Easy Statistics: Linear and Non-Linear Regression
What you'll learn
- The theory behind linear and non-linear regression analysis.
- To be at ease with regression terminology.
- The assumptions and requirements of Ordinary Least Squares (OLS) regression.
- To comfortably interpret and analyse regression output from Ordinary Least Squares.
- To learn and understand how Logit and Probit models work.
- To learn tips and tricks around Non-Linear Regression analysis.
- Practical examples in Stata
Make sure to check out my twitter feed for monthly promo codes and other updates (@easystats3)
Three courses combined. Linear and Non-Linear Regression and Regression Modelling.
Learning and applying new statistical techniques can often be a daunting experience.
"Easy Statistics" is designed to provide you with a compact, and easy to understand, course that focuses on the basic principles of statistical methodology.
This course will focus on the concept of linear regression, non-linear regression and regression modelling. Specifically Ordinary Least Squares, Logit and Probit Regression.
The first two parts will explain what regression is and how linear and non-liner regression works. It will examine how Ordinary Least Squares (OLS) works and how Logit and Probit models work. It will do this without any complicated equations or mathematics. The focus of this course is on application and interpretation of regression. The learning on this course is underpinned by animated graphics that demonstrate particular statistical concepts.
No prior knowledge is necessary and this course is for anyone who needs to engage with quantitative analysis.
The main learning outcomes are:
To learn and understand the basic statistical intuition behind Ordinary Least Squares
To be at ease with general regression terminology and the assumptions behind Ordinary Least Squares
To be able to comfortably interpret and analyze complicated linear regression output from Ordinary Least Squares
To learn tips and tricks around linear regression analysis
To learn and understand the basic statistical intuition behind non-linear regression
To learn and understand how Logit and Probit models work
To be able to comfortably interpret and analyze complicated regression output from Logit and Probit regression
To learn tips and tricks around non-linear Regression analysis
Specific topics that will be covered are:
What kinds of regression analysis exist
Correlation versus causation
Parametric and non-parametric lines of best fit
The least squares method
Beta's, standard errors
T-statistics, p-values and confidence intervals
Best Linear Unbiased Estimator
The Gauss-Markov assumptions
Bias versus efficiency
Zero conditional mean
Regression in logs
Practical model building
Understanding regression output
Presenting regression output
What kinds of non-linear regression analysis exist
How does non-linear regression work?
Why is non-linear regression useful?
What is Maximum Likelihood?
The Linear Probability Model
Logit and Probit regression
Dummy variables in Logit and Probit regression
Odd-ratios for Logit models
Practical Logit and Probit model building in Stata
The computer software Stata will be used to demonstrate practical examples.
The third part provides useful practical tips for regression modelling.
Understanding how regression analysis works is only half the battle. There are many pitfalls to avoid and tricks to learn when modelling data in a regression setting. Often, it takes years of experience to accumulate these. In these sessions, we will examine some of the most common modelling issues. What is the theory behind them, what do they do and how can we deal with them? Each topic has a practical demonstration in Stata. Themes include:
Fundamental of Regression Modelling - What is the Philosophy?
Functional Form - How to Model Non-Linear Relationships in a Linear Regression
Interaction Effects - How to Use and Interpret Interaction Effects
Using Time - Exploring Dynamics Relationships with Time Information
Categorical Explanatory Variables - How to Code, Use and Interpret them
Dealing with Multicollinearity - Excluding and Transforming Collinear Variables
Dealing with Missing Data - How to See the Unseen
Who this course is for:
- Academic students of any level.
- Practitioners who require quantitative knowledge.
- Business users and managers who engage with quantitative reports.
- Government workers who are involved in policy analysis.
- Anyone who has an interest in, or needs to engage, with statistical regression.
Check out my twitter feed for regular promo codes.
Franz is a Professor of Economics at the University of Westminster. Franz joined the University of Westminster in 2006 after completing his PhD in Economics at Lancaster University.
Franz's personal research interests are in education economics, labor economics, and applied econometrics. Franz has made scientific contributions to issues such as social mobility, measuring the returns to education, the effect of weather of happiness and identity formation. He has been involved in numerous funded research projects from research councils and government departments.
Franz has contributed to wide range of projects including policy evaluation and bespoke econometric advice to UK government departments. These include the Ministry of Defence, HM Revenue and Customs, the Department for Education and the Department for Business, Innovation and Skills.
He has published in leading journals such as Economics of Education Review, the Oxford Bulletin of Economics and Statistics, the British Journal of Political Science and the British Journal of Sociology. Franz has also contributed to numerous policy reports and his research has been covered by media outlets such as BBC news, BBC Radio 4, The Economist, The Guardian, The Times, and Huffington Post. Franz also has a monthly radio program called Policy Matters on Share Radio.
Franz is an experienced online educator and has published several online courses including LinkedIn Learning.