h1 - Causation

h1 - Causation - Page 1 of 3 Stat 201 Handout (May 21,...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Stat 201 Handout (May 21, 2003) Concluding Causation Q: When can we conclude " treatment causes a change in response "? e.g. " Drug A provides headache relief ." Here, Drug A is the treatment , and relief of headache (as opposed to no change in headache) is the change in response . A: The following are required to make a valid claim of causation: 1. presence of control group -- needed as a baseline for assessing the effects of the treatment 2. no confounding 3. no bias (from subject or investigator) Q: How to control for confounding? A: The only foolproof way is to ensure the treatment and control groups are absolutely equal in every way EXCEPT the treatment itself. e.g. Does Pentium 4 improve CPU speed for a PC? You get 20 identical PC's (with identical parts and chips and everything else), but the first 10 get Pentium 3 while the last 10 get Pentium 4. Now the only difference between the first 10 (control) and the last 10 (treatment) is the Pentium release. (Note : In real life, it is usually impossible to ensure the 20 PC's are absolutely identical due to uncontrollable random elements in the manufacturing process!) The almost -foolproof way to control for confounding is to use random treatment allocation . Randomness tends to balance out differences outside of the treatment
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 3

h1 - Causation - Page 1 of 3 Stat 201 Handout (May 21,...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online