# Population specific odds ratio and subject specific odds ratio

#### Iris Ilja

##### New Member
Hey guys,

For a school assignment I have to compute the population specific odds ratio and the subject specific odds ratio from the table below.

No MDD 350 (93%) 26 376 (62.7%)
MDD 153 71 224 (37.3%)
Total 503 (83.8%) 97 (16.2%) 600 (100%)

This is what I have on it so far:

The odds of no GAD in the no MDD group are 13.46 (350/26), and the odds of no GAD in the MDD group is 2.15 (153/71). The odds ratio in this case is 6.26 (13.46 / 2.15). So it can be said that the odds of having no GAD and no MDD is 6.12 times larger than the odds of having no GAD and having MDD.

Did I calculate the population or subject specific odds ratio there? And how to do the other one?

#### hlsmith

##### Not a robit
Not familiar with this, how exactly was the problem presented, verbatim?

#### Iris Ilja

##### New Member
I got this question on an assignment from school:

Consider major depression (MDD) and generalized anxiety disorder (GAD). The
interest is now in the question whether more people have GAD or MDD, therefore
look at the marginal distributions of MDD and GAD.
(a) Compute the population averaged odds ratio and interpret it.
(b) Compute the subject speci c odds ratio and interpret it

#### hlsmith

##### Not a robit
Marginal distributions are usually when you don't stratify the data by a group just base things on the totals. Though you are only presenting the data above as a single 2 x 2 table. Is that what they did in the problem?

#### Iris Ilja

##### New Member
No I created the table in order to calculate the odds ratio (I think population odds ratio). Do you need more info and if yes what do you need?

#### hlsmith

##### Not a robit
I hate to sound like a broken record, but what was the exact original question - presented with percentages and counts. So you reiterate the question, but how was data introduced?