use_prop_test.Rd

% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/z_old_statistic.R
\name{use_prop_test}
\alias{use_prop_test}
\title{Test of Proportion}
\usage{
use_prop_test(df_count, eval_first_column = FALSE, limit = NULL)
}
\arguments{
\item{df_count}{A data frame, made by \link{count.column}}

\item{eval_first_column}{Boolean, should the first column be evaluated as well?}

\item{limit}{Integer or numeric - Range of factor levels or minimum count of factor level| e.g. 1:3 or 50}
}
\value{
Returns a vector of p.values obtained by the prop.test
}
\description{
This function applies prop.test on a data frame. When using \link{count.column} the first column represents the first grouping factor. Applying the prop.test on the whole data frame would lead to an error. Therefore, eval_first_column is set to FALSE. Throws warnings if the group size is too small (n<=5).
}
\details{
In the data frame made by count_column every level of the grouped factor is represented. Because the factors should be ordered, the first row corresponds to the level with the most counts.
When you use the prop.test you might want to only look at the 3 most occuring factor levels. Thats why you can set the limit to 1:3 to test only the 3 most occuring levels. When you only want to test factor levels with more than 50 counts
then you set limit to 50.
}
\examples{
df <- data.frame(case_id = c(1,2,3,4,5,6),
                 group = c("one","one","two","one","two","three"),
                 primary_diagnosis = c("Squamous cell carcinoma, NOS", "Adenocarcinoma, NOS", "Adenocarcinoma with mixed subtypes",
                                       "Squamous cell carcinoma, NOS", "Squamous cell carcinoma, NOS", "Squamous cell carcinoma, NOS"))
diagnosis_counts <- count_column(df, factor_one = "group", factor_two = "primary_diagnosis")
pValues <- use_prop_test(diagnosis_counts, FALSE)
}