Can brainwaves be used to measure job aptitude? During the 1980s, a psychologist

Question

Can brainwaves be used to measure job aptitude? During the 1980s, a psychologist tried to use brainwave measurements to identify U.S. Navy recruits who are good at using a rifle (instead of simply watching them shoot). Several times each year, batches of data were analyzed. Each batch of data compared various brainwave measurements with rifle–shooting performance records for a group of recruits. The psychologist was looking for evidence of a relationship between brainwave measures and shooting skill. During 1990, he analyzed 34 of these batches. In 32 of those batches, there was no evidence of association, but the other two batches were statistically significant (at level of significance = .05) in demonstrating a relationship between brainwaves and shooting skill. Based on this, which of the following is true?

The psychologist should toss out the 32 batches of data that failed to demonstrate an association between brainwaves and shooting skill, and keep the two batches that do demonstrate an association. After all, the grant that funds such research and pays the psychologist's salary needs to be renewed.

There's little evidence of an association. After all, since we're testing each batch of data at the = .05 level of significance, by chance alone 1 in 20 batches would demonstrate an association even if there really isn't one. This explains the two “good” batches the psychologist observed.

The psychologist should just pick one of the data sets (at random) and base his analysis on that.

All answers are correct.

Explanation / Answer

Option B explains it well:

OPTION B: There's little evidence of an association. After all, since we're testing each batch of data at the = .05 level of significance, by chance alone 1 in 20 batches would demonstrate an association even if there really isn't one. This explains the two “good” batches the psychologist observed. [ANSWER, B]

***************

Note that alpha = 0.05 means the probability of a type I error, that is, rejecting a true Ho. So that explains why a few batches showed an association.

Related Questions

Navigate

• Browse (All)
• Browse C
• Subjects

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.