In: Statistics and Probability
How do you determine significance based on confidence intervals?
1. If using a ONE SAMPLE t-test, and the null hypothesis is H0: u = 100, if the confidence interval interval contains 100 (for instance, the interval is [70, 136]), is the result NOT significant?
2. If using an TWO INDEPENDENT groups t-test, and the null hypothesis is H0: u1 = u2, if the confidence interval contains 0 (for instance, the interval is [-1, 5]), is the result NOT significant?
Basically, I need to know the difference between which value is important to determine significance using confidence intervals for both kinds of t-tests.
## How do you determine significance based on confidence intervals?
if interval include null value then result is not significant . and if interval free from null value then result is significant .
1. If using a ONE SAMPLE t-test, and the null hypothesis is H0: u = 100, if the confidence interval interval contains 100 (for instance, the interval is [70, 136]), is the result NOT significant?
Answer : Interval is free from null value and u = 100 is include the interval hence the result is significant . at given level .
and we reject Ho .
2. If using an TWO INDEPENDENT groups t-test, and the null hypothesis is H0: u1 = u2, if the confidence interval contains 0 (for instance, the interval is [-1, 5]), is the result NOT significant?
Answer : we have given confidence interval for the difference between means .
interval is -1 to 5 interval include null value hence result is not significant .
here we fail to reject Ho that is population means are equal .
## note : In testing there are p value approach , critical value approach and confidence interval .
we can use above approach and can conclusion .