|
Freethought & Rationalism ArchiveThe archives are read only. |
05-12-2002, 04:38 AM
|
#1 |
|
Contributor
Join Date: Jan 2001
Location: Barrayar
Posts: 11,866
|
Medical language for medical guys: a question
Assuming detection of a significantly increased risk of 2-fold in these genotypes, our study can detect at about 76% power among 100 cases and 196 controls
I found the above sentence in a document I am editing. What does she mean "76% power?" The article is on cervical cancer. Vorkosigan |
|
|
05-12-2002, 05:47 AM
|
#2 | |
|
Veteran Member
Join Date: Mar 2002
Location: anywhere
Posts: 1,976
|
Quote:
In this case, the null hypothesis is probably that cervical cancer is not present. So 76% of the time, their test would be able to detect cancer when in fact there is cancer. In mathematical terms, if the probability of not detecting a cancer when in fact there is cancer is b, then the statistical power is 1-b. Scientiae [ May 12, 2002: Message edited by: Scientiae ]</p> |
|
|
|
|
05-12-2002, 02:02 PM
|
#3 |
|
Contributor
Join Date: Jan 2001
Location: Barrayar
Posts: 11,866
|
Thanks, Sci!
V |
|
|
|
05-12-2002, 06:42 PM
|
#4 |
|
Veteran Member
Join Date: Jul 2001
Location: Dana Point, Ca, USA
Posts: 2,115
|
You got it covered Scientiae. What seems left unsaid is their rate of false positives.
|
|
|
| Thread Tools | Search this Thread |
|
