32. Difference estimation is a special case of regression estimation where the slope is fixed at B = 1 rather than estimated from the data. In accounting, for example, an auditor knows the book value ; for each of the N accounts in a list of accounts receivable, then determines the audited value y; (the actual amount owed) for each account in an SRS of size n drawn from the population. The difference estimator of the population mean yu is Ydiffy+(xu-x). This is similar to (4.16), but with B replaced by B = 1. The sample mean is adjusted by the difference between the population and sample mean of x, (u - x). (a) Show that E [diff] = u. (b) What is V (diff)? (c) Compare V (diff) with the variance of the ratio estimator in (4.10). When will the difference estimator have smaller variance than the ratio estimator?