Online lecture 8 – Session 2: Multinomial regression models by and for historians/historical demographers

Lecturers: Dr. Mayra Murkens, Faculty of Arts, University of Groningen

Content: One of the main issues historical demographers face, is the lack of a proper estimation of the population at risk. Multinomial logistic regression models can approximate differences between groups in past societies without the information on the population at risk. In this lecture, the idea behind multinomial logistic regressions and how to structure your data to perform these models will be discussed.

Objectives:

• Show how multinomial logistic regression analyses can offer a solution when a population at risk is missing. 
• Show what is needed to perform a multinomial logistic regression.

Requirements: Active participation.

Recommended reading:

• Renzo Derosas, Cristina Munno, The Place to Heal and the Place to Die. Patients and Causes of Death in Nineteenth-Century Venice, Social History of Medicine, Volume 35, Issue 4, November 2022, Pages 1140–1161, https://doi.org/10.1093/shm/hkaa050
• Murkens, M., Pelzer, B., & Janssens, A. (2022). Transitory inequalities: how individual-level cause-specific death data can unravel socioeconomic inequalities in infant mortality in Maastricht, the Netherlands, 1864–1955. The History of the Family, 28(1), 95–131. https://doi.org/10.1080/1081602X.2022.2084442