Regression Analysis using Continuous, Binary, Multinomial + Count Outcomes

An in-person event at the University of Hull.

This course is designed for beginner and intermediate users of regression analysis who work with data in which outcome variables are measured on different scales. While much of the data used in economics and social science research is measured on ratio (continuous) and interval (time-series) scales, researchers frequently encounter outcome variables with more complex structures. These include binary, ordinal, nominal, and count variables, as well as censored or limited dependent variables. Such data structures require modelling approaches that extend beyond the standard linear regression framework.

The goal of this course is to move beyond ordinary least squares (OLS) and equip participants with the conceptual understanding and practical skills needed to analyse outcome variables with complex data structures. Participants will gain hands-on experience in estimating and interpreting nonlinear regression models commonly used in applied research.

The course begins with a brief review of linear regression and its assumptions before introducing generalised linear models and nonlinear estimation methods. Participants will then learn how to estimate and interpret models appropriate for different types of dependent variables, including binary choice models (logit and probit), multinomial and ordered choice models, and count data models such as Poisson and negative binomial regression. Attention will also be given to model diagnostics, interpretation of marginal effects, and practical issues in empirical modelling.

Throughout the course, participants will apply these methods using Stata, gaining practical experience in estimating models, interpreting results, and presenting findings in applied research contexts. By the end of the course, participants will have a stronger understanding of how to select and implement appropriate regression models for a wide range of empirical applications in social sciences.

Outcomes:

Participants will:

• Identify appropriate regression models for key research questions and data structures for a variety of outcome variables;
• Formulate and estimate regression models for continuous, binary, multinomial, count outcomes and censored data;
• Interpret and critically evaluate empirical findings across nonlinear models;
• Apply model diagnostics, goodness-of-fit, and marginal effects analysis.

Contributors:

Dr Raymond Swaray is a Senior Lecturer (Associate Professor) at the University of Hull Business School, and a Deputy Director of Advanced Data Analytics (ADA) for the WRDTP. His research focuses on empirical modelling using applied econometric and statistical methods to analyse economic, financial and social science data. He has extensive experience teaching quantitative methods and supervising PhD students. He is also a recipient of the prestigious Charles R. Goeldner Article of Excellence Award.

Important:

• This is an in-person event at the University of Hull (refreshments and lunch will be provided).
• This event is open to members of WRDTP partner institutions only.
• Bookings will close at 9.00am on Friday `1st May.
• When booking, we ask that you use your institutional (.ac.uk) email address and complete all fields of the booking form. Thank you for your understanding.

Please note: The WRDTP is committed to sustainability and to reducing the waste from excess catering at events. A key challenge here is non-attendance at events. From October 1st 2025, the WRDTP will be changing the way we manage the non-attendance of PGR students who have booked place/s at WRDTP Training events. Any PGR student who does not inform the WRDTP (via training@wrdtp.ac.uk) that they will not be able to attend a WRDTP event at least 3 working days before the event takes place will have the cost of their place deducted from their RTSG (if a WRDTP-funded student), or have this charged to their department (if not funded by the WRDTP). This will allow us to better plan for events and to avoid catering waste. Thank you in advance for your cooperation on this matter.