Worked solution
Normalise the parameterised distribution before transforming variance
The probability function of X is
where .
(i) Show that . Find the distribution of X in terms of θ.
(ii) Find in terms of θ and hence show that .
(iii) , with a,b non-zero constants. Given , find θ.
[2017/MJC/II/6]
CJC · Discussion · Question 5
Answer
; distribution above; ; variance ; .
(i) Retain the source’s initial table and normalisation
| x | before substitution |
|---|---|
| 1 | |
| 2 | |
| 3 | |
| 4 |
Summing all probabilities gives , so .
(i) Complete distribution after substituting k
| x | |
|---|---|
| 1 | |
| 2 | |
| 3 | |
| 4 |
The stated interval ensures all four probabilities are positive; other values of X have probability 0.
(ii) Mean with every source expansion term
.
(ii) Second moment and variance
. Hence .
(iii) Cancel the non-zero scale factor
. Thus . Since b≠0, divide by to obtain .
(iii) Apply the parameter restriction
The source calculator roots are θ≈0.0144 and θ≈0.118. The latter exceeds 1/9 and is rejected. Hence θ≈0.0144. An exact expression for the retained root is ; the source’s rounded answer is preserved.
Common mistakes
The second quadratic root is inadmissible under 0<θ<1/9. Do not retain both numerical roots.
Solution notes
All three asks, piecewise support/zero-elsewhere condition, initial and final source tables, every mean/second-moment/variance expansion, quadratic and both source roots/rejection retained. Exact root supplements the printed calculator approximation. No marks printed.