057 1.01 NS
#> # ℹ 14 more rows