If there is a difference in some characteristics would you be able to say that the difference between the two treatment groups is due to the treatment ... or could it possibly be because there was a difference in certain characteristics?
I just have a general question about what the point is of have baseline characteristics and then presenting p-values for these. Is the p-value reflecting any possible differences in the baseline characteristics of the intervention and control group for example? Is this just to prove that the groups are the same at baseline?
Also, what does it really mean if the p-value is statistically significant between intervention and control group at the baseline characteristics? Does this mean that the groups differ a lot on the specific characteristic and therefore are not really similar or comparable?
I guess I'm not understanding what the point is of demonstrating or showing p-values for baseline characteristics. Can someone please explain what the point is?
If there is a difference in some characteristics would you be able to say that the difference between the two treatment groups is due to the treatment ... or could it possibly be because there was a difference in certain characteristics?
I don't have emotions and sometimes that makes me very sad.
But wouldn't randomizing account for this difference??
Thanks though I understand what you are saying
You are correct, randomization functions to make the two study groups equal in characteristics. Providing the p-values shows that the randomization was successful. However there are times when randomization may fail to create equal proportions or comparable samples. Normal instances of significant p-values may be with small sample sizes and affects of outliers or when not using block designs in studies. Much like Dason stated, you can no longer subscribe effects to the intervention/treatment - though, in some instances you can try to account for differences in the analytics.
HS
Tweet |