This site is undergoing review. Some annotations were human-generated, some AI-generated — all are being verified.
Back to papersAbstract Summary Main Finding Notes Links job postings to wages
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
[Claude classification]: This paper is NOT about AI. It studies general skill requirements in job postings from 2010-2015, before the LLM era. The paper uses text analysis to categorize job posting keywords into 10 skill categories (cognitive, social, character, writing, customer service, project management, people management, financial, computer general, software specific). While it discusses technological change and computerization in the literature review, the empirical analysis does not specifically study AI adoption or AI-related skills. The paper makes correlational rather than causal claims. Previous incomplete coding incorrectly listed 'Augmentation' as a topic, but this paper does not study augmentation vs substitution - it studies cross-sectional variation in skill requirements.
Skill Requirements across Firms and Labor Markets
Deming, Kahn
2017NBER Working Paper Series
Other
We study variation in skill demands for professionals across firms and labor markets. We categorize a wide range of keywords found in job ads into ten general skills. There is substantial variation in these skill requirements, even within narrowly defined occupations. Focusing particularly on cognitive and social skills, we find positive correlations between each skill and external measures of pay and firm performance. We also find evidence of a cognitive social-skill complementarity for both outcomes. As a whole, the job skills have explanatory power in pay and firm performance regressions, beyond what is available in widely-used labor market data
Deming and Kahn use text analysis of nearly 45 million job postings from Burning Glass (2010-2015) combined with OES wage data and Compustat firm performance data to study variation in skill requirements for professional occupations across firms and labor markets in the US.
Cognitive and social skill requirements in job postings positively predict wages across labor markets and firm performance, with evidence of complementarity between the two skills; these skill measures account for approximately 15% of variation in firm pay and 5-12% of variation in firm productivity.
- Key Methods
- Text analysis of job postings, regression analysis with occupation/MSA/firm fixed effects, variance decomposition
- Sample Period
- 2010-2015
- Geographic Coverage
- US
- Sample Size
- 44.9 million job postings for professional occupations; 56,611 MSA-occupation cells; 85,695 firms
- Level of Analysis
- Individual, Firm, Occupation, Region
- Occupation Classification
- SOC 2010
- Industry Classification
- NAICS 2012
- Replication Package
- Yes