E leaves more possibility for GM plan to be correct. 0.8*0.7=0.56 so GM had an accurate information about 56% of the market. It seems reasonable for GM to continue producing large cars.
A demonstrates that the GM plan was a an act of wishful thinking.
I vote for A.
Stolyar be careful:
There is nothing wrong with A per se. The percentage size of a sample (so long as it reaches a certain criteria which is beyond the scope of the GMAT) is not so important as whether or not such sample is "representative" of the underlying population. In fact, most statistical surveys and experiments use much much MUCH smaller sample sizes than 10% on a percentage basis and yield remarkable accurate results (high % "confidence levels" in stat terms).
IMO, E is the correct answer because it states that a larger than representative sample of large car affectionados relative to those of small cars answered the survey, thus exagerating the weight of the large car lovers' survey answers.
Here is an example using numbers. Suppose there are 1625000 people in the target population of which 875000 (54%) favor large cars and 700000 (46%), small. Hence, the actual number of those favoring small cars in the actual
population is 46% -- demand which should not be ignored. If 80% of the large car people actually vote for large cars and only 40% of the 750000 small car people actually vote for small cars with the remainder not participating, the final tally would be 700000 to 300000 or 70% to 30% leading those who do not know the underlying demographics and participation rates to incorrectly infer that only 30% rather than the much stronger 46% of the population wants smaller cars. This erroneous interpretation might even be aggravated if previous data had smaller cars at a higher than 46% share, which might be interpreted by a careless marketing executive as a declining demand for smaller cars when, in fact, an increasing demand actually exists.
As an aside, this is why self-selecting surveys such as internet polls typically yield data that is prone to ambiguous interpretation leading to conclusions that are often more self-serving or entertaining than meaningful.
Former Senior Instructor, Manhattan GMAT and VeritasPrep
Vice President, Midtown NYC Investment Bank, Structured Finance IT
MFE, Haas School of Business, UC Berkeley, Class of 2005
MBA, Anderson School of Management, UCLA, Class of 1993