odds ratio

[sitaram]

I need have odds ratio values with CI for 3*2 categorical(age category[>20,20-40,>40] and smoking(smokers,nonsmokers). How to go about calculating the odd's ratio?

 

[hlsmith]

You have to select one group as the reference to compare the other two groups against. It is usually the group with lowest odds.

 

[noetsi]

A stats package does it automatically

 

[sitaram]

You have to select one group as the reference to compare the other two groups against. It is usually the group with lowest odds.

I agree! how ever any specific reason behind selecting the group having lowest odds as reference? Even if a statistical software can help in getting the results, but it would be great to hear from you if there is any concept behind.Please do reply.

 

[hlsmith]

The reason typically is that most people have difficulty interpreting protective ORs, those under a value of '1'. Also, if your scenario above is what you plan to do, you may have some type of immortality bias or survival bias. The rate of smoking in older persons may be higher, but the smoking based sequelae may have manifested itself and they die (lost to follow-up) or are less likely to provide info since they are institutionalized. There are impacts from smoking.

 

[noetsi]

You just reverse what you are maximizing (the 1 or the 0). If you have the value below .1.