university students and I want to estimate the following model for first-year student academic performance at university: WAM = β0 + β1faminc + β2ATAR + β3attend + u where WAM = a student's weighted average mark at the end of the first year faminc = a student's family income in $000s ATAR = a student’s university entry score (0-100) attend = the proportio
I have a random sample of data on university students and I want to estimate the following model for first-year student academic performance at university:
WAM = β0 + β1faminc + β2ATAR + β3attend + u
where
- WAM = a student's weighted average mark at the end of the first year
- faminc = a student's family income in $000s
- ATAR = a student’s university entry score (0-100)
- attend = the proportion of classes the student attended in their first year
I want to estimate the causal effect of class attendance (attend) on WAM. I am concerned there is an omitted variable z that I cannot get data on but I believe it belongs in the model.
It is well known that z has a positive effect on WAM and that z is negatively related to class attendance (attend).
To answer this question, please assume that ATAR and faminc are both uncorrelated with attend and are both uncorrelated with z.
If I estimate the model above without including the variable z, then I can expect that:
b) β^3 will be unbiased
c) None of the other possible answers are correct.
d) β^3 will be positively biased
Trending now
This is a popular solution!
Step by step
Solved in 3 steps
