You need to learn how to use prop.test properly.
Is this a ONE-sample test or TWO-sample test?
Perhaps this will help:
#"hand" calculation
phat <- 58/691
p <- 56/691
q <- 1-p
n <- 691
z <- (phat-p)/sqrt(p*q/n)
z
[1] 0.2787970
pnorm(-z)*2
[1] 0.7804006
#"prop.test"
prop.test(x=58, n=691, p=56/691, correct = FALSE)
1-sample proportions test without continuity correction
data: 58 out of 691, null probability 56/691
X-squared = 0.0777, df = 1, p-value = 0.7804
alternative hypothesis: true p is not equal to 0.08104197
# Note the 'without' continuity correction part
-Peter Ehlers
Roslina Zakaria wrote:
Hi r-users,
�
I would like to use prop.test code and I also calculate manually to test the
proportions for 2 groups.� The problem is the answer for the p-value�calculated
manually are different from prop.test.� Here are the results:
�
## Manually
�
z value: z= (phat-p)/sqrt(pq/n) = (.084-.081)/sqrt(.081(1-.081)/691)=0.289,
pvalue=0.7718
�
## Using prop.test code
low <- c(56,58)
tot <- c(691,691)
prop.test(low, tot, p = c(56/691,58/691), alternative = "two.sided", conf.level
= 0.95, correct = TRUE)
������� 2-sample test for given proportions with continuity correction
data:� low out of tot, null probabilities c(56/691, 58/691)
X-squared = 0.0096, df = 2, p-value = 0.9952
alternative hypothesis: two.sided
null values:
��� prop 1���� prop 2
0.08104197 0.08393632
sample estimates:
��� prop 1���� prop 2
0.08104197 0.08393632
�
�
Thank you so much for any help given.
[[alternative HTML version deleted]]
------------------------------------------------------------------------
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
--
Peter Ehlers
University of Calgary
403.202.3921
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
