Question: Scenario (This scenario might be familiar, based on work from previous semesters. The dataset is entirely new, however.) You work for a small private business
Scenario
(This scenario might be familiar, based on work from previous semesters. The dataset is entirely new,
however.)
You work for a small private business college in California, and are trying to understand the factors
behind student success. Colleges (especially private ones) are businesses, and student success is
important to the success of the institution. Successful students tend to share positive views of the
college with other prospective students, tend to return for further training, etc. Moreover, successful
students lead to happy employers, which increases the value of a college's credentials and further
attracts students.
Thus, student success is an important marketing issue, and marketing tools can be used to foster
student success. If we can understand which factors predict student success, we can do things like
better segment the market, and recruit and admit students who are likely to be successful. If we can
identify issues that negatively impact students' success or times when students are at greater risk of
failure, we can take active steps to help them.
You have collected data on past student performance, in the provided file. Each observation represents
one student. The variables in the dataset are:
PROGRAM
DURATION - length of the program, in years
STATUS - Local ('L') or International ('P') (with a small number of other codes, for rare statuses)
GENDER
ESSAYSCORE - score of an essay required as part of the application process
MATHTEST - score on a math test administered as part of the application process
LOGICTEST - score on a logic test administered as part of the application process
TRANSFER_IN - 1 if the student transferred in from another program, 0 if they started directly in
the program
TOTALSEMESTERS - The total number of semesters enrolled to date
PROBATIONTERMS - The number of terms spent on academic probation
MAXGAP - The longest number of terms taken off by the student, when they would normally
have been enrolled
FINISHED - 1 if they have finished their studies, 0 if not (whether or not they graduated
successfully)
AGE - Approximate age of the student, in years, when they started the program
hsMATH1, hsMATH2, hsLANGUAGE - The student's high school grades in two math courses and
the language course required for admission
MATHCERT - 1 if the student completed a math certificate before applying to the program, 0
otherwise
GPA - The student's accumulated GPA, on a 12 point scale
HASGRADUATED - 1 if the student has successfully graduated, 0 otherwise (whether they are
still taking courses, or have terminated their studies)
Your report should include more than the answers to the specific questions; it should also include both
details of your models and how you built them (further outlined below), and your key findings from the
models. Claims you make should be backed up with appropriate justification from your analysis. You
should include only the relevant numbers in the body of the report itself; attach more complete output
from JMP as appendices, as needed.
You should explore each of the key issues and techniques used this term. Your report should detail your
investigation. Your work should show more than simply building a straightforward model or two, like
you would do for an assignment or test question. It is expected that your work will provide evidence
that you thought carefully about the problems and the dataset, and about how the tools examined in
this course can be applied to address them.
Question 4
Use clustering, considering some or all of the following variables: GENDER, ESSAYSCORE, MATHTEST,
LOGICTEST, AGE, hsMATH1, hsMATH2, and hsLANGUAGE, to try to identify segments in the data.
Characterize the segments found.
Step by Step Solution
There are 3 Steps involved in it
Get step-by-step solutions from verified subject matter experts
