I have a data frame with count numbers and I want to perform a chisq.test for each value of the variable Cluster. So basically, I need 4 contingency tables (for "A","B","C","D") where rows = Category, columns = Drug, value = Total. And subsequently a chisq.test should be run for all 4 tabels.
Example data frame
df <- data.frame(Cluster = c(rep("A",8),rep("B",8),rep("C",8),rep("D",8)),
Category = rep(c(rep("0-1",2),rep("2-4",2),rep("5-12",2),rep(">12",2)),2),
Drug = rep(c("drug X","drug Y"),16),
Total = as.numeric(sample(20:200,32,replace=TRUE)))
Firstly, use xtabs() to produce stratified contingency tables.
tab <- xtabs(Total ~ Category + Drug + Cluster, df)
tab
# , , Cluster = A
#
# Drug
# Category drug X drug Y
# >12 92 75
# 0-1 33 146
# 2-4 193 95
# 5-12 76 195
#
# etc.
Then use apply() to conduct a Pearson's Chi-squared test over each stratum.
apply(tab, 3, chisq.test)
# $A
#
# Pearson's Chi-squared test
#
# data: array(newX[, i], d.call, dn.call)
# X-squared = 145.98, df = 3, p-value < 2.2e-16
#
# etc.
Furthermore, you can perform a Cochran-Mantel-Haenszel chi-squared test for conditional independence.
mantelhaen.test(tab)
# Cochran-Mantel-Haenszel test
#
# data: tab
# Cochran-Mantel-Haenszel M^2 = 59.587, df = 3, p-value = 7.204e-13