Hello,
I am currently busy with my masterthesis and I have a question regarding the bootstrap of the paired samples t test. My data contains a small sample size (n =18). My supervisor advised to do a bootstrap paired samples t test. When I run the test (n = 18), because of missing data.
However, now I have a question regarding the interpretation. In total, I have examined 9 neuropsychological tests (on two time points). Two results seem very contradictory.
One pair has a BCa 95% CI of (-.21, -.07) and a p value of .113
Another pair has a BCa 95% CI of (-.07, -.21) and a p value of .111
Now I was wondering where to check for the significance. The 0 is not in the CI (so it is significant)? However the p value is > 0.05. Does anyone know how to solve this?
Thankyou in advance!
Best,
Anouk
Look in your package manual. You have perhaps entered the data incorrectly . If not then there's no difference between the two tests you checked. However the research design is so poor that it is hard to say much. Your advisor should be ashamed of himself for allowing you to do this and telling you it should be done at all given the described circumstances. David Booth
Hi,
The p values from either of the sets are not significant at or below 0.05. The C.Is is not wide. The null hypothesis is accepted.
@everyone. The null hypothesis is never accepted or at least the power of the test should be given along with the alpha. David Booth
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs