Issue with Understanding Confidence Interval

General
#1

Hello. I need a little hand holding and direction. Just carried out a t.test and ]got the following result in console:slight_smile:

Welch Two Sample t-test

data: Treatment.group and Control.group
t = 2.4139, df = 27.89, p-value = 0.0226
alternative hypothesis: true difference in means is not equal to 0
95 percent confidence interval:
0.05626425 0.68773575
sample estimates:
mean of x mean of y
3.211333 2.839333

I am happy with the p-value and understand what it is saying. I am now looking to report the 95% confidence interval for this test. This is where I need help. Can I calculate this from this result or do I need to run another test? The data of interest is the weight of new born babies in kilogrammes.

Thanks very much if you can assist.

P

#2

Hi portnoy58, welcome to RStudio Community.

t.test() reports the 95% confidence interval for the mean. From your output, the lower confidence bound is 0.05626425 while the upper confidence bound is 0.68773575.

#5

I read the previous posts firsts, and I think this may help you better understand the point
made by siddharthprabhu

Albert Resources – 10 Feb 17

1400×800 52.9 KB

Confidence Intervals: What to Know for Statistics | Albert.io

This article reviews confidence intervals, including how to interpret them and differences between one-sided and two-sided confidence intervals.

In short it suggests the true mean of the data set lies somewhere in the 0.06 and 0.69 range. It suggests we are 95% sure that the true mean is somewhere in that range, but doesn't suggest that 95% of the data set falls within that range (a common mistake).

1 Like
#6

Yes, reading up on confidence intervals is the right approach here to help you understand what those values represent.

Here are a couple of lectures on the subject if video tutorials are more your thing.


#9

An example:
If you draw observations from a distribution repeatedly, your confidence interval will cover the mean in 95% of the time.

#10

The distribution you are sampling from doesn't need to be symmetric.

1 Like
#11

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

#3

Thanks for your help Mr Siddharth, very much appreciated. I am new to both R and statistics so please bear with me! It's really helpful for you to clarify that these are lower and upper bounds. I am not really sure how to interpret this! er.... is this saying that there is between a 0.06 and 0.69 probability of the test achieving the same outcome if we were to run the same trial again or what?! Once more your help much appreciated.

#7

Thanks Mr Robin this is a really helpful steer and thanks for the advice on how you interpret this.

#8

Once more Mr Siddharth thanks for your kindness. These are great links ...just what I need. I think I have got my head round this matter now. Thanks again for your help and support.