This Exam P sample reference tests Joint Distributions. Collapse the joint table by the possible values of the total cost. The resulting total has mean 2.04 and second moment 278.4, so its variance is 274.2384 and its standard deviation is 16.5601, selecting choice D.
Use the collapsed distribution to form the first two moments of the total.
E[W]=w∑wPr(W=w)=2.04
E[W2]=w∑w2Pr(W=w)=278.4
Compute
Compute
Subtract the square of the mean from the second moment, then take a square root.
Var(W)=278.4−(2.04)2=274.2384
SD(W)=274.2384=16.56014493…
Answer
Answer
The standard deviation rounds to 16.6.
16.6(D)
Test the method on unseen problems
Probability Multivariate & Risk Models Practice Workbook
30 original multivariate random variables problems with complete solutions and wrong-choice analysis. Written for Exam P, not copied from the SOA set. One-time $19 PDF.
These notes identify the calculation error associated with each wrong letter when that error is reproducible.
ACounting only one cell from each symmetric off-diagonal pair gives the incomplete raw sums E[W]=1.12 and E[W squared]=153.6. Applying the variance formula anyway gives √(153.6-1.12 squared)=12.3428, which rounds to 12.3.
BThis retains only the one-disabled-person totals 40 and 200 and moves every both-disabled outcome to zero. The resulting standard deviation is 13.7637.
CThis treats the two costs as independent. Their marginal variances sum to 245.9192, whose square root is 15.6818, but the joint table shows positive covariance.
EThis value is impossible under the supplied table: 19.8 squared exceeds E[W squared]=278.4, whereas Var(W) cannot exceed its second moment.
Original practice · fully worked
Original variant: correlation from a shared demand component
Independent centered random variables A, B, and C have variances 4, 5, and 7. Two regional demand changes are modeled by X=A+B and Y=A+C, so the regions share only component A. Calculate the correlation between X and Y.
A 0.040
B 0.364
C 0.402
D 0.444
E 0.676
Answer and calculation
The shared component contributes covariance 4, while the two marginal variances are 9 and 11. Therefore the correlation is 4 divided by the square root of 99, or 0.402015, so choice C.
Setup
Setup
Add the independent variance components within each regional change.
Var(X)=4+5=9
Var(Y)=4+7=11
Model
Model
Only the shared centered component contributes to the covariance; all cross-covariances between distinct components vanish.
Cov(X,Y)=Cov(A+B,A+C)=Var(A)=4
Compute
Compute
Normalize the covariance by the product of the marginal standard deviations.