Major Inquiry Project 1
This assignment is due at the beginning of class on Tuesday, November 1, 2011. You may work
alone or with one classmate (i.e., as a group of two) in the same section of the course. Each
group should submit one completed assignment with the names of each of the group members on
the assignment. Conversation about the assignment must be confined to the individuals in the
group. No communication between groups. The Salary Study Data Set is on the handouts page
of the course website should be used for this assignment. Use = .04 for all statistical tests.
Note: This data is hypothetical. If you want to type Greek letters , , and are a, b, and e in
lower case Symbols font.
Salary is important to many people, and salary equity is a “hot button” in many organizations.
Higher Education is not exempt from the conversation. For this study a researcher collected a
random sample of data from Business faculty at two universities that would be considered peers
(i.e., have similar mission, similar size, similar programs, and similar cost of living areas). The
major difference in the two schools is that one is Public and the other is Private. The researcher
was able to obtain the data through a Human Resource Data Sharing Consortium so the data can
be considered as accurate from the prospective of the employing university. The variables in the
data set include:
Salary Contracted salary expressed in 1000s
Public/Private Whether the university is considered public or private
Rank Faculty Rank (Assistant Professor, Associate Professor, Full Professor)
Gender Male or Female
Yrs Inst Number of years employed by the current university
Yrs Rank Number of years in the current faculty rank (possibly across multiple
Scholarship Number of articles published in career
Teaching Evaluation of teaching (on a scale of 1 to 10 with 10 being best)
Using this data, they wanted to assess the equitability of salaries. Three major theories (each
with more specific sub-theories) have been proposed. [The difficulty level of the associated
model statement is indicated by * Easy, ** Moderate, *** Difficult, **** Hardest.]
1. Can males and females expect to be paid similarly?
Theory 1a: Males make more than females.**
Theory 1b: When you control for rank, there is still a difference in salaries for males and
Theory 1c: Both gender and length of employment with the institution need to be considered
when estimating salary.*
Theory 1d: The difference between salaries for males and females depends on how long the
individual has been employed by the institution.**
2. Is quality of work and loyalty to the profession rewarded (i.e., are faculty who have stayed in
higher education or reached higher levels paid better than new comers?)
Theory 2a: Salaries will be higher for individuals who have been at the same institution
Theory 2b: Number of years at the institution, rank, and years in rank need to be considered
when estimating salary.***
Theory 2c: The impact of years at the institution depends on the rank of the individual.****
Theory 2d: Salary is dependent on quality of teaching, amount of scholarship, years at the
institution, and rank.**
3. Does it make a difference if someone works at a public or a private institution (in terms of
Theory 3a: There is no significant difference between faculty salaries at public vs. private
Theory 3b: When it comes to determining salary, teaching does not carry the same weight at
public and private institutions.*
Theory 3c: Faculty at private schools can be expected to be paid more once you control for
gender, rank, scholarship, and teaching.**
PART 1: Model statements that will direct analysis
Write model statements for each of the sub-theories—(i.e., for Theories 1a, 1b, …, 3b, 3c).
These may be typed using the correct notation or neatly handwritten. Be sure to clearly
define any new variables that you create.
PART 2: Output
Obtain output corresponding to each model statement. You do not need to obtain the
residuals. Position your output on pages so that all of the output for a single model is on the
same page—you may put output for multiple models on the same page. Be sure you’re your
output is in the same order as the model statements. Note: You do not need to submit the
PART 3: Test the theories
Use your output to test each of the theories. For example, statement 1a suggests that males
earn more than females. Is there sufficient evidence to support this statement? Be sure to
fully evaluate each statement. Your evaluation must be supported by reference to specific
numbers on the regression output. (Saying “the p-value is low enough” is not acceptable—
you must quote the specific number.) If you need to create additional output to test a theory
be sure to include this output with your submission and place this output after the output for
the theories provided.
PART 4: Find the best model (of those considered)
Compare the models to determine which model provides the best estimate for salary. Start
with all eleven models—do not eliminate any model based on the results in Step 3. Justify
all comparisons with reference to specific numbers on your output. Clearly label which
models are being compared at each step. Once you select “the best” model, write the
regression equation that is suggested by the output related to this model.
PART 5: Using your “best” model
Based on the model you selected complete the following table to show the estimated salaries
for males and females at both public and private institutions. Assume that you are looking at
individuals with teaching evaluations of 8 and 4 years in rank. Show work to justify your
Public Institutions Private Institutions
Male Female Male Female
Rank = Assistant
Scholarship = 2
Yrs Inst = 4
Rank = Associate
Scholarship = 3
Yr Inst = 10
Rank = Full
Scholarship = 5
Yr Inst = 15
PART 6: Communicating the results
Write an Executive Summary that could accompany a full report of the study. Your
Executive Summary should be no longer than two pages (single spaced Times New Roman
with double spacing between paragraphs). Assume the audience is made up of senior
administrators and faculty of the two universities. (Assume that “model,” “p value,” and any
other statistical “jargon” are NOT part of their vocabulary.) Your Executive Summary needs
to summarize how the study was conducted (in basic language), present results of the study,
and provide limitations. The table from Part 5 should be included in the Executive
Summary. The limitations should include identification of other variables that should have
been considered in the analysis. Note: Executive Summaries are intended to provide busy
people with the essential facts from the analysis in a clear and concise manner—often this is
the only part of the report read by the some of the people making decisions that relate to the
PART 7: Assemble your assignment for submission (NOTE: No Cover Page)
Page 1: Type the names of the people submitting the assignment, type the definition of any
new variables used in the model statements, and provide all of the model statements
(either typed or neatly handwritten).
Next few pages: Computer output for each model (and any additional output you may have
needed to create).
Following the output: Type your results from testing each theory making sure that you
clearly reference specific numbers from the output and clearly communicate your
conclusion in the context of the problem. If you need to do any calculations, you may
neatly handwrite the formulas you use to support your calculations—but these must
appear in the write up of the section rather than on a separate page.
Following the test of theories: Type your results from comparing the models. Be sure to
provide supporting calculations and clearly communicate your results at each step. You
may neatly handwrite the formulas you use to support your comparisons—but these must
appear in the write up of the section rather than on a separate page. Be sure to include the
regression equation suggested as a result of your analysis.
Next: Provide the table from Part 5 along with your supporting documentation for the
Finally: Include your typed and formatted Executive Summary.
Before you come to class on November 1, 2011, staple the entire assignment together.
Completed projects will be collected at the start of class.