For my final year project I'm looking to evaluate pre-existing two-sample test procedures in terms of preliminary tests and their effect on the selection of the most appropriate two-sample test, through simulations. Due to a lack of consensus, I'm hoping to be able to come to some conclusion as to the most appropriate decision making process for certain preliminary tests at least. It would be beneficial to have a few peoples two-sample decision making processes for continuous data with independent samples, to give me an idea of what researchers think about when confronted with the problem of comparing two-samples.
Any help would be appreciated.
You would compare the means using a 2 independent sample t-test. Establish your null hypothesis, check for normality of data (using a residual plot) and for independence (using a Durbin-Watson statistic or Chi squared test), then run the t-test, which will calculate means, standard error, t or z statistic and the p value, which if < 0.05 will indicate a significant difference in the means :)
It is not clear to me the connection between decision making process and sampling design. Try Six Sigma approach by the Minitab software. You'll find a lot of tools there to use with your data.
- Badges
- Science method
- Similar topics
- Mathematical Sciences
- Statistics
- Sampling Design