Stats Exam

profileMaydl
APP516-18Final1.xlsx

Q. 1 Picking tests

There are seven questions. To complete the test, you will need to generate two SPSS data sets,
one for Question 4 and another for Questions 5-7, by importing the Excel data provided.
Please return to me (AR) by e-mail this completed Excel file and the two SPSS OUTPUT files you generate
answering questions 4 and 5-7. (I don't want the data files.) To expedite grading , please delete extraneous
output from your SPSS files and label your results to identify which questions are being addressed.
Question 1: Pick the best test (15 pts.) Score: 0
State what test you would use to answer each of the following questions, and
explain why by identifying the types of variables involved (nominal, ordinal, or
scale), which variables are dependent and which independent, and the assumptions of your test.
(There may be more than one right answer; make sure your list of assumptions is
consistent with your identification of variables.)
a. The difference between 8 privately funded shelters and 9 publicly funded shelters in
the mean annual expenditures on pre-adoption training of dogs.
Type of Variables:
Independent/Dependent:
Test:
Assumptions:
b. The relationship between hours spent in adoption counseling and the number of
failed adoptions in one year for 22 similarly-sized shelters.
Type of Variables:
Independent/Dependent:
Test:
Assumptions:
c. The improvement in obedience test performances (graded 1-100) before and after a six-week
obedience training class in 15 dogs.
Type of Variables:
Independent/Dependent:
Test:
Assumptions:
1 pt. bonus: What control group would you devise to improve the experiment?
d. Differences in belief (on a 1-7 Likert scale) in how much knowledge of wolf biology
one has among 30 people evenly distributed among 3 different geographic regions.
Type of Variables:
Independent/Dependent:
Test:
Assumptions:
e. The association between presence or absence of a smoker and the presence or absence of
chronic respiratory disease in resident cats in 150 homes.
Type of Variables:
Independent/Dependent:
Test:
Assumptions:
0

Q. 2 Short answer

Question 2. Short Answers (12 pts.) Score: 0
Answer the following questions in one or two sentences.
a. What is "power," and why is it important to estimate power before designing your study?
b. Name four things that affect the power of a study design.
c. What is the primary limitation of observational studies?
d. Why should we use ANOVA and multiple comparison tests instead of multiple t-tests
when looking for differences between means of three or more groups?
e. In a simple one-way ANOVA, what part of the F statistic represents the signal, and what part represents the noise?
f. Describe a hypothetical study for which it would make more sense to set alpha = 0.01 rather than alpha = 0.01,
and explain why.
0

Q.3 Foaling

Question 3. Wild horse interlude (14 pts.) Score: 0
Wild horses were gathered by helicopter in February, two months before the mating season begins.
240 mares were randomly assigned to one of three treatment groups.
81 mares were vaccinated with contraceptive treatment 1; 94 received contraceptive treatment 2;
and 64 mares received a sham control. Eighteen months later (horses have an 11-month gestation period),
the number of mares in each group who had foals was counted.
Here are the results:
# Mares: Vaccine 1 Vaccine 2 Control Total
Foaling 13 7 38
Not foaling 68 87 26
Total
You may use whatever software you’d like to answer the questions below.
a. (1 pt.) Do the three treatment groups differ in foaling outcome? State your null and alternate hypotheses.
b. (4 pts.) Conduct the appropriate test, using any software you choose, and paste the results below.
c. (1 pt.) Setting alpha = 0.05, state your conclusion.
d. (1 pt.) Is there a difference between Treatment 1 and Treatment 2? State your null and alternate hypotheses.
e. (4 pts.) Conduct the appropriate test, and paste the results below.
f. (1 pt.) Setting alpha = 0.05, state your conclusion. (No need to compensate for multiple tests.)
g. (2 pts.) If vaccine 1 costs $25 a dose and vaccine 2 costs $150/dose, which should we recommend that people use and why?
0

Q. 4 Dog dentals

Question 4. Token dog question (SPSS lite!) (14 pts.) Score: 0
Reducing the buildup of “calculus,” also called “tartar” in toothpaste ads, is good for a dog’s health. Calculus was
measured by an index which combined both proportion of teeth covered and the thickness of the deposit.
Under the next spreadsheet, "Q.4 Data," are the results of an experiment to test whether different coatings
on dry dog food affected calculus build-up.
Groups of dogs were fed three different dry food diets:
1) Pellets coated with P2O7
2) Pellets coated with HMP (sodium hexametaphosphate)
3) Uncoated pellets, which served as a control
Import the data into SPSS, and then answer the following questions:
a. (2 pts.) Provide a set of descriptive statistics and a boxplot showing each group.
b. (2 pts.) Using the descriptive statistics to defend your decision, explain what test you will use to
determine whether the three groups differ and how these data fit the assumptions of the test.
c. (1 pt.) State the null and alternate hypotheses
d. (4 pts.) Perform the test.
e. (1 pt.) Setting alpha = 0.05, state your conclusion.
f. (4 pts.) Which groups are different from which, if any? Conduct the appropriate tests,, and explain your conclusion.
0

Q.4 Dog Data

Dog Calculus Treatment
1 0.49 Control
2 1.05 Control
3 0.79 Control
4 1.35 Control
5 0.55 Control
6 1.36 Control
7 1.55 Control
8 1.66 Control
9 1 Control
10 0.34 P2O7
11 0.76 P2O7
12 0.45 P2O7
13 0.69 P2O7
14 0.87 P2O7
15 0.94 P2O7
16 0.22 P2O7
17 1.07 P2O7
18 1.38 P2O7
19 0.34 HMP
20 0.05 HMP
21 0.53 HMP
22 0.19 HMP
23 0.28 HMP
24 0.45 HMP
25 0.71 HMP
26 0.95 HMP

Q. 5

For questions 5-7, find the Q5-7 Hastings Deer tab and import the data into an SPSS file,
making sure that the variables are all properly classified and labeled.
Set alpha = 0.05 for all tests
For these questions, do not paste your SPSS output in this Excel document. Instead submit your SPSS output file
and label it with the question numbers.
Question 5. (15 pts.) Deer are difficult to weigh in the field, we would like to determine whether something we can Score: 0
measure with a tape measure instead is a good predictor of body weight. “Hastings deer” contains body weight
measures along with measures of chest girth, body length, and hindfoot length.
a. (4 pts.) Which one of the three linear measures explains the most variability in body weight?
Explain how you decided, referring to your SPSS output.
b. (3 pts.) In SPSS, graph the relationship between that measure and body weight.
c. (3 pts.) Now, do the simple linear regression, and write the regression equation below:
d. (1 pt.) Using the regression equation in (c), what is the predicted body weight for a linear measure of 135 cm?
e. Is the regression statistically significant?
i. (1 pt.) State the null and alternative hypotheses:
ii. (2 pts.) State your statistical conclusion based on your SPSS output.
g. (1 pt.) How do you feel about us using this linear measure of body weight as a surrogate measure of weight?
(Yes, this is a judgment call. Make it, and defend it.)
0

Q. 6

Question 6. (17 pts.) We’d be very reassured if our body condition scores, which are subjective, Score 0
corresponded to something we’ve actually measured. Is there a significant difference between
the two body score classes in any of the three body measurements listed below?
a. (6 pts.) Use “Explore” and other descriptive statistics as needed to characterize the three variables (separating by body
condition score) and fill in the following table to identify and explain what tests you’d use to
find these differences. Refer to your SPSS findings to support your reasons for choosing that test.
.
Variable Test Explanation
Body Weight
Body Length
Chest Girth
b. (4 pts.) Carry out the tests.
c. (2 pts.) State your findings, documented with statistical results.
Is there objective support for our subjective body condition scoring system?
d. (4 pts.) If you used a parametric test to answer the questions in (b), use the non-parametric tests now; or vice-versa.
e. (1 pt.) Do the parametric and non-parametric tests yield the same conclusions or different conclusions?
0

Q.7

Question 7. (13 pts.) Finally, we’d like to know if there are differences in the measurements of Score: 0
deer size and body condition across the three years of the study.
a. (4 pts.) Use “Explore” and other descriptive statistics as needed to characterize the 3 variables below
(separating by year) and fill in the following table to identify and explain what tests you’d use to find these differences.
Variable Test Explanation
Body Weight
Body Length
Body Condition
b. (4 pts.) Carry out the appropriate tests for body weight and body length.
c. (2 pts.) State your findings, documented with statistical results.
d. (2 pts.) We’re also interested in whether body condition score changes by year. Do the test.
(Yes, it’s a different kind.)
e. (1 pts.) State your findings, documented with statistical results.
0

Q5-7 Deer Data

Animal ID Capture Year Weight (kg) Physical Condition Hind-foot Length (cm) Girth (cm) Body Length (cm)
1 2014 68 1 45.5 99 151
2 2014 54 1 47 99 144
3 2014 50 1
4 2014 45 1 46 94.5 146
5 2014 45 1 44 91 134
6 2014 50 1 44 93 135
7 2014 54 1 47 98 146
8 2014 59 1
9 2015 40 2 47 97 141
10 2015 44 2 47 102 146
11 2015 52 1 46 95 134
12 2015 61 1 46 99 137
14 2015 49 2 45 89 143
15 2015 48 1 48 94 137
16 2015 60 1 45 97 148
17 2015 80 1 49 99 152
18 2015 54 2 46 93 151
19 2015 56 2 46 89 141
20 2015 60 2 46 92 151
21 2015 2 45 91 146
22 2015 52 2 45 88 140
23 2015 2 47 91 144
24 2015 2
25 2015 40 2 44 88 128
26 2015 2
27 2015 60 2 45 96 144
28 2015 48 2 45 83 137
29 2015 60 1 48 93 149
31 2016 58 2 46.5 88.2 153
32 2016 1 43.9 88.9 139.7
33 2016 1 44.7 93.5 152.4
34 2016 2
35 2016 1 47.6 104.6 146.05
36 2016 2 43 96.4 139.5
37 2016 2 45 93.9 139
38 2016 62 2 48 91.8 146.2
39 2016 64 2 48.2 94.2 149.8
40 2016 2
41 2016 52 1 43 93.2 137.4
42 2016 2 45.5 94.5 148
43 2016 1 46 102 144
44 2016 63 1 44.4 98.04 154.94
45 2016 64 1 45.5 161
46 2016 44 2 43 92 131
47 2016 40 2 41.5 84.5 129.5
49 2016 62 1 48 101 141
50 2016 72 1 46.2 104 154
51 2016 60 2 43 96 138.2