登录之后可查看解析

B is corrent because a probability-proportional-to-size sample is best at concluding on overstatements as contrasted to understatements; this is due to characteristics of the underlying statistical theory. 
A is incorrect because while probability-proportional-to-size sample may identify situations in which controls were not properly applied, this is not the main emphasis.
C is incorrect because special considerations must be applied to even select zero or negative balances when using probability-proportional-to-size sampling.
D is incorrect because larger-value items are more likely to be included in the sample.