Test Yourself
Take this quiz and see how you do. Select the correct answers by using the drop down menus. Once you are done, click submit to see your score! The ones that you get wrong will have a check!
1. If you have a random sample, the relationship you find in your sample will hold in the population.
-- Choose an answer --
a. True
b. False
2. Which of the following correlations is more likely to be statistically significant?
-- Choose an answer --
a. A correlation of -.70
b. A correlation of +.40
3. You find a correlation of .10. That correlation is most likely to be significant if your correlation is based on ____ observations.
-- Choose an answer --
a. 100
b. 1000
4. You can conduct a
t
test on correlational data.
-- Choose an answer --
a. True
b. False
5. The median split is a common way of analyzing correlational data.
-- Choose an answer --
a. True
b. False
6. The median of the following the following numbers1, 2, 12, 3, 100 is 12.
-- Choose an answer --
a. True
b. False
7. The median split has less power than a test that determines whether a correlational coefficient is greater than zero.
-- Choose an answer --
a. True
b. False
8. A multiple group ANOVA on correlational data allows you to have more power than a t test on those data.
-- Choose an answer --
a. True
b. False
9. If you want to know whether the relationship between outside temperature and aggression is curvilinear (as temperature increases, aggression increases up to a certain point), you should use
-- Choose an answer --
a. A test of the significance of the correlation coefficient
b. Analysis of variance
10. If you want to see whether gender and extraversion predict happiness, you should use
-- Choose an answer --
a. A test of the significance of the correlation coefficient
b. Analysis of variance
11. If an ANOVA finds a significant effect for extraversion on happiness, you can conclude that extraversion causes happiness.
-- Choose an answer --
a. True
b. False
12. If you do 100 analyses using a
p
= .05 level of significance, even if the variables are unrelated, you should expect five significant results. These significant results would be Type 1 errors ("false alarms").
-- Choose an answer --
a. True
b. False
13. Maria obtains several statistically significant correlations. Her results are most likely to be significant at the .05 level in a different sample if, in her original study, she
-- Choose an answer --
a. Computed very possible correlation coefficients
b. Computed only those correlation coefficients that related to her hypothesis
14. You do a study of students with IQs between 115 and 120. You fail to find a relationship between IQ and grade-point average. The most likely reason for your failure is
-- Choose an answer --
a. A nonlinear relationship exists between IQ and grade-point average
b. Restriction of range
You got
out of
correct.
Your Score:
%