登录之后可查看解析

B is corrent because although the use of random sampling makes it possible to calculate statistical conclusions, the wrong size sample will affect the achieved risk of assessing control risk too low. For example, a sample which is too small will lead to a higher level of assessing control risk too low (more sampling risk) than desired. 
A is incorrect because nonsampling risk includes all aspects of ultimate risk that are not due to sampling. However, the error committed in this case is a sampling error.
C is incorrect because discovery sampling, a type of attribute sampling, uses a predetermined sample size to arrive at statistical conclusions.
D is incorrect because inferences can be drawn.