I am going to determine the confidence interval for the proportion based on dataset "CES11" dataset which is included in R. The variable that I have used for the test is Abortion and the values are "No" and "Yes". I want to find the confidence interval for the proportion of "Yes", but R does a hypothesis test and provides a confidence interval for the first level which is "No". How can I change it such that results corresponds to level equal to "Yes"?
Frequency counts (test is for first level):
abortion
No Yes
1818 413
1-sample proportions test without continuity correction
data: rbind(.Table), null probability 0.5
X-squared = 884.82, df = 1, p-value < 2.2e-16
alternative hypothesis: true p is not equal to 0.5
95 percent confidence interval:
0.7982282 0.8304517
sample estimates:
p
0.8148812
CodePudding user response:
A quick way in base R is to reverse the tabulation vector order:
cts_abortion <- table(carData::CES11$abortion)
## p is for rate of "No"
prop.test(cts_abortion)
## 1-sample proportions test with continuity correction
##
## data: cts_abortion, null probability 0.5
## X-squared = 883.56, df = 1, p-value < 2.2e-16
## alternative hypothesis: true p is not equal to 0.5
## 95 percent confidence interval:
## 0.7979970 0.8306679
## sample estimates:
## p
## 0.8148812
## p is for rate of "Yes"
prop.test(rev(cts_abortion))
## 1-sample proportions test with continuity correction
##
## data: rev(cts_abortion), null probability 0.5
## X-squared = 883.56, df = 1, p-value < 2.2e-16
## alternative hypothesis: true p is not equal to 0.5
## 95 percent confidence interval:
## 0.1693321 0.2020030
## sample estimates:
## p
## 0.1851188
Otherwise, you can likely just relevel the factor:
df_ces11 <- carData::CES11
df_ces11$abortion <- factor(df_ces11$abortion, levels=c("Yes", "No"))