During the late seventies when Japan was rapidly expanding its share of the

　　American auto market， GM surveyed owners of GM cars and asked them whether t

　　hey would be more willing to buy a large， powerful car or a small， economica

　　l car. Seventy percent of those who responded said that they would prefer a

　　large car. On the basis of this survey， GM decided to continue building larg

　　e cars. Yet during the'80s， GM lost even more of the market to the Japanese

　　……

　　Which one of the following， if it were determined to be true， would best exp

　　lain this discrepancy.

　　（A） Only 10 percent of those who were polled replied.

　　（B） Ford which conducted a similar survey with similar results continued to

　　build large cars and also lost more of their market to the Japanese.

　　（C） The surveyed owners who preferred big cars also preferred big homes.

　　（D） GM determined that it would be more profitable to make big cars.

　　（E） Eighty percent of the owners who wanted big cars and only 40 percent of

　　the owners who wanted small cars replied to the survey.

　　The argument generalizes from the survey to the general car-buying populatio

　　n， so the reliability of the projection depends on how representative the sa

　　mple is. At first glance， choice （A） seems rather good， because 10 percent d

　　oes not seem large enough. However， political opinion polls are typically ba

　　sed on only .001 percent of the population. More importantly， we don't know

　　what percentage of GM car owners received the survey. Choice （B） simply stat

　　es that Ford made the same mistake that GM did. Choice （C） is irrelevant. Ch

　　oice （D）， rather than explaining the discrepancy， gives even more reason for

　　GM to continue making large cars. Finally， choice （E） points out that part

　　of the survey did not represent the entire public， so （E） is the answer.

　　Analogy

　　To argue by analogy is to claim that because two things are similar in some

　　respects， they will be similar in others. Medical experimentation on animals

　　is predicated on such reasoning. The argument goes like this： the metabolis

　　m of pigs， for example， is similar to that of humans， and high doses of sacc

　　harine cause cancer in pigs. Therefore， high doses of saccharine probably ca

　　use cancer in humans.