Soc 201 - Second Exam
Person | Favorite Food | Weight | Exercise Speed |
Ann | Pizza | 120 | Fast |
Kay | Cake | 160 | Slow |
Bill | Pizza | 140 | Very Fast |
Joe | Pizza | 140 | Slow |
Dan | Pizza | 160 | Fast |
Dave | Cake | 120 | Fast |
Sue | Pizza | 120 | Very Fast |
Mel | Cake | 140 | Slow |
Use the above data to answer all problems on this exam which require
data to answer them (except 12 - 15). You may assume this is a
random sample from a larger population. Show all computations
on the last 5 problems; just giving the answer without showing
how will not count. Be sure to answer all 15 problems.
1. Why or why not use Gamma to answer this question: "What
is the relationship between weight and exercise speed in this
sample?".
a. Yes you can; this is the ideal statistic.
b. Yes you can except it is wasteful to treat weight as ordinal.
c. No you cannot because Gamma is not inferential
d. No you cannot because the variables are not dichotomous and
discrete
e. No you cannot because you do not know which variable is independent
and which is dependent.
2. Why or why not use Chi Square to answer this question: "Can
you generalize the association between food and weight to the
whole population?"
a. No you cannot because Chi Square does not measure association.
b. Yes you can except it is wasteful to treat weight as nominal.
c. Yes you can, it is the ideal statistic.
d. No you cannot because the fe must be 5 or larger.
e. No you cannot because weight is not nominal.
3. Why or why not use the two sample T-test of means to answer
this question, "Can you generalize the difference in average
weights of males and females to the whole population from which
the sample was drawn?'
a. Yes, it is ideal.
b. Yes, except it is wasteful to treat weight as nominal.
c. No, because the sample totals are not 100 or more.
d. No, because the two sample T-test of means does not measure
the significant differences of average weight.
e. None of the above
4. Why or why not use the Phi Coefficient to answer this question:
"What is the relationship between exercise speed and weight.
a. No you cannot because the variables are not dichotomous and
discrete.
b. Yes you can except it is wasteful to treat these data as nominal.
c. Yes, you can this is the ideal statistic.
d. No you cannot because you need an Fe of 5 or more.
e. None of the above.
5. Why or why not asymmetrical lambda to answer this question:
"What is the relationship between favorite food and weight
when you are trying to predict weight from a knowledge of favorite
food.
a. Yes you can; it is the ideal statistic.
b. No you cannot because you do not know which variable is dependent
and which is independent.
c. No you cannot because the variables are not discrete.
d. Yes you can except it is wasteful to treat weight as nominal.
e. No you cannot because you do not have two nominal variables.
6. Interpret a Phi coefficient of -.17 for this table:
Sex | Demo | Rep | |
F | |||
M |
a. Small positive association between being female and voting
Democrat.
b. Small negative association between sex and vote.
c. 17% more disagreement than agreement if the rank order of sex
and vote.
d. p is greater than .05.
e. Small negative association between being female and voting
Democrat.
7. How would you interpret a Chi Square of 26.22 for a table with
3 rows and 8 columns?
a. The chances of being wrong if you generalize from the sample
to the whole population are less than 1 in 100.
b. p = .05
c. p = .01
d. p is less than 5 in 100.
e. You cannot generalize from the sample to the whole population.
8. How would you interpret a Gamma of -.82 for weight and exercise
speed?
a. 82% more agreement than disagreement in the rank order of weight
and exercise.
b. p is greater than .05.
c. Large positive relationship between weight and exercise speed.
d. 82% more disagreement than agreement in the rank order of exercise
speed and weight.
e. 82% improvement when predicting both variables simultaneously
from a knowledge of each other.
9. How would you interpret an asymmetrical lambda of .35 when
weight is dependent and exercise speed is independent?
a. Moderately small negative relationship between weight and exercise.
b. 35% improvement when you are trying to predict both variables
simultaneously from a knowledge of each other.
c. 35% more agreement than disagreement in the rank order of weight
and exercise.
d. 35% improvement when predicting weight from a knowledge of
exercise.
e. 35% improvement when predicting exercise speed from a knowledge
of weight.
10. How would you interpret the result of 2.63 from a 2 sample
t-test of means for the hypothesis, "Males are heavier than
females among the whole population from which this sample was
drawn."
a. You cannot generalize from the sample to the population.
b. The chance of being wrong if you generalize is 5%.
c. The chance of being wrong if you generalize is less than 5%.
d. The chance of being wrong if you generalize is 1%.
e. The chance of being wrong if you generalize is less than 1%.
11. Arrange the data in the proper form and compute gamma for
the relationship between weight and exercise speed.
12 Compute Chi square for the following table:
Food | Very Fast | Fast | Slow | row totals |
Pizza | 11 | 15 | 9 | 35 |
Cake | 21 | 32 | 12 | 65 |
column totals | 32 | 47 | 21 | 100 |
13. Combine categories and collapse the following table in an effort to make it meet the assumptions for Chi Square. Do not collapse it anymore than is absolutely necessary. Do not compute Chi Square just collapse.
Party | 10 | 7 | 6 | 5 | 4 | row total |
Demo | 16 | |||||
Rep | 40 | |||||
Peace | 4 | |||||
Comm | 35 | |||||
Ind | 5 | |||||
col. totals | 32 | 10 | 20 | 18 | 20 | 100 |
14 & 15. Using the table below compute the mean and standard
deviation. Then, using the one sample T-test of means, see if
it is legitimate to generalize to the whole SDSU population. is
22.3 (the mean of the whole population)
Age to nearest birthday | f |
19-20 | 28 |
21-22 | 35 |
23-24 | 22 |
25-26 | 15 |
extra credit: do a 95% Confidence Interval on the above data's mean.