Height and marital outcomes in the Netherlands, birth years 1841-1900

Kristina Thompson*, Xander Koolman, France Portrait

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

6 Citations (Scopus)


Adult body height appears to be significantly associated with marital outcomes: taller men across contexts have been found to be more likely to be married, and more likely to be married at younger ages. We are interested in exploring both outcomes individually and simultaneously, while using an unique, individual-level dataset of Dutch men and their brothers born between 1841 and 1900. To do so, we exploit survival models and cure models. While survival models yield a single estimate for the hazard (or age at) marriage, cure models yield two: one for the likelihood of marriage, and one for the hazard of first marriage. Cure models thus account for selection into marriage, while survival models do not. We find that, in the survival analyses, being in the shortest 20 % of heights is associated with later ages of marriage, relative to being average height. However, when we account for selection into marriage with cure models, we find that height is no longer associated with age at marriage. Instead, we see that height is associated with the likelihood of being married, with being in the bottom 20 % of heights associated with a 56.1 % decreased likelihood of being married, relative to being average height. We therefore conclude that height may be a gatekeeper for access to marriage, but it appears that other factors – likely related to the ability to set up an independent household – are more important in determining the timing of marriage for our research population.
Original languageEnglish
Article number100970
JournalEconomics and Human Biology
Publication statusPublished - May 2021
Externally publishedYes


  • Cure models
  • Height
  • Marriage
  • Netherlands
  • Sibling design


Dive into the research topics of 'Height and marital outcomes in the Netherlands, birth years 1841-1900'. Together they form a unique fingerprint.

Cite this