|
Statistics Take Home Test 4 341_test4b.htm updated 4/18/22 11:25:05 AM |
Michael Hallstone, Ph.D. hallston@hawaii.edu |
Take Home Test 4 for Statistics
Please cut and paste this exam into a word processor such as
Microsoft Word and type your answers and paste your SPSS output in proper
spots. Please do NOT send a pdf due to formatting issues. Please delete the coversheet and any “empty”
problems that you do not answer -- I will send it back if you do not do this.
Since you can lead a student
to water but not make it read here’s a short video on the specifics. You
should watch it as it shows you great hints!
This is due on the date and time specified on the course
schedule (see course syllabus for penalties for late submissions). Keep a copy of this exam for your
protection!
The use of
SPSS for computations is mandatory.
If you do your computations by any other method you will receive zero
points on the problem(s).
In plain English: do NOT do your “step 6” or “TR” calculations by hand or by excel or any other method except SPSS! If you do you will receive a zero on that problem. (Now if you wish to do those calculations on a separate scratch paper so you can “learn” what the SPSS numbers mean, that is fine. This is why I show you the math of step 6 in the lectures. But do NOT include those hand calculations on the test!) Thus, for the “step 6” or “TR” calculations you simply report the number that SPSS gives you like this “SPSS says TR=___”.
Exam 4 problems below
· If you are sending drafts online please do not include “empty” problems. Just send me the problem[s] you wish me to comment upon. It makes our “scrolling life” on computer screens cumbersome.
· Please do not include the cover sheet.
· Please do not edit out my hints inside of the test problems.
· Please type and place your spss in the places indicated
1. The whole exam is worth 100 points and
this question is worth 20 points. In
this question you will do a TWO-sample hypothesis test of means using one of
the variables from your study. This
material is covered in lecture 18. Use an a (or “alpha”) = .05. (You may do
either a one or two tailed test – use the one that makes the most interesting
question to you.)
State the question you are attempting to answer with your hypothesis test in plain English. Hint: you “test the theory” of what is in Ho and “try to prove” what is in H1. For the plain English cut and paste from my examples either from lecture or practice problems and modify it to fit your variables! (5 points)
[Type
your answer here]
Show the 7 steps for this test in your answer. By the way you do not have to draw the picture of the rejection regions in step 4. If you get step 5 correct then I will be able to figure out what your picture would have looked like! (8 points)
[Insert 7 steps and SPSS output by step 6 here]
For Step #6
· report your TR value from SPSS and indicate which of the formulas you used to compute your TR value – for example: “ I used formula #1.”
The top number in SPSS
is Formula 1 (procedure 4 from text and lecture)
The bottom number in SPSS
is Formula 2 (procedure 3 from text and lecture)
[Tell me what formula you used here]
If you did not do so in step #7 above already, make sure you explain what your answer means in plain English! HINT: For the plain English paste from my examples either from lecture or practice problems and modify it to fit your variables! (5 points)
Mention the assumptions you have violated (if any). You
must look at the list of assumptions for this problem from the very top of the
corresponding the lecture. Do not
simply cut and paste them here. Tell me
what you violated and what you assumed (but do not know you violated. HINT: all of you will violate one
assumption for sure and may need to “assume” something about your TWO
populations, depending upon your sample sizes from each of these populations. For
one part of this answer if you say your “test variable is normally
distributed in the two populations” you will receive zero points. See the list of assumptions at the top of the
lecture for a better answer and proper wording. (2 points)
Total out of 20 points ______
2. The whole exam is worth 100 points and
this question is worth 20 points. In
this question you will perform a
one-way ANOVA test using two variables from your study. This material is covered in lecture 19. Use an a (or “alpha”) = .05.
State the question you are attempting to answer in plain English. Hint: you “test the theory” of what is in Ho and “try to prove” what is in H1. For the plain English cut and paste from my examples either from lecture or practice problems and modify it to fit your variables! (5 points)
[Type your answer here]
Show the 7 steps for this test in your answer. By the way you do not have to draw the picture of the rejection regions in step 4. If you get step 5 correct then I will be able to figure out what your picture would have looked like! (6 points)
[Insert 7 steps and SPSS output by step 6 here]
If you did not do so in step #7 above already, make sure you explain what your answer means in plain English! HINT: For the plain English paste from my examples either from lecture or practice problems and modify it to fit your variables! (7 points)
Mention the assumptions you have violated (if any). You must look at the list of assumptions
for this problem from the very top of the corresponding the lecture. Do not simply cut and paste them here. Tell me what you violated and what you
assumed (but do not know you violated.
HINT: all of you will violate one assumption for sure and “assume”
something about your multiple populations. For one part
of this answer if you say your “test variable is normally distributed in the
populations” you will receive zero points. See the list of assumptions at the top of the
lecture for a better answer and proper wording. (2 points)
Total out of 20 points ______
3.
The
whole exam is worth 100 points and this question is worth 20 points.
Perform the seven steps for a Contingency Table Chi-Square test using two
of the variables from your study. This material is covered in the second kind
of Chi Square in lecture 20. Use an a (or
“alpha”) = .05.
State the question you are attempting to answer with your test in plain English. Hint: you “test the theory” of what is in Ho and “try to prove” what is in H1. For the plain English cut and paste from my examples either from lecture or practice problems and modify it to fit your variables! (5 points)
[Type your answer here]
Show the 7 steps for this test in your answer. By the way you do not have to draw the picture of the rejection regions in step 4. If you get step 5 correct then I will be able to figure out what your picture would have looked like! (6 points)
[Insert 7 steps and SPSS output by step 6 here]
If you did not do so in step #7 above already, make sure you explain what your answer means in plain English! HINT: For the plain English paste from my examples either from lecture or practice problems and modify it to fit your variables! (7 points)
Mention the assumptions you have violated (if any). You
must look at the list of assumptions for this problem from the lecture. Do
not simply cut and paste them here. Tell
me what you violated and if you had to assume anything. HINT: all will violate
one assumption regarding the type of sample and may have to look at your
expected counts. (2 points)
Total out of 20 points ______
4.
The
whole exam is worth 100 points and this question is worth 20 points.
Perform a simple linear regression test using two of the variables from
your study. This material is covered in in lecture 21; you’ll find lecture 21c
especially helpful for this problem. Use an a (or
“alpha”) = .05.
Provide the SPSS
regression output for your two variables.
(insert SPSS
output here)
Is your regression
significant and why (2 points)? (Note
you do NOT have to do 7 steps to tell me whether or not it is significant! You may use the p-value method if desired –
it is much easier.)
(insert answer
here)
Write out the equation
for the regression line. (2 points)
(insert answer
here)
What is the r2? (1 point)
(insert answer
here)
Assume you have significant results (even if you do not) and explain what the r2 means in plain English (10 points). Hint: if your results are not significant r2 is actually meaningless so say the following, “If my regression were significant my r2 would mean ________but actually r2 is meaningless.”
(insert answer
here)
Correlation
Provide the SPSS
correlation output for your two variables.
(insert SPSS
output here)
Is your correlation
significant and why (1 points)? (Note
you do NOT have to do 7 steps to tell me whether or not it is significant! You may use the p-value method if desired –
it is much easier.)
(insert answer here)
What is your r (or coefficient of correlation)? (1 points)
(insert answer
here)
Assume you have significant results (even if
you do not) and explain
what the r means in plain English (3
points). Hint: if your results are not significant r is actually
meaningless so say the following, “If my correlation were significant my r
would mean ________but actually r is meaningless.”
Total out of 20 points ______
(insert answer
here)
5.
The whole
exam is worth 100 points and this question is worth 20 points. Perform
a multiple regression test using the variables below from your study. This
material is covered in in lecture 22; you’ll find lecture 22c especially
helpful for this problem. Use an a (or
“alpha”) = .05.
You must use the following variables unless you receive emailed permission from me to use other variables. It is okay to use others, but we need to talk about it first.
Dependent
variable (y) – monthly income in exact numbers (must be ratio level not coded
into categories)
Independent
variables (x)
·
x1 = age
this should be age in exact years (must be ratio level not coded into
categories)
·
x2 = gender (female=1
male=0)
Provide frequency distributions
from SPSS for gender. . Make sure female is coded 1 and male is coded 0.
Provide two versions of your frequency
distributions of gender as I do below so I know you coded gender with females
=1 correctly.
(insert four
frequency distribution tables from SPSS here)
Provide the SPSS
multiple regression output using income as dependent variables and age and
gender as independent variables.
(insert SPSS
multiple regression output here)
Is your overall
multiple regression significant model and why (2 points)? (Here you MUST use the p-value method.)
(insert answer
here)
Write out the equation for the multiple regression line. (3 points) Recall, y-hat= a + b1x1 + b2x2 (where a = y intercept, b1= slope for age, and b2=slope for gender)
y-hat= _____ + _____x1
+ _____x2
(insert y intercept,
slopes or unstandardized coefficients from SPSS in the spaces)
Interpret the unstandardized coefficient or slope of the ratio level independent variable age on the dependent variable monthly income. (5 points) HINT: if your results are not significant say the following, “If this variable were significant the unstandardized coefficient or slope would mean ________but actually this coefficient is meaningless.”
Interpret the unstandardized coefficient or slope of the
“dummy” independent variable gender on the dependent variable monthly income.
(5 points) HINT: if your results are not significant say the following, “If
this variable were significant the unstandardized coefficient or slope would mean
________but actually this coefficient is meaningless.”
Interpret the standardized coefficient or “Beta” of either
age or gender on the dependent variable monthly income. It does not matter which you choose. (5
points) HINT: if your results are not significant say the following, “If
this variable were significant the standardized coefficient would mean
________but actually this coefficient is meaningless.”
Total out of 20 points ______