Thread: Medical statistics. Comparing tumor stages in different partients over time.

Hi,

Below is one dataset with 315 different patients with different tumors staged from stage 0 to 4.
In doubt whether ANOVA is right here! Need to know any stat sign diff distribution of stages within each year and any change over time within each stage!

Thanks
kim

	2007	2008	2009	2010	2011
Stage 0	5	5	11	2	0
Stage 1	6	16	22	20	22
Stage 2	15	17	22	20	11
Stage 3	22	15	20	21	20
Stage 4	13	3	1	3	2

Hey kemorex.

Have you tested the assumptions of your data for the ANOVA (Normality, etc)?

Remember that an ANOVA is just a routine that checks whether there is evidence of different means (or medians depending on whether its parametric or non-parametric).

If you are looking at changes over time you will need to use longitudinal analyses and not a standard ANOVA. Are you familiar with longitudinal techniques and analyses?

I don't think these data are normaly distributed.
Yes. I've worked a bit with repeated measures using linear mixed models, but that was with individual subjects with repeated measurementrs over time. Alle patients within each cell in this table are different subjects. ..
Thanx for your reply
kemorex