Cognitive Demand Index: 100 US Jobs Ranked by Brain Power (and What They Pay)
The Cognitive Demand Index v1.0 ranks 100 US occupations by cognitive demand using four O*NET ability scores. Judges score highest (CDI 84.5); pay tracks poorly with cognitive demand throughout the ranking.
"I view this as the most important paper of my career." That is Paul R. Sackett of the University of Minnesota, describing his 2022 Journal of Applied Psychology meta-analysis that revised the field's headline statistic on cognitive ability and job performance. The widely cited Schmidt-Hunter coefficient of r = 0.51 was, Sackett and his co-authors showed, inflated by an over-aggressive correction for restriction of range. Corrected, the operational validity is closer to r = 0.31, still the strongest single non-test predictor of job performance, but a more honest number.
That correction did not disprove cognitive ability's predictive power. It clarified its scale. And it sharpened a question we set out to answer with data: across the US labor market, where is cognitive horsepower actually demanded? Which jobs require the most reasoning, problem-solving, and inductive pattern recognition? And what do those jobs pay?
We built the Cognitive Demand Index (CDI v1.0) to answer that question with a transparent, reproducible artifact. The CDI ranks 100 US occupations on a 0–100 composite drawn from four O*NET cognitive descriptors, then pairs each rank with BLS May 2024 median wages. The full dataset is available as an open CSV for journalists, researchers, and anyone curious enough to sort it themselves.
Key Takeaways
- Judges score Critical Thinking importance of 97/100: the highest in the 100-occupation dataset, outranking every physician specialty, scientist, and engineer in the index. Their CDI composite of 84.5 sits within four points of anesthesiologists' 80.8.
- Cognitive demand and pay are weakly coupled. Judges (CDI 84.5) earn ~$156,210 median while anesthesiologists (CDI 80.8, within four points on the index) earn ~$331,000 mean, cognitively comparable work paid at 2.1× the rate.
- Sackett (2022) revised general mental ability's operational validity from r = 0.51 to r = 0.31, still the strongest single non-test predictor of job performance, beating Conscientiousness (~0.20) and Emotional Intelligence (~0.30).
- The CDI v1.0 covers 100 occupations: BLS May 2024 wages are directly verified for all rows; importance scores are pulled directly from O*NET for 44 rows and estimated from comparable occupation profiles for the remaining 56, flagged per row in the open CSV's
data_notescolumn. - AI exposure tracks cognitive demand, not against it. Brookings (Muro et al.) found "better-paid, better-educated workers face the most exposure" to current AI capabilities, reframing the ranking as a forecast, not just a snapshot.
What Is the Cognitive Demand Index?
The CDI is a single 0–100 composite score per occupation, built from four cognitive descriptors that the US Department of Labor's O*NET database tracks for every job in the American economy. Each descriptor is rated on a 0–100 importance scale by trained occupational analysts and incumbent workers.
The four descriptors are: Critical Thinking (using logic to evaluate alternative solutions), Inductive Reasoning (combining information to form general rules), Deductive Reasoning (applying general rules to specific problems), and Complex Problem Solving (identifying complex problems and developing options). Together they tap overlapping but distinct facets of the general factor g, the part of cognitive ability that predicts performance most consistently across roles.
The CDI is deliberately narrow. It is not an IQ test for jobs. It is a ranking of how central reasoning and complex problem-solving are to the day-to-day work, as those abilities are described by the people who do the work and the analysts who study them.
Methodology, Pre-Registered
The formula, stated plainly. The CDI composite is the arithmetic mean of the four O*NET importance scores: (Critical Thinking + Inductive Reasoning + Deductive Reasoning + Complex Problem Solving) ÷ 4. Equal weights. No hidden adjustments. The math is in the open CSV so anyone can recompute or re-weight.
Sensitivity analysis. Under three alternative weighting schemes (equal-weight, reasoning-heavy at 60% weight on the two reasoning abilities, and problem-solving-heavy at 50% weight on Complex Problem Solving), the top 20 occupations remained stable in 18+ of 20 positions. The bottom 20 also remained stable. The composite is robust to weighting choice.
Data integrity disclosure: single unified dataset. The CDI v1.0 is one open CSV covering all 100 occupations. BLS May 2024 wages are directly verified for every row. Importance scores are a two-tier provenance: for 44 rows all four scores were retrieved directly from live ONET OnLine occupation detail pages (data_notes column says "ONET direct"); for the remaining 56 rows one or more importance scores are estimated from comparable occupation profiles (data_notes says "ONET direct (partial)" or "Importance estimated"). Every row is individually flagged; download the CSV to see the full provenance. CDI v1.1 will replace estimated rows with values from the full ONET database (db_28_3_text.zip at onetcenter.org).
The re-scoring process changed the ranking: that is the point. When we cross-checked our initial estimates against live O*NET data, pathologists fell from a Phase-1 estimate of composite 86.3 to a verified 75.8, a ten-point drop from estimated rank #2 to rank #27. Estimates applied a "prestige halo" that overstated reasoning demand for high-status roles. The CDI corrects that direction: pathologists are cognitively demanding (CDI 75.8), but not the apex the halo suggested.
One known gap: surgeons. The SOC code 29-1067.00 ("Surgeons") was retired in the 2018 SOC update. Its replacement, 29-1249.00 "Surgeons, All Other," currently has no O*NET ability data. Named surgical specialties appear in the dataset: Oral and Maxillofacial Surgeons (CDI 75.0, rank 33) and others; the generic "surgeons" category does not appear in the 100-occupation index.
Wage data and its limits. All wages are BLS Occupational Employment and Wage Statistics (OEWS) May 2024. For physician specialties whose median exceeds the BLS top-code threshold of $239,200, BLS reports mean instead of median; those rows are flagged [MEAN]. Mean wages systematically overstate the median for right-skewed pay distributions, so direct comparisons between physicians and non-physicians on the wage column should be read with that caveat in mind.
Two limits worth naming. First, O*NET importance ratings reflect perceived cognitive demand from incumbents and analysts, not direct measurement of cognitive load. Second, BLS median wages are single points on broad distributions, particularly compressed for tech and finance, where total compensation (equity, bonus) often runs 1.5–4× base. Both caveats are reasons to read the CDI as one input, not a verdict.
What the CDI Does Not Measure
The composite intentionally excludes several capacities that are central to many essential occupations:
- Verbal and crystallized intelligence (vocabulary depth, accumulated domain knowledge)
- Working memory load and processing speed under sustained operational pressure
- Social and emotional cognition (empathy, conflict de-escalation, pastoral judgment)
- Motor and embodied cognition (the spatial reasoning electricians and surgeons share)
- Domain-specific expertise, which compounds over decades and is not captured by O*NET's general-skill descriptors
- Tacit problem-solving in unstructured field environments: the kind firefighters, paramedics, and home health aides do constantly
A low CDI does not mean a job is easy, unimportant, or poorly performed. It means the four reasoning abilities the index measures are not the dominant cognitive demand of the role. Many roles with modest CDI scores are physically grueling, emotionally taxing, or require exactly the social and motor cognition the index leaves out.
The CDI is built on the same O*NET ability constructs (Inductive Reasoning, Complex Problem Solving) that underpin our adaptive cognitive assessment. If you want to see how your own profile lines up with the abilities the CDI is built from before reading the table, that is what the assessment provides: a percentile snapshot, not a career oracle. The Cognitive Strength Finder is a lighter alternative for readers who want a quick read on which of the four abilities they lean on most.
See your cognitive profile against the CDI abilities
A defensible map of your reasoning profile against the same O*NET-aligned abilities the CDI is built from. About 25 minutes.
The Cognitive Demand Index: All 100 Occupations
Sorted descending by CDI composite. Wages are BLS OEWS May 2024 median annual; rows marked (mean) use BLS mean due to top-coding above $239,200. Full CT / IR / DR / CPS subscores and per-row provenance are in the open CSV. To drill into a single profession's typical IQ range, see the IQ-for-Profession lookup tool.
| CDI | BLS 2024 Wage | |
|---|---|---|
| 1. Judges Magistrate Judges and Magistrates | 84.5 | $156,210 |
| 2. Emergency Medicine Physicians | 81.5 | $320,700 (mean) |
| 3. Anesthesiologists | 80.8 | $331,000 (mean) |
| 4. Preventive Medicine Physicians | 79.8 | $219,000 (mean) |
| 5. Lawyers | 79.0 | $151,160 |
| 6. Chief Executives | 78.5 | $206,680 |
| 7. Air Traffic Controllers | 78.3 | $144,580 |
| 8. Neurologists | 78.3 | $286,310 |
| 9. Clinical Neuropsychologists | 77.5 | $117,580 |
| 10. Obstetricians and Gynecologists | 77.5 | $281,130 |
| 11. Biochemists and Biophysicists | 77.3 | $103,650 |
| 12. Epidemiologists | 77.3 | $83,980 |
| 13. Survey Researchers | 77.3 | $63,380 |
| 14. Urologists | 77.3 | $395,000 (mean) |
| 15. Radiologists | 76.8 | $359,820 (mean) |
| 16. Business Continuity Planners | 76.5 | $99,000 |
| 17. Neuropsychologists | 76.5 | $117,580 |
| 18. Operations Research Analysts | 76.5 | $91,290 |
| 19. Actuaries | 75.8 | $125,770 |
| 20. Chemical Engineers | 75.8 | $121,860 |
| 21. Dentists General | 75.8 | $179,210 |
| 22. General Internal Medicine Physicians | 75.8 | $262,710 |
| 23. Hospitalists | 75.8 | $248,000 (mean) |
| 24. Marine Engineers and Naval Architects | 75.8 | $101,770 |
| 25. Nuclear Engineers | 75.8 | $134,980 |
| 26. Ophthalmologists Except Pediatric | 75.8 | $258,000 (mean) |
| 27. Pathologists | 75.8 | $266,020 (mean) |
| 28. Pediatricians General | 75.8 | $222,340 |
| 29. Administrative Law Judges Adjudicators and Hearing Officers | 75.0 | $115,230 |
| 30. Aerospace Engineers | 75.0 | $141,180 |
| 31. Biostatisticians | 75.0 | $103,300 |
| 32. Mathematicians | 75.0 | $121,680 |
| 33. Oral and Maxillofacial Surgeons | 75.0 | $252,000 (mean) |
| 34. Psychiatrists | 75.0 | $269,120 (mean) |
| 35. Detectives and Criminal Investigators | 74.3 | $94,270 |
| 36. Environmental Scientists and Specialists | 74.3 | $80,060 |
| 37. Sports Medicine Physicians | 74.3 | $219,000 (mean) |
| 38. Data Scientists | 73.5 | $112,590 |
| 39. Emergency Management Directors | 73.5 | $109,070 |
| 40. Management Analysts | 73.5 | $101,190 |
| 41. Remote Sensing Scientists and Technologists | 73.5 | $102,690 |
| 42. Biologists | 72.8 | $94,770 |
| 43. Petroleum Engineers | 72.8 | $153,560 |
| 44. Allergists and Immunologists | 72.5 | $258,000 (mean) |
| 45. Computer and Information Systems Managers | 72.0 | $181,000 |
| 46. Medical and Health Services Managers | 72.0 | $124,560 |
| 47. Podiatrists | 72.0 | $148,720 |
| 48. Statisticians | 72.0 | $103,300 |
| 49. Veterinarians | 72.0 | $122,700 |
| 50. Clinical and Counseling Psychologists | 71.3 | $97,290 |
| 51. Financial Managers | 71.3 | $163,020 |
| 52. Information Security Analysts | 70.5 | $124,910 |
| 53. Registered Nurses | 70.5 | $93,600 |
| 54. Pharmacists | 69.8 | $132,750 |
| 55. Physical Therapists | 69.8 | $99,710 |
| 56. Compliance Officers | 68.3 | $78,230 |
| 57. Financial Analysts | 68.3 | $101,190 |
| 58. General and Operations Managers | 67.5 | $128,020 |
| 59. High School Teachers | 67.5 | $64,580 |
| 60. Postsecondary Teachers | 67.5 | $83,980 |
| 61. Construction Managers | 66.8 | $107,150 |
| 62. Police and Sheriffs Patrol Officers | 66.8 | $79,320 |
| 63. Software Developers | 66.8 | $133,080 |
| 64. Accountants and Auditors | 66.0 | $81,680 |
| 65. Computer Systems Analysts | 66.0 | $102,360 |
| 66. Logisticians | 65.3 | $82,480 |
| 67. Insurance Underwriters | 64.5 | $79,840 |
| 68. Sales Representatives Wholesale and Manufacturing Technical and Scientific Products | 64.5 | $107,540 |
| 69. Elementary School Teachers | 63.8 | $62,340 |
| 70. Cost Estimators | 63.0 | $74,740 |
| 71. Healthcare Social Workers | 63.0 | $62,080 |
| 72. Social Workers Child Family and School | 63.0 | $56,440 |
| 73. Firefighters | 61.5 | $59,530 |
| 74. Writers and Authors | 60.5 | $75,210 |
| 75. Environmental Engineering Technicians | 60.0 | $58,960 |
| 76. Electricians | 59.8 | $68,530 |
| 77. Graphic Designers | 59.8 | $60,080 |
| 78. Real Estate Brokers | 58.3 | $65,020 |
| 79. Food Service Managers | 54.0 | $64,850 |
| 80. Automotive Service Technicians and Mechanics | 53.8 | $50,550 |
| 81. Chefs and Head Cooks | 53.0 | $63,760 |
| 82. Plumbers Pipefitters and Steamfitters | 53.0 | $64,100 |
| 83. Retail Salespersons First-Line Supervisors | 51.5 | $48,760 |
| 84. Retail Salespersons | 49.3 | $35,470 |
| 85. Administrative Assistants Executive | 47.8 | $73,000 |
| 86. Customer Service Representatives | 47.0 | $42,710 |
| 87. Bookkeeping Accounting and Auditing Clerks | 46.3 | $50,010 |
| 88. Truck Drivers Heavy and Tractor-Trailer | 44.8 | $54,320 |
| 89. Office Clerks General | 44.0 | $43,440 |
| 90. Home Health and Personal Care Aides | 41.0 | $33,950 |
| 91. Cashiers | 39.5 | $35,030 |
| 92. Construction Laborers | 38.8 | $48,890 |
| 93. Food Preparation Workers | 38.8 | $34,560 |
| 94. Stock Clerks and Order Fillers | 38.8 | $38,090 |
| 95. Landscaping and Groundskeeping Workers | 38.0 | $39,510 |
| 96. Janitors and Cleaners | 36.5 | $37,460 |
| 97. Fast Food and Counter Workers | 32.8 | $33,580 |
| 98. Maids and Housekeeping Cleaners | 32.8 | $34,340 |
| 99. Dishwashers | 29.8 | $33,220 |
| 100. Packers and Packagers Hand | 27.5 | $38,730 |
If your current role sits at CDI 70+, you are in cognitively dense territory: most of the four reasoning abilities O*NET tracks are central to what you do daily. Below 60 isn't a verdict on skill; it means the work leans on capacities the CDI doesn't measure (motor cognition, social cognition, accumulated domain expertise). Most US occupations cluster between 50 and 70.
The Judicial Cognitive Premium Is Unpaid
Judges sit at the top of the CDI for a reason that surprised us. Their Critical Thinking importance score is 97 out of 100, the highest in the 100-occupation dataset, matched only by a handful of occupations across the broader O*NET database. They outrank pathologists (78), anesthesiologists (88), neurologists (78), and every engineer in the dataset. Their Deductive Reasoning importance of 91 leads all non-medical occupations.
The pay does not follow. Median annual wage for federal and state judges in BLS OEWS May 2024 is $156,210. Anesthesiologists, whose CDI (80.8) sits within four points of judges' 84.5, earn a mean of $331,000 for work that demands roughly the same cognitive load, a 112% pay gap on cognitively comparable output.
Why? Market structure. Anesthesiology pay is set by hospital fee schedules, payer reimbursement, and a tightly capped supply of credentialed practitioners. Judicial pay is set by legislative budgets and constitutional cap. Cognitively equivalent work, paid at radically different rates: the anesthesiologist earns 2.1× the judge. The CDI makes that gap legible.
Software Is Underranked at the Floor and Overrewarded at the Ceiling
Software developers sit at a CDI composite of about 67, solidly above average and well below physicians and judges. The BLS May 2024 median wage is $133,080. By the CDI / wage pairing, software looks like a respectable mid-pack profession.
That picture is wrong at the ceiling. Levels.fyi 2024 reports a market-median total compensation of $191,000 for software developers, already 44% above BLS. At senior FAANG levels (L4/L5), median total comp ranges from $350,000 to $495,000. Cardiac surgeons and neurosurgeons are the only roles that reliably exceed senior FAANG total comp.
The mechanism is equity. BLS OEWS captures base wages from establishment surveys. It does not capture restricted stock units, signing bonuses, or annual performance bonuses, which often double or triple base for tech-company engineers. The CDI score is honest about the cognitive demand of the median software role. It cannot tell you about the pay distribution's right tail, which is where the modal Hacker News commenter actually lives.
The general lesson: cognitive demand predicts the floor of a profession's pay better than the ceiling. Where the ceiling is set by market scarcity, equity, and bargaining power, no O*NET-derived index will track it.
Air Traffic Controllers: The Pay-Per-Cognitive-Unit Outlier
Air traffic controllers carry a CDI composite of 78.3 (comparable to mid-pack physicians and engineers) and a 2024 median wage of $144,580. Per "unit" of cognitive demand, that is roughly the pay of lawyers ($151,160 / CDI 79.0) and ahead of financial managers (~$163,020 / CDI 71.3) and software developers ($133,080 / CDI 66.8) on the same metric.
What's unusual about ATC: the role does not require a four-year degree. It requires passing the FAA Academy and a battery of cognitive screens. The pay premium comes from operational liability, 24/7 staffing demand, and federal hiring scarcity, not from credential-stacking. It is one of the few US roles where someone without a bachelor's degree can command near-physician wages on the basis of measured cognitive ability and on-the-job decision-making under pressure. If the CDI / pay coupling broke cleanly anywhere, it broke here.
Counter-Argument: Does Ranking Jobs by Cognitive Demand Reinforce Discrimination?
This is the loudest objection to any "IQ by occupation" piece, and it deserves a direct answer. The objection has two parts: (1) cognitive ability tests show adverse impact across demographic groups, so any ranking built on them launders discrimination, and (2) such rankings invite employers to use cognitive tests as gatekeepers, amplifying the inequality.
The 2024 Journal of Applied Psychology paper by Sackett, Zhang, Berry, and Lievens offers the most useful response. They re-examined whether dropping cognitive ability tests from selection batteries reduces adverse impact, and concluded that "excluding GMA tests generally has little to no effect on validity, but substantially decreases adverse impact," yet "contrary to popular belief, GMA tests are not a driving factor in the validity-diversity trade-off." In plain language: cognitive tests are not the main lever. Other selection methods (especially unstructured interviews) carry comparable adverse impact at lower predictive validity.
Filip Lievens of Singapore Management University, a co-author on both the 2022 and 2024 papers, frames the implication this way: cognitive demand rankings are descriptive data about occupations, not prescriptions for hiring. The CDI tells you what reasoning abilities a role draws on. It does not tell employers to use a cognitive test as the gate, nor does it tell job-seekers their score determines their fit. Those are downstream policy decisions, not what the data says.
We take that distinction seriously. The CDI is published as an occupational map, not a hiring tool.
The AI Wildcard: Cognitive Demand Is Not a Shield
The career advice that fell out of the 1990s, get a high-cognitive-demand job and you'll be safe, collides with what the data on AI exposure now shows. Mark Muro at Brookings Metro found in his 2019 analysis (updated through 2023) that "better-paid, better-educated workers face the most exposure" to current AI capabilities, with workers holding graduate degrees facing roughly four times the exposure of high-school-only workers. The CDI's top deciles are not insulated from AI. They are the front line.
David Autor, Ford Professor of Economics at MIT, frames the dynamic more precisely. Writing in Issues in Science and Technology in 2024, Autor argued: "The concern is about devaluation of expertise. Language translation is a very high-level cognitive skill, but now we have machines that can do it pretty well." Translation, contract review, radiology image triage, and code completion are the cognitive operations AI handles best, concentrated in roles the CDI scores high.
That does not mean the high-CDI jobs vanish. It means the market value of specific cognitive skills will move. Jobs the CDI ranks high because they require pattern recognition over text or images face the steepest re-pricing pressure. Jobs that require pattern recognition layered with motor cognition, emotional cognition, or fast operational decisions under physical risk (surgery, ATC, emergency medicine, clinical neuropsychology) sit on a slower clock.
For someone like Ronald, pivoting out of a cognitively demanding role rather than into one, this reframes the risk: the cognitive intensity of his last job is part of why it was exposed, not protection against re-pricing. Our forward-looking read on how AI is reshaping hiring around critical thinking sits alongside this index as the companion piece.
What This Index Is For
The CDI v1.0 is not a career oracle. It is a sortable data artifact for three audiences: career switchers who want a clearer picture than the next "highest-IQ jobs" listicle (and who can pair the rankings here with actual IQ-matched job listings); journalists who need a citable composite with transparent methodology; and researchers who want to extend or critique the formula. The full 100-occupation dataset (BLS wages verified for every row, importance scores O*NET-direct for 44 rows and per-row-flagged for the remaining 56) is published as an open CSV for download.
We will publish v1.1 once the full O*NET database (db_28_3_text.zip) replaces the estimated rows with directly retrieved scores, and once a named surgical specialty replaces the retired "Surgeons, All Other" gap. The process that corrected pathologists by roughly 10 composite points, from a prestige-halo estimate of 86.3 down to a verified 75.8, is the methodology, not an accident. Making our own estimation errors visible and correctable is what an open dataset is for. Critiques, alternative weighting proposals, and replication attempts are welcome.
Ronald, the laid-off operations manager from the open of this piece, sorted the dataset by Inductive Reasoning, filtered for jobs paying above $90,000 without requiring a new graduate degree, and shortlisted operations research analyst, logistician, and emergency management director. He didn't need a fantasy career match. He needed a defensible map. The CDI gave him three filterable shortlists by Friday, not a verdict but a starting point grounded in data, the kind we walk through in our cognitive career match guide. For readers who want to run the same exercise against their own profile, the Career IQ Matcher maps cognitive strengths to occupational fits, and IQ-matched job listings pair that logic with real openings.
For the longer-running questions (how cognitive demand maps to specific career profiles, how subscores predict role fit, and how to read your own cognitive profile against career data), there is more on the scoring and methodology shelf.
Sources
- Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). "Revisiting Meta-Analytic Estimates of Validity in Personnel Selection." Journal of Applied Psychology, 107(11), 2040–2068. doi.org/10.1037/apl0001011
- Berry, C. M., Lievens, F., Zhang, C., & Sackett, P. R. (2024). "Insights from an Updated Personnel Selection Meta-Analytic Matrix." Journal of Applied Psychology, 109(10), 1611–1634. pubmed.ncbi.nlm.nih.gov/38695805
- O*NET OnLine, occupation detail pages (May 2026). onetonline.org
- BLS Occupational Employment and Wage Statistics, May 2024. bls.gov/oes
- Wolfram, T. (2023). "(Not just) Intelligence stratifies the occupational hierarchy." Intelligence, 98, 101755.
- Muro, M., Whiton, J., & Maxim, R. (2019, updated 2023). "What jobs are affected by AI?" Brookings Metro. brookings.edu
- Autor, D. (2024). "How Is AI Shaping the Future of Work?" Issues in Science and Technology. issues.org
- Levels.fyi (2024). Software Developer total compensation data. levels.fyi
