Calculates the probability of winning (winP). In the referenced article Zou et al (2022) proposes a method for calculating probability of winning with a confidence interval an p-value testing.
Usage
win_prob(
data,
response = NULL,
group = NULL,
alpha = 0.05,
beta = 0.2,
group.ratio = 1,
sample.size = FALSE,
print.tables = FALSE,
dec = 3
)Arguments
- data
A data frame containing the response and group variable.
- response
The name of the response variable. Takes first column if empty.
- group
The name of the group variable. Takes second column if empty.
- alpha
The alpha level for the hypothesis test. Default is 0.05.
- beta
The beta level for the sample size calculation. Default is 0.2.
- group.ratio
The ratio of group sizes. Default is 1.
- sample.size
Flag to include sample size calculation. Default is FALSE.
- print.tables
Flag to print cumulative tables. Default is FALSE.
- dec
Numeric for decimals to print. Default is 3.
Examples
win_prob(data=stRoke::talos,response="mrs_6",group="rtreat")
#> Zou et al's winP (doi: 10.1161/STROKEAHA.121.037744)
#>
#> Probability of a random observation in Placebo group
#> will have a higher response score than a random
#> observation in Active group:
#>
#> winP: 0.400 (0.612, 0.372) p=0.0125
#> --------------------------------------------
#>
#> The numbers needed to treat (NNT) are: -9
#>
#>
#>