FINAL DATA ANALYSIS PROJECT THE PROBLEM : Universities and colleges do not like to admit students who
FINAL DATA ANALYSIS PROJECT
THE PROBLEM : Universities and colleges do not like to admit students who do not perform well. It is expensive and unpleasant for both the student and the school.
Since you work in the Department of Education and are a trained statistician, your supervisor has asked that you assist the Head of the Admissions Department. You have been given the MIDWEST SCHOLASTIC DATA file to develop a detailed statistical plan that the Admissions Officer can use to determine which students are most successful at her university.
THE DATA : Your data, the MIDWEST SCHOLASTIC DATA file, contains a number of variables. The database contains information on all undergraduate students at the University of West Erlham County - (UWEC). 13 variables are considered for each student: sex (SEX), high school percentile (HSP), cumulative GPA (GPA), age (AGE), total credits earned (CREDITS), classification (CLASS), school/college (COLLEGE), primary major (MAJOR), residency (RESIDENCY), admission type (TYPE), ACT English score (ENGLISH), ACT math score (MATH), and ACT composition score (COMP). YOUR INSTRUCTOR WILL SHOW YOU HOW TO GET THE DATA INTO DATA DESK.
HOW TO BEGIN: AS a WORKING HINT explore the data set before you try to formulate a plan.
- Review each of the variables individually, especially graphically. Look for anything unusual such as outliers.
- Review the variables in pairs – look for variables that are closely related.
-
Decide which variables will be useful predictor variables.
3. Fit a linear model (i.e. use regression) to all the variables that appear to be important. - When you are making your recommendations to the Admissions Officer, use your work to see if differently qualified students do better in certain majors and divisions
-
Prepare to document your work in two parts.
YOUR DOCUMENT: The first part will be a non-technical write-up of work in a letter to the Admissions Officer. Explain your findings, recommendations and reasons to her using words understood by someone with NO statistical training. Do not exceed three to four pages.
The second part will be an appendix describing your technical work. This will be the most substantial part of your document. Although this will be second in the presentation of your work, you will likely write up the appendix first; then, using the information in the appendix, you will write up the non-technical document that accompanies it.
In the documentation of your work, use the suggested outline below for guidance. Label the sections for clarity.- Introduction and description of the problem.
- Description of the Variables.
- Description of your Process and Preliminary findings.
- Interpretation of the Preliminary Work and Discussion of Your next Steps.
5. Summary and Conclusions. (In this part of the write up, include only relevant graphs.)
Deliverable: Word Document
