Occupation and industry are essential measures in many social surveys. Traditionally, occupation data are collected through open‑ended questions on job title and duties, which are later coded to standard codeframes by expert coders. Measuring occupation is inherently challenging, as respondents may describe the same job in different ways or provide insufficient detail for accurate coding. These difficulties are often magnified in online self‑completion surveys, where participants complete questions without interviewer support, increasing the likelihood of vague or incomplete responses.
As survey research increasingly moves towards online and mixed‑mode designs, identifying approaches that enable accurate, consistent and efficient occupation and industry coding has become essential. This workshop, part of Survey Futures’ Research Strand 5 (‘Complex Measurements’), will explore a wide range of current and emerging methods designed to address the challenge of collecting and coding occupation data in web surveys. It will cover in‑survey coding techniques such as look‑up tools and closed‑list questions, the growing use of large language models both during and after data collection and qualitative research focused on respondent understanding of occupation and industry questions.
The workshop is aimed at survey methodologists, researchers, practitioners, and data users with an interest in improving the collection and classification of occupation and industry information.
This is an in-person event and lunch will be provided.
Please find the agenda below:
Occupation and Industry Coding in Online Surveys: Evidence, Practice, and Challenges
4 June 2026 | London | 10:30 – 16:00
Presentation format: 15-minute presentation + 10-minute discussion
Chair: Lisa Calderwood (Centre for Longitudinal Studies, UCL)
Welcome (10:30 – 11:10)
10:30 – 11:00 — Welcome coffee
11:00 – 11:10 — Introduction
Session 1: Mode Comparisons and Respondent Self-Coding (11:10 – 13:00)
11:10 – 11:35 — Helena Koerber (Centre for Longitudinal Studies, UCL) – An Evaluation of the Look-up Tool for Occupation Coding
11:35 – 12:00 — Cristian Domarchi (University of Southampton) – Industry and Occupation Coding: A Comparison of Office-based Coding and a Closed-list Approach
12:00 – 12:10 — Break
12:10 – 12:35 — Danielle Watson and Beth Jones (Office for National Statistics) – Using Mental Models Research to Review How We Collect SIC and SOC Information
12:35 – 13:00 — Cristian Domarchi (University of Southampton) – How Consistent is Occupational Coding Across Data Collection Modes? Findings from the Census Non-Response Link Study
Lunch (13:00 – 13:45)
Session 2: AI Approaches to Occupation Coding (13:45 – 15:00)
13:45 – 14:10 — Olga Kononyhina (LMU Munich) – Can Large Language Models Advance Occupational Coding? Evidence and Methodological Insights
14:10 – 14:35 — Patrick Sturgis (London School of Economics) – SOCbot: Using Large Language Models to Measure and Classify Occupations in Surveys
14:35 – 15:00 — Katrina Tyrell (Office for National Statistics) and Jennifer Arkell (Office for National Statistics) – Using Generative AI to Help Respondents Complete Industry and Occupation Questions
15:00 – 15:15 — Break
Session 3: Discussion (15:15 – 16:00)
15:15 – 16:00 — Discussion with introduction from Matt Brown (Centre for Longitudinal Studies, UCL)
When registering please let us know if you have any dietary requirements.