We adjusted for sex and age in our first model and for sex, age, race/ethnicity, educational level, marital status, region of residence, and smoking status in our second model
theskinnyduck wrote:The beauty and degree of information of correlational studies...
The relation was only true when controlling for age and sex. I wonder why they dont mention conclusions for their second model.We adjusted for sex and age in our first model and for sex, age, race/ethnicity, educational level, marital status, region of residence, and smoking status in our second model
Darren72 wrote:theskinnyduck wrote:The beauty and degree of information of correlational studies...
The relation was only true when controlling for age and sex. I wonder why they dont mention conclusions for their second model.We adjusted for sex and age in our first model and for sex, age, race/ethnicity, educational level, marital status, region of residence, and smoking status in our second model
Actually, it is the opposite. If you look at their Table 2, the relationship between obesity and cannabis use is only statistically significant when they don't control for anything. Once they control for these factors, the odds ratios are not statistically different. Strangely, they don't note this in the text.