If the effect modifier is categorical (video mins 17:44), could you show how to calculate the smoking effect for men and women separately? This might be straightforward but not sure which formula to use and what to plug in?.
I'm a bit unclear on the question. here are a few things that may help clarify. first, confounding and interaction as different things. you may consider checking the (age*smoke) interaction if you believe that the effect of smoking should change depending on ones age. if you think this makes sense conceptually, you can test the interaction between the two.
While calculating the interaction between Age and Smoke, why haven't we considered the coefficient of Age? for ex- why haven't we added 50*0.093367 in the interaction equation?
If the effect modifier is categorical (video mins 17:44), could you show how to calculate the smoking effect for men and women separately? This might be straightforward but not sure which formula to use and what to plug in?.
for confounding variables, Age - should we also check with Interaction term Age*Smoke
I'm a bit unclear on the question. here are a few things that may help clarify. first, confounding and interaction as different things. you may consider checking the (age*smoke) interaction if you believe that the effect of smoking should change depending on ones age. if you think this makes sense conceptually, you can test the interaction between the two.
@@marinstatlectures thank you...I now do understand that confounding and interaction are different things..
While calculating the interaction between Age and Smoke, why haven't we considered the coefficient of Age?
for ex- why haven't we added 50*0.093367 in the interaction equation?
if you write down the equation, you will find the the cofficient of Smoking which is we care about