Job quality

Understanding job quality – the aspects of work that affect employee wellbeing – is crucial for improving workforce conditions.

While research typically gathers insights through employee surveys and reviews, this ESCoE project explores whether indicators of job quality can be extracted automatically from online job adverts. Building on previous work, we developed an open-source Python package to identify and categorise job quality indicators in job postings.

Using a taxonomy based on the Measuring Job Quality Working Group’s seven dimensions of job quality, we designed a pipeline that classifies relevant sentences and maps them to categories such as employment terms (e.g. offers to work flexible hours), benefits (e.g. life assurance) and job design (e.g. mentions of career progression).

As part of this project, we also undertook a programme of modernisation of related Nesta-ESCoE projects in order to make the code readily available for others to use. This means we now have open source code libraries that can extract information such as skills and company descriptions from job adverts, and map these to existing taxonomies.

The project consists of two workstreams (1) extracting selected drivers of job quality and (2) Machine learning (ML) infrastructure & maintenance:

• Job quality: We have extracted benefits and other aspects mentioned in job adverts, evaluated the resulting data and compared a selection of job quality dimensions across occupations, industries and regions. We have also tested our findings and algorithms to a specific sector; the early years careers.
• Machine learning infrastructure: We have developed an ML Infrastructure layer for all our data science and labour market projects. This included building maintainable approaches using python libraries and project refactoring to modify codebases of any existing projects. Our ultimate aim is to create models that are easy to use for our stakeholders. For more information about the ML Infrastructure, read the technical blog.

We applied this approach to a random sample of job adverts from Nesta’s OJO database in order to investigate offered job quality across sectors in the UK. Some example findings include that mentions of Learning and Development are negatively correlated with salary across occupations, that flexible location is mentioned most commonly in the IT, Human Resources, and Marketing sectors, and that the sectors most likely to offer flexible hours are Logistics & Transport, Health & Social Care and Hospitality & Catering.

We have also used this methodology to carry out a detailed analysis of the advertised working conditions in the early years sector and comparable industries. Nesta’s Fairer Start Mission seeks to close the gap in early outcomes between young children growing up in disadvantage and the national average. A key challenge faced by the mission is a shortage of early years practitioners, and it is thought that early years professionals are leaving the sector for better pay and working conditions. Our analysis suggests that the pay for waiting/retail staff is similar to that for early years practitioners, but that those roles have less secure contracts and are less likely to offer Continued Professional Development (CPD). Meanwhile, vacancies for school teachers are advertised with better pay compared to early years practitioners, and similar other benefits (career progression and Learning and Development).

These indicators provide the first-ever timely and low-cost signal of job quality dimensions. The UK currently has no official measuring system for job quality, and the main indicator is an annual survey.

While there have been several efforts to define the dimensions of job quality, there have been far fewer efforts to measure these dimensions. Our innovative methods offer an approach to quickly detect shifts in the prevalence of different benefits within jobs. They also provide a complementary source of data compared to other research that focuses on measurable outcomes of job quality in terms of worker well-being.

Applied to a wider sample of job adverts, it gives researchers and policymakers a way to monitor aspects of the quality of jobs offered and could be used to track offered job quality in target sectors, on a specific dimension, or across the market as a whole.

By releasing the code, both the Office for National Statistics and government departments (who hold their own datasets of job adverts) will be able to monitor aspects of job quality in any industry or occupation, and these novel measures will provide a useful complement to more traditional approaches.

Read more about the impact of this project in this case study.

Oxbury, R. and Gallagher, E. ‘Automatically identifying drivers of job quality in online job adverts’ Poster session, ESCoE Conference on Economic Measurement, King’s College London, 21-23 May 2025

Oxbury, R. ‘ Using Python to extract job quality from online job adverts’ Data science at Nesta, Medium, 31 March 2025

Oxbury, R., ‘Identifying drivers of job quality in online job adverts‘, ESCoE blog, 30 October 2024

Vines, J., ‘Measuring job quality using ojd_daps models: Technical blog‘, ESCoE blog, 30 October 2024

• See the open codebase
• See full analysis
• Read the case study on early years practitioners