Randomisation baseline testing

Test 1 is telling us that the treatment means are comparable at baseline, which is what we would expect to see given that this is a randomised trial. Of course chance differences can sometimes occur. Indeed in a randomised trial we would expect to see p < 0.05 for such a baseline comparison 5 per cent of the time. See Section 6.9 for a further discussion on this point. 

We will return to the example where we have just a single baseline variable, size of primary tumour, predicting the outcome, time to disease recurrence, but now in addition we have randomised the patients to one of two treatment groups, test treatment and placebo. 

Statistical testing for baseline imbalance has no role in a trial where the handling of randomisation and blinding has been fully satisfactory. ... 

---