Now posted.
Question 2b asks us to add 0.5 to all of the values in the hair_EtG variable before transforming it by the natural log to make a new variable. This is fine. However after this sub-question we are asked to extensively use the variable hair_EtG variable. We are never asked to subtract the 0.5 away from hair_EtG. In other words we are never asked to return the hair_EtG variable to the state it was before sub-question 2b. All of the subsequent questions will be using the hair_EtG transformed by the addition of 0.5 the way the exam is presently written. Thanks.
Please use the original variable for the analyses of the hair_etg variable.
Do you mean that after sub-question 2b, you'd like us to transform hair_EtG back into it's original form).
In questions subsequent to 2b that ask us to use the ln transformed EtG. Is this using the ln transformed original hair_EtG or the hair_EtG that had 0.5 added to it?
you should not use etg +.5 anywhere. It is only so that you do not get missing values when you take the natural log (because ln(0) is neg infinity).
When it says to use the ln etg, it means the new variable you made using ln(etg+.5).
When it says to use the raw data, use the original data.
You should never write over you original variables, by the way.
Question 2.h
bullets 4
- What other test would yield the same conclusion? Run that test
As I recall (maybe I'm wrong), the exam does not include material covered on lecture 11 (last). Does this question refer to another test that yield same conclusion e.g. correlation (bullet3) or mean values (bullet2) or both.
Bullet 5
Are the assumptions required for these analyses met? Why or why not?
Which one of the analyses performed in this question are referenced?
Thank you
Meir
The answer to this question is not from lecture 11, except perhaps if you look at the tables of tests (but it is review, not new material).
Maybe it would help to think about the type of varlables -- what type of variable is the dependent/outcome variable? What type is the independent/predictor variable?
The question about the assumptions applies to both the linear regression and the equivalent test for this case.
Hi I just wanted to have something clarified. When we say regress Y on X, I thought that this would mean we want to predict values of Y given values of X which would make X the independent variable and Y the dependent variable. Is that not correct?
I've noticed multiple times that the exam has been saying: "number of days drinking alcohol in the prior 21 days regressed on EtG in hair (raw data) (EtG in hair is the dependent variable)" so it is clear that we mean for EtG to be the dependent variable but by saying number of days regressed on EtG it makes it sound like EtG is the independent variable.
Could you clarify which is correct? Thank you.
hair etg is the dependent variable
hair etg is the dependent variable
Exam should be mailed to the TA mailbox (biostat200.2013@gmail.com), correct?
Thank you
Meir
Yes.
Yes.