I have a system where I'm studying 3 different diseases, however they share some pathological similarities. There is no data on them being similar in the properties that I am studying, but of course it cannot be ruled out that the three will share that specific characteristic. So, in this situation should I consider the three diseases to be independent? Or should I assume dependency for the sake of statistical analisys? Is there any rule of thumb in this situation?

Thanks! ]]>

I have 7 different properties, all independent, that are being mesured. Each property has a control and three disease groups. All independent. With each property, I compare the 3 disease groups to the control, but not the disease groups to eachother. I do this with ANOVA + Dunnett's test. So I have 7 ANOVAs.

My problem is that I'm not sure how I should correct for multiple testing because I have 7 ANOVAS. Right now I do a Benjamini-Hochberg process for 7 measurements to correct the alpha for the p-values of the ANOVAs, then I just do the Dunnett's tests (in the ANOVAs that still have significant p-values) with no further correction. Is this correct?

Thank you! ]]>

I'm not the best at stats so please keep terms and approaches in layman's terms please - I currently use SPSS too :)

I'm not too sure which test to conduct, so any help would be great.

I have record of the number of deaths since 1998 - 2016, each having a month when recorded - I subsequently subdivided the months into 4 seasons, and am now trying to test the number of deaths against the seasons. I have divided each by the number of months in each season to compensate for the different number of months in each season..

The data is non-parametric...

I have been recommended to carry out logistic regression analysis, but as the data is not proportional I don't see this as appropriate. I tried GLMZ, poissons regression as it is count (isn't it?) but that worked before I divided it but didn't work after I divided it.

I have attached an image of an example of the data.

I'm just going round in circles and I really don't know which test to conduct...

Any help would be gratefully appreciated!

Regards

J