Testing Proportions
Lecture
R reference
Hypothesis testing functions
| Function | Test | Usage | Notes |
|---|---|---|---|
binom.test() |
Run an exact binomial test | binom.test(k, n, null_p |
Remember to specify the null probability of success as the final argument |
chisq.test() |
Run a goodness of fit test | chisq.test(<observed counts>, p = <null frequencies>) |
chisq.test(c(100, 150, 32), p = c(1/3, 1/3, 1/3)) |
chisq.test() |
Run a contingency table analysis | chisq.test(<contingency table>) |
cont.table <- rbind(c(45,54), c(67,88)) chisq.test(cont.table) |
fisher.test() |
Run Fisher’s exact test | fisher.test(<contingency table>) |
Usage is the same as for contingency chisq.test() |
2x2 contingency table calculations
-
Odds ratio = (or reciprocal)
-
Relative risk for Row1 over Row2 =