Soc 201 - Final Exam

 

 

 

 

Person/Gender Age to nearest birthday Social Class Income Prejudice Score Weight
Joe/M 3 yrs Lo $2 10 2 lbs
Bill/M 2 Lo 4 7 7
Sam/M 3 Hi 2 14 3
Sue/F 2 Hi 4 2 2
Kay/F 3 Hi 2 3 5
Ben/M 2 Lo 4 9 4
Jill/F 3 Lo 2 5 3
May/F 2 Lo 4 6 5
Ken/M 3 Hi 2 11 7

 

 

 

Use the above data for all questions and problems on this exam which need data. You may assume this is a random sample from a population which is normally distributed with respect to all interval variables, and that any interval variable plotted against any other interval variable will give a linear relationship and homoscedasticity and that all categories in all nominal or ordinal variables have similar variances with respect to all interval variables.

Be sure to work all 20 problems!

 

 

_____________________________

Name

 

If you are willing to tutor in this class next semester, put your phone number below. I will distribute names and phone numbers of willing tutors, it is up to the student and tutor to agree on hours and prices.

 

__________________________

Phone

 

1. Why or why not use the Pearsonian R to answer this question: "Would you expect to find the association between age and income in the whole population from which this sample was drawn?"

a. Yes, it is the ideal statistic for this question and these data.

b. No, because you do not have a linear relationship.

c. Yes, it is OK, but it is wasteful to treat age as ordinal.

d. No, because you need two interval variables.

e. No, because you need an N of 10 or more when there are ties in the data.

f. None of the above.

2. Why or why not use the asymmetrical lambda to answer this question: "What is the association between prejudice score and social class?"

a. Yes, it is the ideal statistic for this question and these data.

b. Yes, except it is wasteful to treat these variables as nominal.

c. No, because prejudice score is not a nominal variable.

d. No, because this is not an inferential statistic.

e. No, because you do not know which variable is independent and which is dependent.

f. None of the above.

3. Why or why not use the phi coefficient to answer this question: "What is the relationship between age and social class?"

a. Yes, it is the ideal statistic for this question and these data.

b. Yes, it is OK but wasteful to treat age as a nominal variable.

c. Yes, it is OK except it is wasteful to treat social class as nominal.

d. No, because the fe need to be 5 or larger.

e. No, because you need to know which variable is independent and which is dependent.

f. None of the above.

 

4. Why or why not use partial tau to answer this question: "What is the relationship between social class and prejudice score with age held constant?"

a. Yes, it is the ideal statistic for this question and these data.

b. Yes, except it is wasteful to treat variable #3 as nominal.

c. Yes, except it is wasteful to treat variables #1 and 2 as interval.

d. Yes, except it is wasteful to treat variable #3 as interval.

e. No, because you do not need a third variable which is held constant.

f. None of the above.

5. Why or why not use Fisher’s F-test to answer this question: "Can you generalize the relationship between income and social class to the whole population from which this sample was drawn?"

a. No, because you need 5 or more cases in each nominal category.

b. No, because you need an N of 10 or more.

c. No, because social class is not interval.

d. Yes, except it is wasteful to treat social class as nominal.

e. Yes, except it is wasteful to treat income as nominal.

f. None of the above.

6. Why or why not use Kendall’s Tau to answer this question: "What is the association between age and prejudice score in this sample?"

a. No, because you must have two ordinal variables.

b. Yes, it is ideal for this question and these data.

c. No, because you need an N of 10 or more.

d. No, because you do not have linear relationship or homoscedasticity.

e. No, because you do not have a third variable which is held constant.

f. None of the above.

 

7. Why or why not use the test of significance for tau to answer this question: "Would you expect to find the association between social class and prejudice score in the whole population?"

a. No, because this is not an inferential statistic.

b. No, because this is not an inferential question.

c. No, because you need an N of 10 or more.

d. Yes, it is the ideal statistic for this question and these data.

e. Yes, except it is wasteful to treat prejudice as ordinal.

f. None of the above.

8. Why or why not use partial r to answer this question: "What is the relationship between income and age with social class held constant?"

a. Yes, it is the ideal statistic for this question and these data.

b. Yes, it is OK except it is wasteful to treat variable #3 as nominal.

c. Yes, it is OK except it is wasteful to treat variables #1 and 2 as interval.

d. Yes it is OK except it is wasteful to treat variable #3 as interval.

e. No, because you do not have a third variable which is held constant.

f. None of the above.

9. Why or why not use the test of significance for r to answer this question: "Would you expect to find the association between prejudice score and income in the whole population from which this sample was drawn?"

a. Yes, it is the ideal statistic for this question and these data.

b. No, because you need and N of 10 or larger.

c. No, because you need 5 or more cases in each nominal category.

d. No, because this is not an inferential statistic.

e. Yes, except it is wasteful to treat prejudice as interval.

f. None of the above.

 

l0. Why or why not use eta to answer this question: "What is the relationship between age and prejudice score?"

a. Yes, it is the ideal statistic for this question and these data.

b. No, because prejudice score is not interval.

c. Yes, except it is wasteful to treat age as nominal.

d. No, because you do not have linear relationship and homoscedasticity.

e. Yes, it is OK except it is wasteful to treat prejudice score as nominal.

f. None of the above.

 

 

11. How would you interpret a tau of -.60 for the variables of social class and prejudice score?

a. 36% of the variance in social class can be explained by prejudice.

b. Moderate or little larger than moderate negative relationship between prejudice score and social class.

c. In 60% of the comparisons made, persons in different social classes showed systematic differences in their prejudice scores.

d. 60% more disagreement than agreement in the rank order of social class and prejudice score.

e. Tau cannot be negative; impossible answer.

f. None of the above.

12. If you computed an r of -.834 for income and age, what would be the results of the test of significance for r for income and age?

a. p is greater than .05

b. p = .05

c. p is less than .05

d. p = .01

e. p is less than .01

13. How would you interpret an F of 5.99 for the variables of income and social class?

a. You cannot generalize from the sample to the whole population.

b. p is less the .01

c. p = .05

d. The chances of being wrong if you generalize from the sample to the whole population are less than 5 in 100.

e. The chances of being wrong if you generalize from the sample to the whole population are exactly 1 in 100.

f. None of the above.

14. How would you interpret an r of -.50 for the variables of income and age. (You need to compute r2 in order to know all interpretations of r.)

a. Moderately small relationship between age and income.

b. 50% reduction in error when predicting age from income.

c. 25% of the variance in income cannot be explained by age.

d. 50% of the variance in age can be explained by income.

e. 25% of the variance in age can be explained by income.

f. None of the above.

 

 

 

 

 

 

 

 

 

 

 

 

 

15. Rearrange your data in the proper form and compute Pearsons R for weight and income. Can you generalize this to the whole population from which the sample was drawn? Why or why not?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

16. Compute partial gamma for social class and weight controlling for income.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

17. Compute eta for gender and weight. Explain all possible interpretations.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

18. Predict the age of a person whose income is $1.50 (Show all computations. Correct guesses will not count.)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

19. Make a frequency distribution for weight, using categories with a width of one year. Then compute the median weight, using the data you just arranged. Be very clear about what you think is the median by completing this sentence: "The median weight is

_________."(Do NOT compute the median on the raw or ungrouped data; compute it on the frequency distribution.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20. Compute the mean and standard deviation for weight. You can use the frequency distribution from 19.