topic Re: Covariate balance randomized trials in Statistical Procedures

Covariate balance randomized trials

bollibompa — Sat, 05 Nov 2016 15:40:54 GMT

Hi, this may a question posted the wrong forum but I trying here.

If patients are randomzied in a clinical trial one is expecting that both measured and unmeasured baseline covariates are balanced between treated and controls.

My question is how you mathematically/statisticaly can derive that balance is increased when the study popualtion is increasing?

Thanks in advance!

Thomas

Re: Covariate balance randomized trials

PGStats — Sat, 05 Nov 2016 20:44:27 GMT

In my mind, it goes like this.

Because of randomisation, the expected value of both control and treated groups are the same for all covariates. When the sample sizes increase, the standard errors of estimated means decrease. Thus the expected difference between treated and control group means decreases.

Re: Covariate balance randomized trials

bollibompa — Wed, 16 Nov 2016 16:26:06 GMT

Thank you PGStats for your quick reply!

/Thomas