Standard Occupational Classification (SOC) System 2018
Source Overview
The SOC is a federal classification system that organizes all occupations in which work is performed for pay or profit. The 2018 revision contains 867 detailed occupations organized hierarchically into 23 major groups.
Core Framing
The SOC is designed as a task-based classification: it categorizes workers by work performed, not by job titles, credentials, or educational requirements. This design choice enables consistent statistical comparison across federal agencies while remaining agnostic to credential inflation and title variation.
The system balances competing demands: statistical comparability, adaptability to economic change, backward compatibility with historical data, and practical usability for coding millions of workers.
Key Design Decisions
Classification Principles (structural rules):
- Each occupation assigned to only one category at the most detailed level
- Classification based on work performed (and sometimes skills/training needed)
- Managers defined by planning and directing duties
- Supervisors classified by time spent supervising (80% threshold in certain major groups)
- Time series continuity maintained to extent possible
Coding Guidelines (operational rules):
- Code based on work performed, not job titles
- When work fits multiple occupations, code to highest skill requirement
- If skill requirements equal, code to where most time is spent
Extracted Content
Atoms and molecules extracted from this source:
- 02-atom—task-based-classification
- 02-atom—all-other-escape-valve
- 06-atom—classification-principles-vs-coding-guidelines
- 06-atom—direct-match-concept
- 02-molecule—disambiguation-by-skill-or-time
- 02-molecule—hierarchical-aggregation-levels
- 06-molecule—crosswalks-for-interoperability
Notes
The SOC revision process (2010 → 2018 → 2028) demonstrates how controlled vocabularies evolve through formal governance: multi-year cycles, public comment periods, interagency coordination, and explicit attention to backward compatibility.
The “Direct Match Title File” concept (where a job title maps to exactly one occupation) is particularly relevant to entity resolution and knowledge graph construction.