ME-637 Panel data analysis using Stata

The course is connected to the following study programs

• School of Business and Law, PhD programme

Recommended prerequisites

It is assumed that the students have basic knowledge in statistics and quantitative researchmethods. For PhD candidates at the University of Agder, it is assumed that they have taken thecourse ME-612-1 Survey of Statistical Methods or a course equivalent to this. Since this is anapplied statistics course, students must also have knowledge in Stata which is the software forthis course. Students without previous knowledge in quantitative research methods and Stataare expected to utilize more hours than the required workload

Course contents

This course introduces linear panel data analysis for research within business, management, andadministration. The primary objectives for this course are

1) to provide the student withrelevant knowledge in panel data techniques and

(2) how to apply those techniques in Stata so that the students can practically apply the knowledge in their own research.

The course is organized into five parts activities in Stata.

Part l introduces panel data, particularly how panel data structurelooks like, benefits of panel data over cross-section or time series, and the basic panel datamodel.

Part 2 demonstrates how to import excel data in Stata, declare the data as panel, testwhether panel data modelling is really needed for the data sample, and how to keep track of activities in Stata.

Part 3 presents basic (static) panel data models including the famous randomand fixed effects models and how to determine which of the two models is appropriate for anypanel data sample.

Part 4 introduces advanced (dynamic) panel data models based ongeneralized method of moments (GMM) - which is often used to mitigate endogeneityconcerns.

Part 5 presents two-way error model, introduces time-series and multilevel analyses,of learning

Below are the topics covered in each part:

Part 1 Introduction to Panel Data: 

• Data types and benefits of panel data
• Basic linear panel data model

Part 2 Panel Data Set-up in Stata:

• How to create log and do files in Stata
• How to import data into Stata
• Panel data set up & LM test in Stata

Part 3 Static panel data models:

• Random effects model
• Fixed effects model
• Random versus fixed effects models
• Between effects model
• Hausman and Taylor model

Part 4 Dynamic panel data models:

• Introduction to GMM
• Difference GMM model
• System GMM model
• Two-step GMM
• Specification options

Part 5 Other issues:

• Two-way error component
• Introduction to time series analysis
• Introduction to multilevel analysis
• Stata output export to Microsoft Word

Learning outcomes

Upon successful completion of the course, the students should be able to:

• differentiate data types and, in particular be able to organize data in panel data andnested data structures.
• explain assumptions underlying each panel data model.
• choose appropriate panel data model for different data samples.
• utilize panel data techniques to address endogeneity problems.
• apply panel data techniques to different research problems including theirs
• apply Stata to execute different panel data models
• interpret the output of each panel data model
• export Stata output directly to Microsoft Word in different ways

Examination requirements

• Approved individual assignment which involves data analysis using Stata. The data will be provided to students.
• Mandatory 80% class attendance

Teaching methods

• Lectures
• Stata lab session following each topic introduced
• Pre-class assignments.

Offered as Single Standing Module

Yes

Assessment methods and criteria

The final examination is a 100% written research paper where the students apply the paneldata techniques to answer their own research problems using Stata. Thus, students are encouraged to bring their own data during the course (if available) to get help regarding correct data structure and model.

The grading is Pass/Fail, where pass equals the letter grad B or better.