Ilhomjon98 wrote:
Yet, I couldn't completely comprehend why (B) is correct.
Could you please explain why (B) is correct?
Yes, E is wrong because the argument is only about averages. As long as tips are larger on average, the servers make more money, and it doesn't matter if every tip is 3% larger, or if one tip is huge and 99% of tips are unchanged.
Here we have a small-scale experiment which produced a 3% increase in tips. If we're going to translate that to a large-scale experiment (where servers always write 'thank you' on bills), then if we're expecting we'll still see a 3% increase in tips, we're assuming nothing will change as we scale things up. That is, we're assuming customers will behave the same way if they always see 'thank you' on their bill as they did in the small-scale experiment where they were seeing 'thank you' on their bill for the first time (or at least when they only rarely saw 'thank you' on a bill). So that's why B is right.
There are a lot of similar arguments, in the real world and on the GMAT, which make a similar assumption. For example, and this is not based on any actual facts about limes, I'm just inventing a scenario: if a two-week-long study proved that eating a lime every day improved people's moods, and based on that data, dieticians suggested everyone should eat limes every day to improve mood long-term, that conclusion would be assuming that people don't build up a tolerance for or immunity to the happiness-producing effects of limes. That is, the conclusion is assuming when you scale things from a short two-week study to a permanent behaviour, people continue to respond identically. That may or may not be true.