Interpretation ODDs RATIO

[Tommaso Salvitti]

Good morning everyone,
i have a dubt that i'll try to explain.

logistic y b

.

y=1 (illness)
y=0 (healthy)

b=1 (clinical score)
b=0 (not clinical score)

I got an odds ratio of 16.00 (1.82-142.92) p-value=0.0023
Is it correct to say that the probability (the risk) of obtaining a clinical score is 16 times more frequent in sick/illness people than in healthy people? Thanks to everybody in advance

[Tommaso Salvitti]

If my interpretation is not correct, what is the correct one? thanks a million in advance

[Maarten Buis]

It is not correct. The odds ratio is a ratio of odds and not a ratio of risks or probabilities. The odds is the expected number of "successes" (1s) per "failure" (0s), while a risk or probability is the expected number of "successes" per attempt. Moreover, you switched your dependent and independent variable in your description: the outcome is illness and the "cause" is clinical score; in your text you suggest the opposite.


So the odds of an illness is 16 times higher among people with a clinical score compared to people without a clinical score.

[Tommaso Salvitti]

i ll try to explain better

logistic groupASD clinicalscore

my dep var is groupASD, Where (groupASD=1 if people is autistic, groupASD=0 if the people is not autistic);

in this way is correct my interpretation???

thanks a lot Maarten Buis and thanks to everybody

[Tommaso Salvitti]

my outcome is the "groupASD"

[Tommaso Salvitti]

Thanks Maarten Buis, when you wrote "in your description", do you mean my interpretation?
my depvar is groupASD.


so with this code

logistic  groupASD clinicalscore

so the correct interpretation is the phrase in bold " the odds of an ASD is 16 times higher among people with a clinical score compared to people without a clinical score.???

thanks in advance

[Maarten Buis]

correct

[Tommaso Salvitti]

thanks, only one thing if I said the odds of an ASD compared to Typical development people is 16 times higher among people with a clinical score compared to people without a clinical score is correct the same?
thanks a lot Maarten Buis

[Maarten Buis]

I am getting worried: You are relying on the opinion of a random person on the internet. That is not a good way to do research. Instead you should understand the argument, make up your own mind on whether you agree with it or not, and once you have done that you can apply that argument to minor variations of a sentence. You don't need my seal of approval; it is not worth much anyway.

[Tommaso Salvitti]

no, i rely on the most experienced people. Anyway, thanks a million