We are looking for the most helpful explanation for the apparent discrepancy in the statistics.
The apparent discrepancy is the following:
Fact 1: Of the people who responded to the survey, 75% said they had started for their respective team.
Fact 2: However, the “actual number” of people who had started was 50% (of the actual players).
The “scope shift” in the basis upon which the statistics are based will most likely help to explain this discrepancy. Fact 1 is a percentage based on the people who “responded” to the survey. This number may very well be different than the total number of “actual” players on the team.
If most of the respondents were ppl who had started, then it would make sense that Fact 1 is skewed more towards “starters” than is the actual number.
(A) suggests that a very small number of survey participants were incorrect in their response to the survey. Depending on the size of the teams and amount of players involved, this may or may not be enough to account for the apparent discrepancy in the statistics. It does not seem to be strong enough to explain the significant difference.
(C) tells us that not all starting players responded to the survey. This fact, if true, would deepen the paradox. If there are some starters who didn’t even bother to respond to the survey, then how can it be that the percentage of starters, as registered by the survey, is larger than than the ACTUAL percentage of starters (based on all the players)? We would expect the survey percentage to be less than the actual percentage.
(D) also deepens the paradox. If almost all relevant people responded to the survey (and assuming they did so truthfully), then the percentage provided by the survey would seem to be based on a good representative sample of the entire population of players. We would expect the statistic in Fact 1 to be almost equivalent to the statistic in Fact 2.
(E): Wether or not good players were “held out” until later in the game does not help us understand why there exists a discrepancy in the statistics describing the percentage of players who had started.
However, if what (B) tells us were true - a disproportionately high number of starters responded to the survey - then this fact suggests that the statistic provided by the survey is based on a skewed sample of the population. If most of the respondents were starters, we would expect the percentage of respondents who said they had started to be greater than the percentage of actual players who had started.
Answer (B) helps the most in explaining the apparent paradox.
Posted from my mobile device