Your analysis of the legality of using IQ tests to hire employees is undercooked.
Two cases of the Supreme Court—Griggs v. Duke Power Co. (which you discussed) and Albemarle Paper Co. v. Moody—helped make the use of IQ tests prohibitively expensive. In Albermarle, the Supreme Court applied the relevant EEOC regulations to a company that used the Wonderlic test, the Revised Beta Examination, and the Bennet Mechanical Comprehension Test for hiring:
"The EEOC has issued ‘Guidelines' for employers seeking to determine, through professional validation studies, whether their employment tests are job related. 29 CFR pt. 1607. These Guidelines draw upon and make reference to professional standards of test validation established by the American Psychological Association. The EEOC Guidelines are not administrative regulations promulgated pursuant to formal procedures established by the Congress. But, as this Court has heretofore noted, they do constitute '(t)he administrative interpretation of the Act by the enforcing agency,’ and consequently they are ‘entitled to great deference.’ Griggs v. Duke Power Co., 401 U.S., at 433-34, 91 S.Ct., at 854. See also Espinoza v. Farah Mfg. Co., 414 U.S. 86, 94, 94 S.Ct. 334, 339, 38 L.Ed.2d 287 (1973). Albemarle Paper Co. v. Moody, 422 U.S. 405, 430-31 (1975).
The message of these Guidelines is the same as that of the Griggs case—that discriminatory tests are impermissible unless shown, by professionally acceptable methods, to be ‘predictive of or significantly correlated with important elements of work behavior which comprise or are relevant to the job or jobs for which candidates are being evaluated.’ 29 CFR s 1607.4(c)." [end of Supreme court quote]
The current rules—see 29 CFR section 1607—make using IQ tests for hiring prohibitively expensive, although the regulations allow their use under certain circumstances. Any company that wants to use IQ tests in their hiring process must rigorously test this process and has the burden to prove that the test works. In other words, IQ tests are presumed to cause an “adverse impact” and discriminate in violation of the Civil Rights Act under the current regulations. And if a company fails to meet this burden in court, it could face severe penalties, such as punitive damages. Given that the Court gives “great deference” to these regulations in interpreting the equal employment opportunity provisions of the Civil Rights Act, these regulations are the main hindrance to the adoption of IQ tests in hiring employees.
Wow, the rulers are VERY averse to being open about IQ in the workforce. No doubt the execs would be exempt from such testing, but the biggest egos who are presumed to have slightly above average IQs would still be pissy about having hard proof that many of their employees are gifted. And gifted women? Rage ensues. They already know it by interacting (well, some of them are bright enough to correctly estimate others' IQs ..), but this just chisels it in stone. Owie.
Furthermore, employee discussion about IQ testing would over time reveal that the most globally intelligent people are being *picked on and driven out* by middle and upper management.
Everyone already knows it. This would just make it undeniable.
The rulers really only wanted to utilize IQ testing to identify children whose intellect (and often accompanying sense of justice) would make them a threat someday, to get those children into a special program to see which ones would be obedient vs which would *refuse* to behave as show ponies jumping hoops, and then to shoo the top bootlickers into "intelligence" or "intelligence"-adjacent career paths.
Schmidt & Oh say no, 0.48 reliability for job knowledge vs. 0.65 for IQ, or 0.78 for IQ + integrity tests. Job knowledge tests + IQ have the same reliability, 0.65 as IQ alone.
The ceiling on selection validity for actual hiring practices isn't going to be much higher than the least-reliable required criterion, likely years of experience or education, 0.16 or 0.10, respectively.
The right question is not income but productivity. With mass tests (not individually-administered or computerized adaptive) of decades past and hiring practices that only select for IQ indirectly, income is at most a 0.3 correlation with IQ. (Some disagree, but I doubt their statistical methods; they are trying to prove IQ is important while granting hiring practices a validity they do not have, so they try to get a higher correlation coefficient.) Better tests won't help since the hiring decisions are not made ofln the basis of the tests.
For productivity, the Schmidt & Oh paper I cite in “Nobody wants geniuses”, the validity for hiring selection for IQ is 0.65, and 0.78 with integrity test added, IIRC. (That paper is unpublished, and above Schmidt (same Schmidt) and Hunter's highly-cited prior estimates of 0.5-0.6 going back to the 1980s.) To translate that to productivity, some additional numbers are needed: number of applicants for each job, s.d. of productivity of the job, base rate productivity of the job. The top 2% most highly-paid and cognitively demanding jobs dominate the calculation of potential gains in productivity from hiring from the top on IQ and integrity.
The literature on integrity tests is much more sparse, I think they're trainable, but less advanced in development, so hard to know the future course of their validity, but they added avout 0.13 to the 0.65 IQ validity. Other personality measures will be lower-impact and variable, only conscientiousness has much support, but for more routine jobs. Geniuses will be low conscientiousness and high disagreeability, but highest productivity (if measued right, which is anothet big problem).
Anyway, overall, I think (estimate, guess) if hiring were on the best constructable test criteria to the greatest practical extent, then the IQ-lifetime income correlation would be about 0.5 - 0.6 and the correlation with the criterion including other factors such as personality would be about 0.1-0.15 higher.
I appreciated the distinction you drew at the end between talent, energy, character, and polish that tracks with what I see clinically and organizationally. Over and over, systems over-index on polish and underweight actual cognitive talent and character.
The legality/practicality section was also helpful. In practice, what I see inside organizations isn’t so much “IQ is illegal” as “anything with potential disparate impact feels too risky and too hard to validate properly,” so they retreat into softer, less predictive tools that feel safer but are often less fair.
The piece also quietly surfaces a hard truth: once you start measuring anything well, IQ, work samples, or real productivity, you expose hierarchies people would rather keep blurry. That human discomfort may be just as powerful a barrier as regulation.
High-paying tech companies already select for IQ in software engineer hiring. They just launder it as a time-constrained "live programming exercise".
That exercise has the trappings of a work sample test but is very heavily G-loaded.
100 IQ folks can study and study and not pass. 130+ folks can show up and pass
There has been lots of Goodharting on it, but the (hated) practice survives because it works to select for cognitive ability (often called "problem solving ability" by hiring managers).
That was from s a UK-based survey, right? Live programming interview steps are much more common in the US, though more common in the UK than other European nations.
They're also most prevalent at firms paying closer to $200k/yr for senior SWEs than for firms paying closer to $100k, despite BLS data often showing mean programmer earnings closer to $100k. The programming test is how companies stratify, though most don't understand that. You're right that it's mostly spread via memetics.
For silicon valley firms paying $300k+ it is very very hard to find a firm that doesn't do a challenging live programming interview.
They are pretty darwinian internal hierarchy that selects for performance output, communication and getting along with others - which together is more important than IQ. So, the article’s argument holds.
“Administering work sample tests” is a little iffy: If I recall correctly from the last time I dug into this, work samples refer to samples produced by people literally already on the job. As in, supervisors using “work samples” to predict job performance would mean they are essentially predicting future performance from past performance.
E.g., your Github repo of code does not, I believe, count as a work sample when applying to coding jobs.
If anything this goes to show how powerful IQ (or “GMA”) tests are that they’re basically as good as past job performance.
I like the idea but certain jobs should be female-based and based on a hotness criteria. Stewardesses, waitresses, secretaries, etc. should all be hired on attractiveness and not I. Q. If anything it should a non qualifier. Instead, we live in a system that prioritizes obese hideous women and gay guys. Where are my ditzy blond bimbo stewardesses!
I have a quick question that is not related to the information presented in this article.
I seem to have a wide disparity between the scores that I've obtain from standardized tests such as the PSAT and ACT, vs IQ tests that I've taken online.
On the PSAT and ACT, I've scored in the 85th and 90th percentile respectively. (1230, 29 (23E/30M/30S/32R)
However, on online IQ tests such as CORE, I've scored in the 25th to 35th percentile, 90 - 95.
I'm somewhat nervous of pursuing a bachelor's degree because of the scores that I've obtained, and this article scares me of my bleak future. I wanted to know if there any career paths that someone of my caliber can prosper in.
I'm pretty much open to anything, since I currently work a dead-end job at a small municipal organization.
I apologize if this post comes off as neurotic, but I am very concerned about my future.
Your analysis of the legality of using IQ tests to hire employees is undercooked.
Two cases of the Supreme Court—Griggs v. Duke Power Co. (which you discussed) and Albemarle Paper Co. v. Moody—helped make the use of IQ tests prohibitively expensive. In Albermarle, the Supreme Court applied the relevant EEOC regulations to a company that used the Wonderlic test, the Revised Beta Examination, and the Bennet Mechanical Comprehension Test for hiring:
"The EEOC has issued ‘Guidelines' for employers seeking to determine, through professional validation studies, whether their employment tests are job related. 29 CFR pt. 1607. These Guidelines draw upon and make reference to professional standards of test validation established by the American Psychological Association. The EEOC Guidelines are not administrative regulations promulgated pursuant to formal procedures established by the Congress. But, as this Court has heretofore noted, they do constitute '(t)he administrative interpretation of the Act by the enforcing agency,’ and consequently they are ‘entitled to great deference.’ Griggs v. Duke Power Co., 401 U.S., at 433-34, 91 S.Ct., at 854. See also Espinoza v. Farah Mfg. Co., 414 U.S. 86, 94, 94 S.Ct. 334, 339, 38 L.Ed.2d 287 (1973). Albemarle Paper Co. v. Moody, 422 U.S. 405, 430-31 (1975).
The message of these Guidelines is the same as that of the Griggs case—that discriminatory tests are impermissible unless shown, by professionally acceptable methods, to be ‘predictive of or significantly correlated with important elements of work behavior which comprise or are relevant to the job or jobs for which candidates are being evaluated.’ 29 CFR s 1607.4(c)." [end of Supreme court quote]
The current rules—see 29 CFR section 1607—make using IQ tests for hiring prohibitively expensive, although the regulations allow their use under certain circumstances. Any company that wants to use IQ tests in their hiring process must rigorously test this process and has the burden to prove that the test works. In other words, IQ tests are presumed to cause an “adverse impact” and discriminate in violation of the Civil Rights Act under the current regulations. And if a company fails to meet this burden in court, it could face severe penalties, such as punitive damages. Given that the Court gives “great deference” to these regulations in interpreting the equal employment opportunity provisions of the Civil Rights Act, these regulations are the main hindrance to the adoption of IQ tests in hiring employees.
Wow, the rulers are VERY averse to being open about IQ in the workforce. No doubt the execs would be exempt from such testing, but the biggest egos who are presumed to have slightly above average IQs would still be pissy about having hard proof that many of their employees are gifted. And gifted women? Rage ensues. They already know it by interacting (well, some of them are bright enough to correctly estimate others' IQs ..), but this just chisels it in stone. Owie.
Furthermore, employee discussion about IQ testing would over time reveal that the most globally intelligent people are being *picked on and driven out* by middle and upper management.
Everyone already knows it. This would just make it undeniable.
The rulers really only wanted to utilize IQ testing to identify children whose intellect (and often accompanying sense of justice) would make them a threat someday, to get those children into a special program to see which ones would be obedient vs which would *refuse* to behave as show ponies jumping hoops, and then to shoo the top bootlickers into "intelligence" or "intelligence"-adjacent career paths.
The military also, basically, hard selects for IQ for any specific role using ASVAB scores:
https://www.mometrix.com/academy/wp-content/uploads/2024/02/United-States-Military-Jobs-and-which-ASVAB-Scores-Qualify.pdf
Job knowledge tests beat general intelligence tests. However they usually require specialized skills to make so most companies don't use them.
Schmidt & Oh say no, 0.48 reliability for job knowledge vs. 0.65 for IQ, or 0.78 for IQ + integrity tests. Job knowledge tests + IQ have the same reliability, 0.65 as IQ alone.
See: table 1, p. 65 https://home.ubalt.edu/tmitch/645/session%204/Schmidt%20&%20Oh%20validity%20and%20util%20100%20yrs%20of%20research%20Wk%20PPR%202016.pdf
(Same Schmidt as Hunter & Schmidt).
The ceiling on selection validity for actual hiring practices isn't going to be much higher than the least-reliable required criterion, likely years of experience or education, 0.16 or 0.10, respectively.
If we had perfectly reliable tests of IQ and personality, what percent of the variance would they explain with regard to lifetime income?
Just more than half, may post on this later
The right question is not income but productivity. With mass tests (not individually-administered or computerized adaptive) of decades past and hiring practices that only select for IQ indirectly, income is at most a 0.3 correlation with IQ. (Some disagree, but I doubt their statistical methods; they are trying to prove IQ is important while granting hiring practices a validity they do not have, so they try to get a higher correlation coefficient.) Better tests won't help since the hiring decisions are not made ofln the basis of the tests.
For productivity, the Schmidt & Oh paper I cite in “Nobody wants geniuses”, the validity for hiring selection for IQ is 0.65, and 0.78 with integrity test added, IIRC. (That paper is unpublished, and above Schmidt (same Schmidt) and Hunter's highly-cited prior estimates of 0.5-0.6 going back to the 1980s.) To translate that to productivity, some additional numbers are needed: number of applicants for each job, s.d. of productivity of the job, base rate productivity of the job. The top 2% most highly-paid and cognitively demanding jobs dominate the calculation of potential gains in productivity from hiring from the top on IQ and integrity.
The literature on integrity tests is much more sparse, I think they're trainable, but less advanced in development, so hard to know the future course of their validity, but they added avout 0.13 to the 0.65 IQ validity. Other personality measures will be lower-impact and variable, only conscientiousness has much support, but for more routine jobs. Geniuses will be low conscientiousness and high disagreeability, but highest productivity (if measued right, which is anothet big problem).
Anyway, overall, I think (estimate, guess) if hiring were on the best constructable test criteria to the greatest practical extent, then the IQ-lifetime income correlation would be about 0.5 - 0.6 and the correlation with the criterion including other factors such as personality would be about 0.1-0.15 higher.
I appreciated the distinction you drew at the end between talent, energy, character, and polish that tracks with what I see clinically and organizationally. Over and over, systems over-index on polish and underweight actual cognitive talent and character.
The legality/practicality section was also helpful. In practice, what I see inside organizations isn’t so much “IQ is illegal” as “anything with potential disparate impact feels too risky and too hard to validate properly,” so they retreat into softer, less predictive tools that feel safer but are often less fair.
The piece also quietly surfaces a hard truth: once you start measuring anything well, IQ, work samples, or real productivity, you expose hierarchies people would rather keep blurry. That human discomfort may be just as powerful a barrier as regulation.
Love the thumbnail for this one
High-paying tech companies already select for IQ in software engineer hiring. They just launder it as a time-constrained "live programming exercise".
That exercise has the trappings of a work sample test but is very heavily G-loaded.
100 IQ folks can study and study and not pass. 130+ folks can show up and pass
There has been lots of Goodharting on it, but the (hated) practice survives because it works to select for cognitive ability (often called "problem solving ability" by hiring managers).
>High-paying tech companies already select for IQ in software engineer hiring. They just launder it as a time-constrained "live programming exercise".
Empirically, SWEs only have a mean IQ of 111, so it's either not a common practice or an effective one.
That was from s a UK-based survey, right? Live programming interview steps are much more common in the US, though more common in the UK than other European nations.
They're also most prevalent at firms paying closer to $200k/yr for senior SWEs than for firms paying closer to $100k, despite BLS data often showing mean programmer earnings closer to $100k. The programming test is how companies stratify, though most don't understand that. You're right that it's mostly spread via memetics.
For silicon valley firms paying $300k+ it is very very hard to find a firm that doesn't do a challenging live programming interview.
They are pretty darwinian internal hierarchy that selects for performance output, communication and getting along with others - which together is more important than IQ. So, the article’s argument holds.
“Administering work sample tests” is a little iffy: If I recall correctly from the last time I dug into this, work samples refer to samples produced by people literally already on the job. As in, supervisors using “work samples” to predict job performance would mean they are essentially predicting future performance from past performance.
E.g., your Github repo of code does not, I believe, count as a work sample when applying to coding jobs.
If anything this goes to show how powerful IQ (or “GMA”) tests are that they’re basically as good as past job performance.
"work samples" and "work sample tests" are different. Work sample test is doing part of the job.
I like the idea but certain jobs should be female-based and based on a hotness criteria. Stewardesses, waitresses, secretaries, etc. should all be hired on attractiveness and not I. Q. If anything it should a non qualifier. Instead, we live in a system that prioritizes obese hideous women and gay guys. Where are my ditzy blond bimbo stewardesses!
I have a quick question that is not related to the information presented in this article.
I seem to have a wide disparity between the scores that I've obtain from standardized tests such as the PSAT and ACT, vs IQ tests that I've taken online.
On the PSAT and ACT, I've scored in the 85th and 90th percentile respectively. (1230, 29 (23E/30M/30S/32R)
However, on online IQ tests such as CORE, I've scored in the 25th to 35th percentile, 90 - 95.
I'm somewhat nervous of pursuing a bachelor's degree because of the scores that I've obtained, and this article scares me of my bleak future. I wanted to know if there any career paths that someone of my caliber can prosper in.
I'm pretty much open to anything, since I currently work a dead-end job at a small municipal organization.
I apologize if this post comes off as neurotic, but I am very concerned about my future.