Logical reasoning PrepTest 142 · Section 2 · Question 20

Incorrect. Even if the cardiologists made few obvious mistakes, they could have made more than the computer program. They also could have made a large number of non-obvious mistakes that the computer program didn't. This answer doesn't undercut the conclusion that the computer program is better at reading EKGs.

Incorrect. While the practice of medicine overall might be an art rather than a science, the argument and conclusion are just about diagnosing heart attacks using an EKG, and there's a right and wrong answer there.

Correct. Argument or Facts:
Argument

Valid or Flawed:
Flawed

Question Type:
Weaken

Stimulus Summary:
A computer program took on a cardiologist for who could interpret EKG data better. The program correctly diagnosed more incidents that were later confirmed as heart attacks than the doctor, so EKG interpretation should be left to the machines.

Answer Anticipation:
When diagnosing something, there are two relevant considerations. The first is how many times you correctly diagnose the disease—in this case, a heart attack. The computer program won that hands down ("correctly diagnosed a significantly higher proportion").

However, the other consideration—equally important—is whether when the patient didn't have a heart attack if that was correctly diagnosed. There's no information about who was better at diagnosing someone who wasn't having that heart attack.

To bring up an analogous situation, think about someone who always calls "Heads" in a coin flip. They will be right 100% of the time that the coin comes up heads. However, that doesn't mean that they're the best at predicting coin flips—they're wrong every time it comes up tails!

Same here—imagine a computer program that always says someone is suffering a heart attack. That program would always correctly diagnose a heart attack. It would also always incorrectly diagnose someone who wasn't having a heart attack, and thus it probably wouldn't be a great program!

To weaken this argument, therefore, the correct answer will likely establish either the doctors were better at determining when someone wasn't having a heart attack, or the computer program was worse at that.

Answer Explanation:
This answer reflects the other side of the "competition"—correctly diagnosing no heart attack. If the cardiologist was better at that, then we'd have to dig into the numbers and which diagnosis is more important to figure out which is better—weakening the conclusion that it was the computer program.

Key Takeaway:
First, accuracy isn't just about true positives (here, saying there was a heart attack when there was), but also true negatives (here, saying there was no heart attack when there wasn't). Whenever comparing accuracy, both of these considerations need to be established.

Second, "neutral"-language answer choices in Weaken and Strengthen questions are generally wrong because their impact on the argument changes based on which "direction" the answer choice goes in reality. For example, if the answer establishes that something has an effect on blood magnesium levels, without knowing if that effect is an increase or decrease, it's probably impossible to know if that strengthens or weakens the argument. Look out for neutral language like that as trap answers in Strengthen and Weaken questions.

Incorrect. For the cases where there is sufficient information, however, either programs or cardiologists would be more accurate, and that's the group that we want interpreting our EKG. More generally, this answer applies equally to both groups, and the conclusion is comparative (one thing is better than another), so a similarity is unlikely to change that conclusion.

Incorrect. These "neutral" answer choices—ones that don't give you enough information to know which direction it goes in—are almost always wrong. Was this cardiologist more experienced and skilled than the average, or less? If she were more experienced and skilled, then the argument is strengthened; if less, weakened. "Neutral" answers generally cut in different directions depending on which direction the neutral language goes, so they don't strengthen or weaken the argument.