Statistics · Regression · Medium
A variable left out
The true relation is y = 2x + 3z + noise, where x and z each have variance 1, their correlation is 0.5, and the noise is independent of both. With a very large sample you regress y on x alone, with an intercept. What slope do you get?
a fraction, an expression or a decimal to two significant figures · 5/8 · ln 2 · pi · inf
Answer and solution
Answer
7/2 ≈ 3.5
Solution
A regressor takes credit for whatever it is correlated with that has been left out.
cov(y, x) = 2 × 1 + 3 × 0.5 = 3.5, and var(x) = 1, so the slope is 3.5: the true 2, plus 3 times the slope of z on x.
Regression comes up in 11 firms' roundsHudson River Trading · Goldman Sachs · D. E. Shaw · +8
These 16 rounds ask this kind of problem live, with follow-ups. Every round on every track is in Premium.