In: Statistics and Probability
24.
Let S be the educational attainment of individuals in a town, with values S=0 for less than high school and S=1 for high school or above. Also, let Y be their individual annual income with values Y=0 for less than $20,000, Y=1 for between $20,000 and $40,000, and Y=2 for above $40,000. Consider now the following joint probabilities:
S\Y Y=0 (less than $20K) Y=1 ($20K-$40K) Y=2 (more than $40K)
S=0 (Less than HS) 0.05 0.03 0.01
S=1 (HS or more) 0.22 0.36 0.33
Determine the expected value (mean) of the conditional expected level of educational attainment (given the various levels of individual annual income).
Consider the two events
S : Level of educational attainment.
Y : Levels of individual annual income.
Y | Total | ||||
Y = 0 ( Less than $ 20 K) | Y=1 ( $ 20 K - $ 40 K) | Y=2 ( more than $ 40 K) | |||
S | S =0 ( less than High School) | 0.05 | 0.03 | 0.01 | 0.09 |
S=1 ( High school or more) | 0.22 | 0.36 | 0.33 | 0.91 | |
Total | 0.27 | 0.39 | 0.34 | 1 |
The Marginal probability distribution of S is
S | 0 | 1 | Total |
P(S=s) | 0.09 | 0.91 | 1 |
The marginal probability distribution of Y is
Y | 0 | 1 | 2 | Total |
P(Y=y) | 0.27 | 0.39 | 0.34 | 1 |
We have to find
E ( S / Y=0) , E(S / Y=1) and E(S / Y=2).
by definition of conditional distribution
P ( S / Y=y) = P ( S=s, Y=y) / P ( Y=y).
Hence the conditional distribution of S given Y = 0 is
S | 0 | 1 | Total |
P(S/ Y=0) | 0.1852 | 0.8148 | 1.0000 |
by definition of conditional expectation
E ( S / Y=0) = 0 * 0.1852 + 1 * 0.8148
E(S / Y = 0) = 0.8148.
The conditional distribution of S given Y =1 is
S | 0 | 1 | Total |
P(S / Y=1) | 0.0769 | 0.9231 | 1.0000 |
E ( S / Y=1) = 0 * 0.0769 + 1 * 0.9231
E ( S / Y=1) = 0.9231.
The conditional distribution of S given Y =2 is
S | 0 | 1 | Total |
P(S / Y=2) | 0.0294 | 0.9706 | 1.0000 |
E ( S / Y=2) = 0* 0.0294 + 1 * 0.9706
E(S/ Y=2) = 0.9706.