#5 - Researchers asked 100 fifty-year-olds and 100 twenty-

[Administrator]

Please post below with any questions!

[mokkyukkyu]

Hi, How is the last sentence an alternative explanation?
I chose C...

Thank you

[Adam Tyson]

In this case you can read "alternative explanation" as "alternative conclusion" or even "alternative cause". The data is fine - our author is not questioning the results of the study, and accepts that twice as many older folks as younger folks said they give blood. The difference here is in the interpretation of that data.

The researchers have concluded that older folks are more altruistic - they are more likely than their younger counterparts to be giving and generous to help others. They apparently based this conclusion on an assumption that the data they got is good, accurate, truthful data. That is, they believe that the older folks actually DID donate blood twice as much as the younger ones.

Our author suggests a different explanation for the data - she suggests that the data may be inaccurate, untruthful. Perhaps that means the older folks lied about donating blood, because they were ashamed to admit that they don't? Maybe the younger folks lied about not donating, because they were embarrassed to admit that they do? Whichever group gave false answers, and especially if both groups did, the study would be less reliable - maybe completely unreliable - because of bad data. That alternate explanation for the results of the study calls into question the researchers' conclusion.

This language is typical of many weaken questions in causal arguments, but can be found many other places, such as this Method of Reasoning question about a survey. It can be viewed as causal - what caused one group to differ from the other in the way they answered? Was it their different behavior, or was it caused by lies? Or, without a causal view, you can just look at it as a classic survey problem - the data isn't always accurate.

Take a look at both Method of Reasoning and also Survey Flaws in the Flaw in the Reasoning material, and see if those don't help clear things up for you. Good luck!

[Khodi7531]

Ok i'm very confused on how it's B over A. Sure, the data is interpreted differently and concluded it's skepticism due to societal expectations. But isn't the issue of the argument "unrepresentative" data? I mean they call into question the conclusion coming from the potential hesitancy of survey participants of essentially telling the truth and not admitting their behavior.



So how does that not create an "unrepresentative sample"? These people are not correctly representative and isn't that what this answer is getting at?


Please help so I know before the exam on monday lol

[Adam Tyson]

An unrepresentative sample, khodi, is one where the folks you surveyed or tested do not accurately reflect the type of people that you drew conclusions about. That could be because the sample is too small, or it could be because the people surveyed or tested differed in some important way from the people we drew conclusions about. Their answers aren't the issue - who they are, or how many of them there are, is what matters in a question about whether they are representative.

If we surveyed these same 100 50-year-olds and 50 20-year-olds and drew a conclusion that "as you age, you become more likely to donate blood", THEN we might have an unrepresentative sample problem because we only looked at two specific ages. Maybe 35-year-olds donate more than 50-year-olds? Maybe 75-year-olds donate less than 50-year-olds? The two snapshots of specific ages do not adequately represent the entire range of the aging spectrum.

Here, though, we surveyed the two specific groups, and we drew conclusions about only those two groups. Unless the groups were too small, or unless these particular people being surveyed were in some important way different from typical people of their ages, then an unrepresentative sample is unlikely to be the problem.

More to the point here, the author in this stimulus did NOT point to any potential problem with the sample. He had no problem with WHO we asked about blood donations. Rather, he was concerned about what ANSWERS we got. Pointing out that the answers may be inaccurate has nothing to do with whether the group we surveyed was representative of the group we drew conclusions about!

When looking at surveys, flaws are likely to be found in one of three categories: WHO we asked (was the sample representative?), WHAT we asked (were the questions unbiased and well constructed?) and what ANSWERS we got (did people give us accurate, truthful responses?) Don't mix those up!