Estimating time spent on work tasks

Research from arXiv highlights a new method for estimating the time spent on work tasks, a critical factor in understanding how technological advancements impact occupations. This development addresses previous limitations in task-weighting methodologies, which often relied on imprecise data or opaque models. The proposed approach offers a principled way to quantify time allocation across approximately 18,000 tasks, covering a significant portion of U.S. employment.

Why it matters

Accurate estimation of time spent on tasks is crucial for understanding the impact of technology on labor markets and occupational structures. This research provides a more robust foundation for strategic workforce planning, enabling better anticipation of skill shifts and resource allocation in response to technological change.

Key insights

• Prior methods for weighting tasks in economic models of occupations have been identified as idiosyncratic or poorly justified.
• Existing time share estimations for tasks often use coarse ONET data, which is not designed for this specific application, or rely on 'black-box' language models.
• A new principled method has been developed to estimate time shares for nearly 18,000 tasks that encompass a substantial portion of U.S. jobs.
• The task-based framework in economics is a standard model for analyzing technology's effects on work by observing changes in task-level requirements.
• The aggregation of task-level effects to occupation-level impacts requires accurate task weighting.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2608.05172