Enterprise AI Analysis
Leveraging large language models for career mobility analysis: a study of gender, race, and job change using U.S. online resume profiles
This study introduces an innovative AI-driven framework, FewSOC, to enhance career mobility analysis by improving occupational classification accuracy from resume data. We analyzed 228,710 U.S. career trajectories, revealing that intra-firm occupational changes are the strongest drivers of upward mobility. Crucially, we found significant gender and racial disparities: women and Black college graduates experience lower returns from job changes than their male and White peers, while Asian college graduates show a higher likelihood of upward mobility. Our multilevel sensitivity analyses confirm these disparities are robust, highlighting the need for targeted interventions to address structural barriers in career progression.
Key Executive Impact Metrics
0.72
FewSOC Precision Score for SOC Classification
228,710
Career Trajectories Analyzed
53.71%
Individuals Achieving Upward Mobility
Deep Analysis & Enterprise Applications
Select a topic to dive deeper, then explore the specific findings from the research, rebuilt as interactive, enterprise-focused modules.
Methodology
Delve into the AI-powered techniques that underpinned our groundbreaking analysis.
SOC Classification Accuracy
The FewSOC framework achieved a significantly higher precision score of 0.72, representing a 10.42% improvement over Lightcast's default labels (0.65). This enhanced accuracy ensures more reliable and valid findings in our career mobility analysis.
0.72 FewSOC Precision ScoreFewSOC Framework Overview
FewSOC, our LLM-based occupation classification method, processes job titles and company names through a two-step pipeline: initial SOC generation via few-shot prompting, followed by SOC mapping and matching to ensure accuracy and consistency with the O*NET-SOC 2019 taxonomy.
Initial SOC Generation (LLM few-shot prompting)
→
Auxiliary Binary Predictions (Non-occupational roles, multiple roles)
→
SOC Mapping & Matching (Crosswalk tables, word overlap similarity)
| Model | Accuracy |
|---|---|
| Gemini 2.5 Flash | 0.7425 |
| GPT-3.5 Turbo | 0.7195 |
| GPT-4.1 Mini | 0.6073 |
| Llama 3.1 8B Instruct | 0.4722 |
Career Trajectory Construction Steps
This outlines the systematic process for constructing linear career trajectories from raw resume data, ensuring consistency and relevance for mobility analysis.
Job Records Filtering (valid dates, non-occupational titles excluded)
→
Education Records Filtering (Bachelor's degree or higher, valid dates)
→
Post-graduation Gap Threshold (to reflect actual first position)
→
Timeframe Filtering (1999-2022 job records)
→
Linear Trajectory Construction (Remove overlaps, sort by start/end dates)
→
Career Length Standardization (5-year window)
Findings
Uncover the core insights and disparities shaping career trajectories in the U.S. labor market.
Gender and Racial Disparities in Upward Mobility
The study revealed significant disparities in upward career mobility outcomes by gender and race. Women consistently experienced lower returns from job changes compared to men, regardless of the change type. Black college graduates also faced greater disadvantages than White peers. In contrast, Asian college graduates were more likely to achieve upward mobility.
Women and Mobility
Women consistently face greater challenges in attaining upward mobility than men (p < 0.001), experiencing lower returns from job changes, particularly intra-firm occupational changes (Type-2) and intra-firm lateral moves (Type-3).
Black College Graduates
Black college graduates are at a greater disadvantage in achieving upward mobility compared to their White peers (p < 0.001).
Asian College Graduates
Asian college graduates are more likely to achieve upward mobility than White college graduates (p < 0.001). However, specific negative interaction effects were observed for Asian women in Type-1 job changes and Asian men in Type-2 job changes in multilevel analyses.
Implications
Explore the broader significance of our research for policy, diversity, and future work.
This section will contain further analysis of the policy recommendations, diversity initiatives, and future research directions stemming from the study's findings on career mobility and disparities. Our work underscores the critical need for organizations and policymakers to address systemic barriers affecting career progression for women and minority groups, leveraging precise AI tools like FewSOC for data-driven interventions.
Predict Your AI's Impact
Estimate the potential savings and reclaimed hours by integrating AI into your enterprise operations, tailored to your specific context.
Estimated Annual Savings
$0
Hours Reclaimed Annually
0
Implementation Roadmap
A phased approach to integrate AI solutions seamlessly into your operations, ensuring maximum value and minimal disruption.
Phase 1: Discovery & Strategy
Comprehensive analysis of existing workflows, identification of AI opportunities, and development of a tailored implementation strategy.
Phase 2: Pilot & Proof-of-Concept
Deployment of AI solutions in a controlled environment to validate effectiveness, gather feedback, and demonstrate tangible ROI.
Phase 3: Full-Scale Deployment
Rollout of AI solutions across relevant departments and processes, with continuous monitoring and support to ensure smooth integration.
Phase 4: Optimization & Scaling
Ongoing performance optimization, identification of new AI applications, and expansion of solutions to further enhance operational efficiency.
Ready to Transform Your Enterprise with AI?
Book a personalized consultation with our experts to discuss how our AI-driven solutions can address your specific challenges and drive tangible results.
Ready to Get Started?
Book Your Free Consultation.
Let's Discuss Your AI Strategy!
Lets Discuss Your Needs
Select a Date
Sun
Mon
Tue
Wed
Thu
Fri
Sat