SAS代写 - PowCoder代写

Econ 453 Fall 2018 — Final Exam Course Project Requirements second release version

Note: This document may be revised to increase its clarity or to provide tips or hints.

The final exam for the course consists of students turning in a technical PowerPoint document that discusses the OECD PISA exam and shows the result of a logistic regression model that the student has built.
This will be individual work. A student may consult a course colleague for tips, but the work is to be the student’s.
This assignment will serve as the end-point for the mastery based learning objectives in the class.

Beginning of Submission Date: Due Date:
Submission Location:

Data Sets for Project (pick one)

Thursday, December 7, 2018 Wednesday, December 12, 2018 D2L

OECD 2015 PISA student data set.
OECD 2015 PISA Financial literacy student data set
OECD 2015 PISA Collaborative problem solving data set The data can be found at: http://www.oecd.org/pisa/data/2015database/ The goal is for students to demonstrate the SAS skills and understanding of logistic regression models from this semester. In addition, the project is to serve as a technical writing sample for a student’s work portfolio or as a technical writing sample. Lastly, students should make sure their work is unique by choosing interesting independent variables and an interesting dependent variable.

The Project will have the following sections:

Part 1. An introduction that discusses the OECD PISA exam and motivates the forthcoming analysis and statistical model that will be developed. The binary dependent variable is defined and described. The independent variables being used will be described. The data merging takes place at this step. (See data requirements later on.)

Part 2. Descriptive statistics of the dependent and independent variables are provided. Both the description of variables and an analysis of variables is provided. Use one page for each variable you discuss. Provide a tabular report, a graph, the variable’s definition and some insight about the distribution of the variable.

Part 3. Conditional analysis is provided. Use one page for each variable that you analyze. (Hint: See mini-project as an example for this part.) Provide a tabular report and a graph.

Part 4. The modeling variables are created—categorical variables must be turned into dummy variables for the model. Missing values and how they are dealt with are discussed here. Discuss the reasoning about the choices you make and use information from Part 3 to support your choice.

Part 5. The logistic regression model is estimated and results are shared. Discuss which variables are statistically significant and if the direction and magnitude are what you expect (based on Part 3).

Part 6. Create and use predicted probabilities from the model to provide intuition for the reader about the insights from the model in Part 5. This should be done in several ways. Provide a simple conclusion to your work.

Appendix. Code and technical information is provided here.

Other Requirements

Additionally, students will have some common elements to their projects and some unique elements.

Students need to create a smaller permanent data set to make their work easier to do. This could be done multiple times depending if variables change over time.
Your dependent variable must make sense and have little or no missing values.
Dependent variable—PISA data set o ItisrecommendedthatstudentsusePV1Math,PV1Read,orPV1Sciasthefoundational variable for the binary dependent variable. o Therearealsosubjectsub-scoresthatmaybeinteresting. o Itwillbesimilarforthoseworkingwithfinancialliteracyorcollaborationdatasets
Independent variables o Thesamplesizetobeanalyzedwillbeatleast50,000.

o Countries—studentsneedtohaveatenoughcountries(n)intheiranalysistomeet

sample size requirement
 Pick countries that are similar based on geography  Pick countries that are similar based on culture
 Pick countries based on GDP per capita
 Pick countries based on similar population size

o Allstudentswillusegenderasavariableandtwootherdemographicvariables,aswell as CNT.

 Dummy variables to identify n-1 countries will be created. (In Part 4).
o Studentsshouldhaveuniqueprojects.Thatmeansthatthedependentvariables,

countries, independent variables can’t be the same.
 Don’t do the same work as a classmate.
 It could be considered an academic integrity violation.

o Studentsmustselectatleastthreeinterestingvariablestolookatasindependent variables.

o Anyindependentvariableswithmissingvalueswillneedtobekept.  The missing values need to be imputed.

 Expect to have missing values in your interesting independent variables.
 Choose independent variables so that you don’t have multicollinearity issues as a result of missing value imputation.

Writing needs to be professional and combine both technical and non-technical language. Where possible use non-technical language.
Graphics should be produced with Proc SGPLOT both for descriptive statistics and conditional statistics.
Students will not make in-class presentations of their work.
More variables than the minimum number can be used.
All variables identified must be present in Part 2, Part 3, Part 4, and Part 5. ??? = To be determined and announced to class

Page 3 of 6

Learning Objectives

The following learning objectives will apply to each section of the project. Separate grades will be assigned when a section is associated with two or more learning objectives.

Type	Learning Objectives	Part 1	Part 2	Part 3	Part 4	Part 5	Part 6	Appendix
Mastery Based Topics	Creating Tabular Reports for Descriptive and Conditional Statistics	U	U	U
Data Creation, Manipulation, Merging, and Cleaning	W	W	W
Data Visualization	S	S	S
Logistic Regression	Z	Z	Z
Working with Large Data	Y	Y	Y	Y	Y
Code Reusability and Organization	T
Non- Mastery Based Topics	Creation of Documentation	T
Learning Technology
Ordinary Least Squares
Professional Development
Professional Writing	X	X	X	X	X	X
Problem Solving with Technology
Time Management	V	V	V	V	V	V	V
Grades per Part	2	7	5	4	4	7	2

Illustrative Example

A mini-project example provided by Prof. Sugiyama can serve as a partial model for students to understand some of the formatting guidelines. (It doesn’t contain an example of missing value imputation.) It is helpful to use it as an example for what the different parts of the project should “look like.”

Formatting

Students are to create a technical PowerPoint.

Have a title page
You are writing a research paper, but it a PowerPoint form. Use one idea per page in order to think about how to think about writing your document.
Please make sure you PowerPoint is formatted to a landscape 8.5 x 11 page. This is not always the default.
Please make sure you use a small font of 12 or 14 for all of your text.
Please use page numbers.
Do not use design elements or backgrounds. You want to have a blank page.
Use about 1 inch margins on your pages. You have to manually set you margins by moving your text boxes and charts.

Submissions:

Students will be required to turn in their work by uploading it to D2L. D2L: Please upload three items

PowerPoint PPT file
SAS code used for your work
SAS output in PDF Grading Expectations Students are to do their best at this assignment given the time that they have. Unlike many assignments you may have experienced, this is a “stretch” assignment to see what students in the class can do. It is not expected that students will do this assignment perfectly. Please try your best and if you have difficulties, denote what they are in your PowerPoint presentation. Additional Resources Additional SAS or Methodological handouts (to the ones already given in class) may be provided to students depending on student questions. Additionally, handouts that provided more detailed steps for students may be provided or may be discussed in class.

Tips and Points for Students

Below you will find some quick pointers to topics and documents.

Issue	Tip
How to create dummy variables for logistic regression	Handout “Data Step (Creating Variables)
Reducing the size of your data set	Handouts “Data Step (Reducing SAS Datasets)”
Changing variable names	Handouts “Proc Datasets” “Data Step (How to do things similar to Proc Datasets)”
Understanding or changing Formats	Handouts “Proc Format”
How to read in a SAS data set	Handouts “How to Read in a SAS Dataset”
How to create conditional tables or describe statistic tables	Handouts “Proc Tabulate”, “What Do You See”
How to estimate a logistic regression	Handouts “Proc Logistic”
How to visual or intuitively show predicted probabilities from a Logistic Regression Model	Handout “Kernel Density Estimation”
How to Investigate Variables for Usage in Predictive Model	TBA or class discussion

Version Information

Version 1: Draft version released on Nov 15, 2018 Version 2: Dec 3, 2018

Page 6 of 6

Related Posts