Clinical trial analysis - have I got it right?

Two-way ANOVA (with replication), how robust is robust?

Hello all,

I would like to see what the pain of a patient is before and after an opperation. But my data is not fitting the assumptions of ANOVA. I have read that ANOVA is robust, but how robust?

I have data from patients who were opporated on in 2006, 2007 and 2008. Each patient completed a total of three different questionaires. I would like to test the questionaires seperately.

The repeated measures (questionaire after opp.) were completed at the same time by all the patients, therefore the patients that were opporated on in 2006 had a longer recovery period. Hence the reaason I would like to use Year as one factor.

The second factor I would like to use is Sex, the oppoeration was back related and I am interested to see if work may have had an effect. History suggests men are more likely to be heavy lifting, etc. than women in the work place, this is especially true as thepatients are all above 65 yrs.

My data is Normally Distributed now, I transformed it!

Non-continuous data!

The questionaires are scored as follows:
RMS 0-24
ODS (%) 0-100
VAS 0-10

Not a balanced data set!

Numer of individuals tested (all questionaires were completed, no missing data)
2006 - 34
2007 - 20
2008 - 49
Total - 103

Is it even worth running the ANOVA, my SPSS package does not have non-parametric suitable.

I am looking forward to replies as I am finally getting back into stats. this work is for my girlfriends diss. I am teaching her as I go but got stuck!

Advice would be greatly appreciated.

Kind Regards


PS This is a edit after reading the guidlines and continued my research to answer the questions I had previously asked.
Last edited:
Hi there just to let people know that I have edited my post having re-read the guidlines. I have also changed the question because I have found out a few of the things I asked.

UPDATE: I will make all my data continuous by converting into percentages! I am guessing/thinking this is ok?

Kind Regards