I tested the activity of some prepared formulations of mosquitocidal bacilli through 10 months storage. these formulations was 40 and i tested the activity of these formulations every month against mosquito larvae and calculated the mortality %.
Data were statistically analyzed according to SPSS system using one-way analysis and the Duncan’s multiple range test to determine the significance between means. Data were expressed as mean values ± standard error. during publication of these data the reviewer still question the appropriateness of Duncan's multiple range test for post-hoc comparisons, much less for the large number of comparisons indicated in each column.
He did not suggest which test can i use.
My question is which test can i use to statistically compare these data.
Duncan's multiple range test is especially protective against false negative (Type II) error at the expense of having a greater risk of making false positive (Type I) errors.
May be the review is more aware of Type I error?
Tukey's range test is commonly used to compare pairs of means and this approach controls better Type I error.
Another solution is to perform Student's t-test of all pairs of means, and then to use false discovery rate (FDR) controlling procedure (to control the expected proportion of incorrectly rejected null hypotheses). See http://www.stat.cmu.edu/~genovese/talks/hannover1-04.pdf https://en.wikipedia.org/wiki/False_discovery_rate regarding this scenario.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS