Hi,
I'd like to ask you for some help with the interpretation of the outputs of mi OLR. I've been looking through several manuals and I can't find clear answers...
I ran a model with an ordinal dependent variable (1 to 5 levels of a scale that goes from negative/bad to positive/good consequences for health) and a main categorical independent variable that has 4 categories, plus a few control variables (age, sex, occupational level...).
I was reading the Odd Ratios resulting from the ".ologit ... , or" instruction as it follows:
When my practice in relation to health equals "category A" of the independent variable, the OR of being located in the following best level of the dependent variable increases in X points in comparison to the "Reference category" of the independent variable.
When my practice in relation to health equals "category B" of the independent variable, the OR of being located in the following best level of the dependent variable increases in X points in comparison to the "Reference category" of the independent variable.
...
Also, I'd say that if the OR of practicing B versus the Reference category are higher than the OR of practicing A versus the reference category, then B is better for health than A.
I dont want to transform the odds into probabilities as some manuals say. Instead I want to be sure that the OR provided by Stata for each category of the independent variable refers to the odds of being in the next superior level of the dependent variable, which in this case indicates a better level health.
I have Stata v.12
Thanks a lot.
PS: This is my first post, sorry in advance if something is against the forum's conventions.
I'd like to ask you for some help with the interpretation of the outputs of mi OLR. I've been looking through several manuals and I can't find clear answers...
I ran a model with an ordinal dependent variable (1 to 5 levels of a scale that goes from negative/bad to positive/good consequences for health) and a main categorical independent variable that has 4 categories, plus a few control variables (age, sex, occupational level...).
I was reading the Odd Ratios resulting from the ".ologit ... , or" instruction as it follows:
When my practice in relation to health equals "category A" of the independent variable, the OR of being located in the following best level of the dependent variable increases in X points in comparison to the "Reference category" of the independent variable.
When my practice in relation to health equals "category B" of the independent variable, the OR of being located in the following best level of the dependent variable increases in X points in comparison to the "Reference category" of the independent variable.
...
Also, I'd say that if the OR of practicing B versus the Reference category are higher than the OR of practicing A versus the reference category, then B is better for health than A.
I dont want to transform the odds into probabilities as some manuals say. Instead I want to be sure that the OR provided by Stata for each category of the independent variable refers to the odds of being in the next superior level of the dependent variable, which in this case indicates a better level health.
I have Stata v.12
Thanks a lot.
PS: This is my first post, sorry in advance if something is against the forum's conventions.
Comment