## Wednesday, April 22, 2009

### Psychology GRE Study Guide Page 12

I’m too lazy to link all of the previous pages, but you can find them in the archives and under the label “GRE”. Page 1 has a link to the official Practice GRE from which this guide is developed.

80. Correlations can help researchers predict how behaviours occur, but they can not determine causality, why two factors occur together.

a. Teachers have nothing to do with the experiment (if their grading scheme was a factor, that might play into the specifics of the explanation).

b. If students uniformly underestimated the duration of their sleep, the correlation would still be the same as students with more sleep would perform better than students with less sleep. That’s all the results are looking at, more vs. less, not a specific number of hours of sleep.

c. A perfect correlation would have a value of 1.0. See GRE Study Guide pages 10 and 11 for more on correlations.

d. Hours of study would be a third variable, a link between cause and effect. For For example, perhaps students who sleep more are more rested and able to study more which might, though not necessarily, increase their grades. But it’s still related to sleep.

e. There might be a predisposition for students who perform better to sleep more which thereby makes the sample population less random.

81.

a. Having an outside party repeat a measurement can be a good way of checking someone else’s measurement.

b. This could help, but it’s not a very reliable source of information. Just because a person strongly believes they had six hours of sleep, doesn’t mean they actually did. Similarly, a student may be unsure about their sleep (they may even be extremely unconfident in all aspects of their life), when their estimation is quite accurate.

c. In two weeks, the situation may be different for some, or all, of the students. Some females may estimate longer hours of sleep than previously, but only because they are in their menstrual cycle. Yet, the researcher may label that woman an unreliable estimator when in fact she knows exactly when she sleeps. So, just because the duration of a person’s sleep changes, that doesn’t mean their ability to judge how much sleep they got is any less reliable.

d. If it was known that males/females were better at estimating hours of sleep, than this could be used. But there is no evidence that that gender difference is true.

e. If there was a reliable correlation, all this would tell you is that older students tend to sleep more, for example. As above, there is no reason to believe age difference would result in estimation differences.

82.

a. A perfect correlation with a coefficient of 1 implies a linear relationship. Since the correlation is less than 1, the relationship will be non-linear.

b. This is the natural bell curve of academia. Averages of the measurements were taken and we are told that the population is a random sample which is exactly the opposite of a biased sample.

c. The difference between the maximum and minimum values means nothing on its own. How do you know it’s an insignificant difference? The standard deviation can provide this information; if the standard deviation is greater than the calculated difference, the values are not significant. But here, the sample size is large, so the standard deviation is likely to be small since the standard deviation is the average of the measurements divided by the square root of the number of subjects. For another explanation and example of standard deviation, see GRE Study Guide page 8.

d. The size of sample doesn’t tell you much on its own, either. An educated guess would be that the error/standard deviation of the measurement is likely smaller than in the groups with fewer subjects, but even this is not necessarily true. But since averages have been calculated, and we were told in the introduction of the question that the results were reliable (the values are significant even with the consideration of errors), the number of subjects in each category does not affect the results.

e. Again, we don’t know what the error bars are, but from a quick glance at the numbers, the standard deviation is going to be greater than 0.1, which means the A and B group get the same hours of sleep. Also, there is no reason to assume that the relationship will be linear; maybe getting too much sleep results in lower grades.

Subscribe to:
Post Comments (Atom)

Thanks for great sharing post. Its very comprehensive information for everyone,This tips is very useful for me and I will share it to my friends keep posting GRE Study Plan

ReplyDelete