Timeline for Converting OR to percentages and interpretation
Current License: CC BY-SA 3.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 12, 2012 at 10:32 | comment | added | Michael R. Chernick | @TryingHard My answer regarding R square did not directly answer your question which is probably the reason someone chose to downvote it. My purpose was to point to a different approach to the problem that did not involve proportions of women. Instead of looking at the relationship between SLE and smoking among women, I was suggesting that the relationship of smoking to SLE can be viewed through a statistical model where R square represents the amount of variability in the occurrence of SLE that is explained by whether or not the woman is a smoker. | |
| Jun 12, 2012 at 9:28 | comment | added | Trying Hard | I am really sorry Michael, but I agree with "whuber" comments as I dont understand your answer however it is probably that this subject is really new to me and I need "baby" talk to help me understand. So sorry. | |
| Jun 11, 2012 at 14:00 | comment | added | Michael R. Chernick | I recognize that the question asked for a conversion from odds ratios to proportions and addressed that in my second answer. My feeling here was that this conversion idea is not a very good one and I think that address the question in terms of variance explained is a better way to look at things. The contingency table expresses counts for SLE as a function of smoking or not and therefore can be looked at as a simple ANOVA model. I don't see what is unclear about my response. R square provides an an estimated of percentage of variance explained by the factor smoking (yes or no). | |
| Jun 11, 2012 at 13:36 | comment | added | whuber♦ | I did not downvote, Michael, but I am inclined to convert this reply into a somewhat mysterious comment to the original question. I am puzzled why you are invoking a "linear model" and "a partial $R^2$" when the question appears to concern a 2 x 2 contingency table and, by virtue of its homework tag, it begs us for elementary and clearly explained responses. | |
| Jun 11, 2012 at 2:16 | history | edited | Macro | CC BY-SA 3.0 |
TeXed R^2
|
| Jun 11, 2012 at 1:22 | comment | added | Trying Hard | So then would the following be the correct anser for question b)??An odds ratio of 1.0 means there is no increase of decrease in risk of developing SLE. Therefore for an odds ratio of 1.80, there is an increase in risk for developing SLE due to smoking of 0.90 or 90/100 over the general population risk. | |
| Jun 11, 2012 at 1:17 | comment | added | Michael R. Chernick | Would the person that downvoted this response explain their reasoning? | |
| Jun 11, 2012 at 1:06 | history | answered | Michael R. Chernick | CC BY-SA 3.0 |