How do we Randomize?

By Amos Odeleye-Ajakaye posted 04-24-2014 16:18

  
From my discussions (on-line & off-line) with several statisticians and researchers, some believe Randomization infer Causation and others disagree.

Either you disagree or agree, the next question is:How do we Randomize?

I have 2 consensuses on this issue?
#1. Randomize in sample selection
#2. Randomize in allocation of treatment

Will both methods produce the same or different outcome (s)?

What about estimator bias and confounding?

Please provide answer(s) with references (weblink, journal article, book, etc), whenever possible.
1 comment
45 views

Related Content

Permalink

Comments

Zachary Scott
10-13-2014 14:57

Randomization alone definitely does NOT infer causation. If that was the case, many of these bogus studies around would have validity. You can randomize using some sort of prng, with numbers meticulously assigned. Randomization should not be performed for a perceived outcome, only for randomness and desirable inclusion probability and prevention of bias and preservation of statistical properties.