MET4 Empirical Methods
Autumn 2022Spring 2023
This course builds on and extends the methodology and key subjects from the first year. Students will be trained in the use of empirical analysis for decision making. Special emphasis is given to interpretation of economic and behavioral data. Students will learn to distinguish random variation from systematic variation and causality from correlation. Methodological issues are integrated with other economic subjects through examples and specific applications. While the main focus in the first year course in statistics is univariate analyses, this course also covers multivariate methods.
The following topics are covered:
1. Introduction to scientific methods in social sciences
- Qualitative vs quantitative analysis
- Research ethics
2. Descriptive statistics
- Population and sample
- Types of data and information
- Central location, variance and co-variance
3. Comparing two populations
- Sampling distributions
- Comparing two means
- Comparing two variances
- Comparing two proportions
4. Chi-squared tests
- Goodness-of-fit (more than two proportions)
- Test for independence in a contingency table
5. Regression analysis
- Simple regression
- Multiple regression: Modelling and residual analysis
- Panel data
- Categorical regression (logit and probit)
6. Introduction to machine learning
7. Time series
- Explain the theoretical foundation of the statistical methods that are covered in the course.
- Substantiate the choice of a statistical method in a given realistic problem, and then apply the method.
- Use software to make descriptive statistics using numerical and graphical methods.
- Perform inference about one and two populations based on one or two samples.
- Use linear regression to identify the linear dependence structure based on a set of explanatory variables and a response variable, for cross-sectional and panel data.
- Distinguish between correlation and causality in empirical problems.
- Interpret and precisely describe the result of a statistical analysis.
- Know recent methods in developments in machine learning.
- Trade bias against variance in order to optimize the predictive power of machine learning techniques.
- Build an empirical model using variables and functional form in order to solve specific problems.
- Identify basic time series models and make predictions
- General competence:
- Recognize empirical arguments in public discourse and in the news, and criticize the choice of methods, execution and interpretation.
- Perform basic data analysis using modern computer tools.
Teaching consists of interactive sessions and lectures given at campus. Most of the curriculum will be supported by online based modules containing short videos, exercises and notes. The students will work with data labs containing exercises and cases from the textbook. Student will have to hand in an assignment to document competence in the use of statistical software and reporting of results.
It is recommended that the students are in command of the contents of MET3.
It is assumed that the students are in command of the contents of MET2.
Credit reduction due to overlap
Approved hand-in assignment.
The final grade is based on a written individual home exam (4 hours), which counts 70 %, and a take home exam (3 days, from 09:00 on day 1 until 14:00 on day 3), which counts 30 %. The take home exam is done in groups of 2-4 students.
The two exam elements can be taken independently, but it is recommended to take them within one term.
Students can retake each exam element independently or together.
A-F for both assessment elements and the overall course grade.
Keller: Statistics for Management and Economics, 2th Edition, Cengage
Parts of the curriculum are also covered in Jan Ubøe, Statistikk for økonomifag, which is used in MET2.
- ECTS Credits
- Teaching language
Autumn and Spring. Offered Autumn 2022.
Associate Professor Håkon Otneim, Institutt for foretaksøkonomi (Main course responsible)
Associate Professor Geir Drage Berentsen, Institutt for foretaksøkonomi