Census Linking


Census Linking with UK Data (23 August 2023)

Hillary Vipond, Organiser
Sponsored by the Economic History Society

Session 1 - Ryah Thomas
Missing mobility: Selection bias in nominal linking techniques'. It compares the selection bias and error rate of standard techniques to my preferred method (network persistence linkage), analyses the impact of different selection effects on estimates of population social mobility, and offers some historical explanations for the bias introduced by standard techniques. 

Session 2 - Gill Newton
Different methods for record linkage of a select group (employers) and a whole population sample across time through historical censuses, compared to the select group linked from trade directories to censuses.

Session 3 - Eilidh Garrett and Alice Reid
From parishes to towns to the Nation; some thoughts on census linkage.

'Best Practices' Discussion

Session 4 - Hillary Vipond
A multi-model specification assessment of error rates in census linking, comparison between UK and American level data, exploring impact of model choice on results.

Session 5 - Guy Solomon/Turing Group
How/why we are hoping to build a user community around our implementation, and how we have applied it to a historical case study (bicycle manufacturing in Coventry, 1881-1911). 

'Best Practices' Discussion

Concluding Thoughts