The 1964 Civil Rights Act in the 21st Century

Rand Paul's remarks about the 1964 Civil Rights Act brought forth lots of talk about libertarians and lunch counters but almost no discussion of how the Civil Rights Act actually works in the twenty-first century. Yesterday provided a nice reminder.   

I won't comment on Lewis v. City of Chicago directly because it was decided on technical matters (the Supreme Court ruled that black firefighters in Chicago did not miss a deadline to argue that a test disproportionately hurt their chances of employment). The underlying facts, however, are of interest not because they are especially unusual but because they are common.  From Fire Law

The case, Lewis v. Chicago, involved alleged discrimination against African American applicants for the Chicago Fire Department who took a test in 1995.

The department set a passing score of 64 on the exam. Applicants who scored at least 64 but below 89 were informed that they passed the test, but would probably not be hired given the number of candidates who scored 89 or above. [26,000 applied and there were only a few hundred jobs, AT]  Applicants scoring 89 and above were classified as “well qualified”.

The majority of “well-qualified” applicants were white. Only 11 percent were black… 

The trial court sided with the black applicants, and ordered the city to hire 132 randomly selected African American applicants who scored above 64. The court also ordered the city to divide backpay owed among the rest of the black applicants.

White, Asian and Hispanic applicants who also scored above 64 but below the 89 standard were not offered employment or backpay.

Perhaps you are wondering about the tests?  You would be hard pressed to find any obvious racial bias.  I haven't found the Chicago test online but you can find similar tests from New York (also the subject of lawsuits) here.  Here is a sample questions from New York.

Nowadays the Chicago fire department simply gives everyone an easy test and then they hire randomly.


Not necessarily firefighters because the are the real deal. But for most government workers we should deny employment to people scoring above 64.

In Sweden there are physical requirements for becoming a firefighter. These tests are fairly tough and selective, as the physical requirements on firefighters sometimes help save lives.

Needless to say, the requirements are lower for women, as it would be "discriminatory" otherwise. Sometimes you just have to wonder about the mental health of people deciding these things...

My own views are that one should be objective in testing and that the test should match the job.

There is one point that is not discussed and that was aluded to by one of the commentators: nepotism in local or city government. I think that has more to do with discrimination than any test, and often there are no tests, just interviews. In library systems, parks and rec, police and fire, if you did a social network analysis or even just a "who is related to whom", you would uncover a network that has bias characteristics because of its initial state. Some of this probably is inherent: you know of an opening, and you tell your niece or nephew. But, I suspect, there is more there than just that.

Again, I am for objective tests that match the job. But, for the reason above.

Well, if there is a passing score, then something should happen if you pass (for example further testing). Moving the passing score after the test implies some kind of foul play. Now, had the fire department said "top 300" or "top 10%" or something similar, such foul play accusations would have less merit.

I'm sure the test does fine as a test of reading comprehension skills, but how does it do as a predictor of real-world firefighting performance?

I mean, I could answer questions like this all day, but it doesn't make me qualified to fight a fire. Why can't the reverse be true? There could be highly skilled firefighters who are not academically gifted, or even borderline illiterate (and since African-Americans generally receive an inferior education in this country, there's obviously going to be more of them among that racial group).

You don't need to put an explicit racial bias in the test because there's already a race/class bias in the education system that you can piggyback on while maintaining a front of objectivity. (See also literacy tests for voting.)

Now in order to provide everyone the freedom to get a job they haven't prepared for everyone has to be an expert in firefighting tests. Great.

Actually it's even more basic than that. Tests are there to reduce the cost of evaluation. But, the hoi palloi still want to believe in this fairy tale of unlimited funds.

If you want to remediate the fact that certain groups don't have skills, please don't do it by sending them into burning buildings.

Scalia did not agree with the candidates but ruled that the law forced him to rule the way he did. Typical Scalia, bad laws give bad results, if you don't like the results tell Congress to change the law.

At one point, and perhaps still, psychological exams for Chicago police officers were race normed. Why? disparate impact on minorities if they were held to the same results as white candidates.

But also because minority candidates had a high failure rate for the drug screen. White candidates with good scores and clout made it through the system. But so few minorities were left on the hiring list the city needed to eliminate the many remaining white candidates for trivial and often arbitrary reasons. Many of these minorities hired actually came from more privileged backgrounds then the white candidates being eliminated.

As far as representing the population of the city, the racial makeup of the police and firefighters often reflect populations that existed twenty years ago. For example, the average police officer will serve 30 years. Since the time he or she was hired the racial makeup of the city has changed but he don't fire officers after every census to achieve racial balance (not yet).

Plus after adjusting for education and criminal records, a minority candidate has a much better chance of being hired. Sadly a large percentage of minority males in our cities have criminal records. BTW white candidates for the police department are usually eliminated for criminal records that would not eliminate a minority candidate (usually non violent crimes). As one Alderman in Chicago claimed, it is almost impossible for an inner city black male to avoid criminal arrest and so they should be held to a different standard when they apply for city jobs. A problem is that screening for criminal records can have a disparate impact on minorities and can you prove that a criminal record prevents people from being good police officers or fire fighters?

"I am getting tired of the Cult of the Emergency Worker."

Me too, especially the drug enforcement cops who are better armed and armored than our troops in Iraq. But not for firefighters. I'm also tired of the cult of the teacher etc. I'm not comparing fire fighters to garbage men, I'm comparing them to social workers. You don't hear firefighters whining nearly as much as these others.

Back to the issue, what is the percentage representation of blacks in government overall? Why single out firefighting just because they at least have the semblance of an objective test?

What does the question as posed have to do with firefighting? E.g., can someone here answer why the 60 minute tank does not give one 28 minutes of work-time and 6 minutes to exit?

The post has a good idea: What is the actual impact in 2010 of the Civil Rights Act?

Well, who wants to discriminate based on race these days? ---the people who favor affirmative action. If we allowed it, I doubt we'd see much discrimination against blacks, but we certainly would see racial hiring of minorities come out in the open instead of being thinly veiled, as it must be now.

foosion: yes, it is irrational. First of all, please give evidence for (1); I did not see anyone argue that there was no connection between test results and firefighting performance--even the plaintiffs seem to focus on the argument that white firefighters were at an advantage because they largely came from firefighting families. (While the example question is silly, I'd question the basic competence of anyone who got it wrong. I would like firefighters to have basic reading comprehension skills.)
Secondly, it depends what you mean by "biased"--disparate impact is not the same as disparate treatment, and I would argue that a test can result in disparate impact without being "biased." The facts of the case as stated above seem to closely parallel Ricci v. DeStefano (which, interestingly enough, was also decided based on the 1964 Civil Rights Act, but went the other way). The debate in that case focused heavily on the disparate impact vs. disparate treatment distinction and how it should affect antidiscrimination law.

Ummm, why shouldn't people familiar with fire-fighting, who even demonstrate it on a test be favored over people who don't?

That isn't racism.

the mask vibrates at pre set levels. If you know how large a tank is and the vibration goes off, you know how much time you have left to get out. The question just asks if you understand that the smaller the tank, the less time to get out after the mask vibrates.

I'm glad other firefighters don't after to worry about looking out for your sorry ass in a fire.


You pay high wages to attract high quality candidates. But of course if you prefer to just give a couple of guys off the street a hose, an axe, and a six pack, I guess they can fight fires.

BTW the Rand corporation did a study on the motivations of people who become police officers. They found very different motivations based on race. White candidates, in general, viewed the job as a calling, higher status, and almost noble (and ran in families). Minority candidates, in general, had a lower view of the job and viewed it as just another government or civil service job - an alternative to the post office


I believe the 36% was for Chicago.

"Nowadays the Chicago fire department simply gives everyone an easy test and then they hire randomly."

That question wasn't easy?

Ignoring the technical grounds that the Supreme Court ruled on, can anyone summarize the argument that the plaintiffs made - how exactly is a test such as this racially biased? What is the argument for this?

Louis Menand had an interesting review of the history of testing for jobs, school admissions, etc -- New Yorker a few years ago. The central issue is that tests rarely measure what you want. Think of SAT tests. Does the test measure relevant skills here?

The court's call was bad. Qualified for the job is qualified for the job. The army has health requirements. Restaurants have cooking skill requirements. Airlines have flying ability requirements.

It would be one thing if everyone in the pool was /roughly equally skilled/. How can you call bias when you get low grades on the test? Tests aren't sentient, unless something has changed since last I checked.

What the firefighters were really arguing about was "equality of upbringing." This one isn't as hard to fix as one may think:

1) Only state-sponsored reproduction is allowed. Individual-choice reproduction would be illegal.

2) Women (and eventually, with improved technology, non-human birth machines) chosen to reproduce to are compensated for their labors, given full pre-natal care packages, and whatever else they need for a successful birth. Women are randomly, blindly chosen based on health factors alone. A few without perfect health records are chosen as well
to avoid a monoculture. While in labor, they stay in a high quality, federal pre-natal facility for monitoring/testing and support.

3) At birth, the new person is entered into a federally run "upbringing" program. Birth parents are never involved.
The program ensures that all children receive the same education, nutrition, and environment, which would be of very high quality relative to your typical suburb or city. Caretakers are rotated throughout the different federal facilities so no caretaker gets undue influence. They are also to work in pairs or groups, and a policy compliance monitor (also rotated) are always present to watch the caretakers. Caretakers would also be very well compensated, but cannot leave the program for a set number of years (built in clawback provisions in their contract, rather than pensions). The number born per year (if needed) would be calculated based on a zero population growth formula.

4) No children are initially allowed to leave the facility until they can pass appropriate "regular" tests and maturity tests. They can start taking the tests at 18 years of age. Those that pass are given (financial, logistical) support for a number of years after they leave to ease their transition. Those that fail a certain number of times are given modest exit package and discharged.

We'd get equality of education, nutrition, security, environment, and culture.

No more inequality of upbringing.

"Well, if there is a passing score, then something should happen if you pass (for example further testing). Moving the passing score after the test implies some kind of foul play. Now, had the fire department said "top 300" or "top 10%" or something similar, such foul play accusations would have less merit."

The problem is, they set the "passing" score lower than the "automatically get considered for the job" score. Anyone who scored over a 65 the city considered qualified to do the job, anyone who scored over 88 entered into a pool from which they randomly selected who they were going to hire. This raises the question: if the people who scored over 65 were considered qualified by the city and fully capable of doing the job, why keep them out of the pool? Why *not* prefer putting all the qualified candidates in the pool, especially when that avoids racially disparate impacts? I mean, it would be one thing if they took the candidate who scored the highest first, then next highest, etc. That might indicate that the test isn't just about proving a base level of competence but is a nuanced aptitude test that realistically ranks candidates by their ability to do the job. But if they are just going to randomly pick the hire out of a pool of people with acceptable scores, as they did, why not make that pool the number of people *with acceptable scores*? By their own admission they appear to be randomly drawing a line in the middle of the qualified candidates, a line that leads to racially disparate outcomes. I don't think there's any problem with forbidding that. It's just a matter of holding people to a robust standard if they argue that disparate racial outcomes in hiring are the result of business necessity.

More broadly, given that the Title VII is designed to function as a prophylactic, I don't think you can just look at the current state of work-place discrimination suits to determine the effects of Title VII. You also have to find a way to look at the universe of discrimination that Title VII is discouraging. Not that I have idea how big that universe is or how to asses it.

Anyways, I take from this post that you are kind dubious about making a hirer prove business necessity if his hiring criteria lead to disparate impacts, and would prefer that the system ignore disparate impacts all together. Do you also oppose Title VII's proscription of disparate treatment? Do you think government should be in the business of trying to fight private racial discrimination at all? If you do accept that their is some basic right to not be irrationally discriminated against in the work place because of your race or that policing that discrimination reaches good outcomes, do you think you can craft legislation that effectively catches the bulk of discrimination WITHOUT putting the burden on the employer to show business necessity in the case of disparate outcomes? It's awfully hard to prove discriminatory intent, after all, even where it exists.

IMHO, intelligence does matter and should be a big factor.

It's called the "effects test." If the effect of an act (here a civil service/firefighter test, more recently subprime mortgage loans)derives a statistical, racial imbalance (note I don't say inequality this isn't about equality it's about advancing balcks), it is discrmination. That is the bogus legal principle applied here. Don't know where that is in the Constitution: that a white man that scored higher than a black is denied life, liberty, property (a job) without due process. But I'm not a lawyer, not a black, not a liberal.

The guvmint does not apply such a 'test' to the NFL. Go figure.

Official reponse for the "highly qualified" white that wasn't hired so Chicago could meet its quota: drop dead.

T. Shaw:

I think the NFL is a good analogy because firefighters need many of the skills that football player do, and I doubt most of the players in the NFL would score in the top 89% on this test. But unlike the city of Chicago the NFL does not use written tests to pick their players, they pick payers that can do what is required.

The 1964 Civil Rights Act is practically irrelevant today. Other than a blatant attempt to subvert Rand (I'm all fine and well with that; the guy strikes me as a dunce), what difference does it make. If the law were gone tomorrow, does anyone here beleive it would be particularly Earth shifting?

Actually it's even more basic than that. Tests are there to reduce the cost of evaluation. But, the hoi palloi still want to believe in this fairy tale of unlimited funds. If you want to remediate the fact that certain groups don't have skills, please don't do it by sending them into burning buildings.

How to become a Chicago firefighter

Minimum Qualifications:

Valid Drivers License
High school diploma or GED
Proof of residency in the City of Chicago (At time of employment)

Submit an application with the City of Chicago Department of Human Resources online at Only apply during an open hiring periods. Submitting an application requires a nonrefundable payment of $20.
Take a Civil Service exam. Dates and times are randomly assigned. No reschedules will be permitted for anyone under any circumstances. A study guide will be given out after you submit a completed application.
Applicants that pass the exam will require a background investigation, medical examination, drug screen, and a Physical abilities test. The Chicago Fire Department provides information on the Physical Abilities Test. A series of Physical Abilities Test Preparation videos are available on YouTube. These were uploaded and created by the Chicago Fire Department.
Earn your spot on the eligible list.

Don't like results sue us

Funny that you made that post immediately after DanC's, DWhite

I think it is good that you are writing about how civil rights laws work today instead of what was happening fifty years ago. From this layman's perspective after reading about the Lewis case and the Ricci case it looks like it is impossible for employers to design a test for job applicants and not risk getting sued.

Universities and the military are given virtual blanket immunity against disparate impact lawsuits, so academics (Alex excepted) tend to be clueless about the current state of discrimination law. For example, economics grad programs are allowed to give substantial weight to the GRE test, even though blacks average at the 14% percentile of the white distribution of scores. Law schools are heavily dependent upon the LSAT, even though blacks average at the 10th percentile of white scores.

For numbers, see:

Likewise, the military. The Pentagon's lower-bound of a 92 IQ for enlistment has dramatic disparate impact on blacks and Hispanics. That's one reason why whites have been dying at a per capita age adjusted rate in the Iraq War about 75% more often the minorities.

In contrast, disparate impact law is a very big deal for much of the civil service and much of private enterprise. It often leads to the quiet adoption of de facto racial quotas to avoid lawsuits under the EEOC's Four-Fifths Rule that says that the worst performing group has to do at least Four-Fifth's as well as the best performing group or the federal government will want to know the reason why.

That's why in 2006, Chicago shifted to an absurdly easy firefighter test that 96% of whites passed, because if 77% of blacks passed, then the city met the Four-Fifths Rule. However, the 2009 decision by Judge Garaufis against the Fire Department of New York in the Vulcan Society case claimed that even that is not a safe harbor against being declared guilty of disparate impact discrimination.

But who wants to talk about discrimination law in 2010 when it's so much more fun to talk about what would happen if the whole country got into a giant time machine and went a half century into the past?

A few thoughts:

1. A quick read of the decision does NOT sound like what Tyler said it was. The SC only ruled on the clock for a 300 day limit would start (the day the policy was announced versus the day the policy was implemented, opting for the latter).

2. The case was remandered back for further consideration which means there is no 'order' to hire 132 blacks. The case will have to be considered in lower courts and, quite possibly, an order to hire 132 blacks may end back up in the SC yet again.

3. In 2009 the SC ruled on the New Haven Fire Department case. That ruling against 'reverse discrimination' makes me very skeptical that what Tyler said would happen could survive scrutiny.

4. Basically from what I've read by following some links on the econolog debate is this: There are no quotas or formal 'reverse discrimination' allowed. There is scrutiny of 'adverse impact'. You compute adverse impact by basically looking at the acceptance rates of various groups (by race, gender etc.) and look at the highest to lowest. The rule of thumb is the lowest should be 80% of the highest or more. BUT that by itself proves only adverse impact which is not in itself against the law. Note wikipedia:

An important thing to note is that adverse impact is not illegal.[6] Adverse impact only becomes illegal if the employer cannot justify the employment practice causing the adverse impact as a "job related for the position in question and consistent with business necessity" (1964/1991 Civil Rights Act, Section 2000e-2[k] [1] [A]). For example, a fire department requiring applicants to carry a 100 lb (50 kg) pack up three flights of stairs. The upper-body strength required typically has an adverse impact on women. The fire department would have to show that this requirement is job-related for the position.

While you can argue that informal quotas might be used to avoid the hassel of getting scrutiny, the fact is the law does not require it, in fact prohibits it. What I suspect is happening here is that for political reasons the fire department would rather not have a 'disparate impact' selection process even if they could legally justify it (and keep in mind a legally justified process that generates a disparate impact does not mean it's the best process). So they would rather juggle criteria around until they can find one that is at least somewhat related to the job but doesn't generate a disparate impact.

5. Now clearly the department can argue that disparate impact is justified here in that the test is reasonably related to the job of being a firefighter (it doesn't seem to have to prove that the test is the best way to measure qualification).

What I would point out though is that the test seems like a good measure of the qualification for teaching firefighting. It also seems like a good measure of applying theoretical knowledge to theoretical problems, reading comprehension, abstract reasoning etc. I'm kind of sketpical that this is actually about being a good firefighter.

A good test IMO would be to put someone in a simulated situation and make them choose three tanks for three different fires. A smaller tank is probably lighter and would be a factor when carrying heavy weights is a major issue. The larger tank probably makes sense if you have to go into a burning building for a long period of time but you're not carrying a lot of other equipment. The good firefighter should be able to make these choices in real time by grabbing a tank....not checking a circle with a #2 pencil.

Which leads me to think the test is hyped because its easy for potential firefighters to 'study' for it and cheap for the city to administer and grade it. My ideal test would be hard to administer and difficult to grade. In the regular private sector a job like this would probably be evaluated using multiple criteria including performance reviews, interviews etc and heavy on real life physical simulated activities. City jobs, though, can't really do that without becomming vulnerable to corruption and patronage so they try to come up with an objective and transparant 'test' to do the picking.

Taking Steve's points about the test seriously (those who pass it are either smart people who can think on their feet or dedicated people who are motivated to study all things firefighting), let's make a few points:

A. This is NOT the army's test. There's no reason to assume high scorers on this test is equilivant to scoring high on the army's test.
B. Is firefighting really a high skill, high leadership job all the time? Should we base its hiring only on a test score? Wouldn't it make sense to use the test to weed out the totally unqualified people & then hire randomly from the population left? Then reevaluate after a year or two based on actual performance. Even if the test does a good job predicting who is going to be a good firefighter it's R2 is probably going to be less than other words there are good firefighters who won't be detected by the test and bad firefighters who will ace the test, a bit of a random element mixes things up a bit.

C. Is dropping a process because it generates a disparate impact really 'reverse discrimination'? IMO you can re-jiggle the selection criteria all you want but at the end of the day you have to have something that is rational (i.e. you can't say 'well take the low scores' unless there's a good reason to believe low scores mean better firefighters) and doesn't generate a disparate impact in itself without justification. White or black you're not entitled to a job as a firefighter. It seems to me the city can play around with their process till the cows come home, even decide not to hire any firefighters provided at the end of the day they use a process that can be defended. In other words, if the city wants to ditch the test score system or move the cut offs up and down they are perfectly free to do so. I would say, though, selecting only black applicants who scored lower is going too far.

It's amazing how many posters in this thread talk about causation in ways that imply that "skin color" or genetics are the cause of disparities, when it's clear to those with functioning brain cells that "upbringing inequality" is the actual cause.

Skin color is irrelevant.

"it's clear to those with functioning brain cells that "upbringing inequality" is the actual cause."

This is utterly false and thoroughly disproven, although a popular and politically correct belief.

Amazing how Boonton could get so much wrong so quickly.

To summarize Vulcan Society:

On July 22, 2009 Judge Nicholas G. Garaufis, a Clinton appointee, gave Bush a belated going-away present by swallowing the Gonzales Department of Justice allegations wholesale and deeming FDNY’s 1999 and 2002 paper-and-pencil employment exams discriminatory against minorities solely on the grounds of Disparate Impact.

Garaufis's opinion is an amusing compendium of the sophistries that comprise the conventional wisdom of Disparate Impact legal theories.

The judge found the tests discriminatory despite being unable to find any evidence of actual discrimination. Even John Coombs, the head of the plaintiff, the black firefighter’s Vulcan Society, couldn’t identify specific problems. Newsday reported:

“Asked to point out questions he considered discriminatory on the exam, Coombs said, ‘I'm not going to answer that. It's irrelevant.’ He added, ‘It's a bad exam when the exam gives you †¦ results that are abysmal for diversity.’ [Federal judge calls FDNY recruiting exams discriminatory, By Michael Frazier, July 25, 2009]

Judge Garaufis allowed Coombs’ Vulcans to join the case midway through. Yet, he banned the main union, the Uniformed Firefighters Association, from participating even though they argued that New York’s Bloomberg Administration was too politically ambivalent to adequately defend firemen’s interests.

Stephen Cassidy, head of the UFA, pointed out:

“Basic intelligence is an important asset for firefighters, and ignoring that fact will imperil the safety of the firefighting force. †¦ The job requires not only physical strength, but also an alert and keen mind. †¦ Firefighters are now extensively trained to deal with hazardous materials, possible terrorism and environmental issues unknown years ago. †¦There is no doubt that intelligence and ability to read and understand are important traits for firefighters.†

The predictive validity of written tests for performance on jobs much like firefighting has been documented by many decades of study by the U.S. military. But the Judge mentions none of this overwhelming evidence.

Garaufis’s decision in Vulcan may seem bizarre in the wake of Ricci. But as I've pointed out, it's much easier for the government to discriminate against whites before they have been hired and get union and civil service protections.

The Bush Administration’s claim that the test “is not job related for the position in question† was laughable. Each question is flagrantly job related.

Thus Diane Cardwell of the New York Times reported on July 23 in "Judge Finds Racial Bias in Fire Dept. Recruiting:"

“New York City used tests that discriminated against black and Hispanic applicants to the Fire Department and had little relation to firefighting, a federal judge in Brooklyn ruled on Wednesday.†

Yet, Cardwell’s accompanying feature article in the same issue of the NYT, "Racial Bias in Fire Exams Can Lurk in the Details," included numerous complaints that the problem with the FDNY test was that it had too much relation to firefighting:

“†¦ firefighter entrance exams have tended to favor applicants already steeped in the ways of the job, like ‘people whose dads and uncles are firefighters,’ said Richard Primus, [Email him]a professor of constitutional law at the University of Michigan. †¦ Besides, Professor Primus added, some of that knowledge is not needed to become a good firefighter. ††¦ some of it tends to be knowledge that “firefighting junkies have, even though it is not really necessary for fighting fires.’†

Those darn “fire buffs† keep studying in their spare time how to save our lives. It’s discriminatory!

Cardwell’s explanation in the NYT of what’s wrong with the test can most charitably be read as heavy sarcasm:

“Each exam consists of 85 multiple-choice questions about firefighting practices: the order in which a firefighter should don gear in an alarm; what the rear of a building would look like, based on its facade; the right situations in which to say ‘mayday’ rather than ‘urgent’ over the walkie-talkie.

“Nevertheless, a closer look shows that the exams also required applicants to read and understand long passages, often containing technical terms, and then answer questions about them.†

Cardwell complains:

“One question, for instance, follows a 250-word description of the use and maintenance of portable power saws †¦†

I’m sorry, but portable power saws, especially the hellacious ones used by firemen to cut through steel and concrete, come with massive instruction manuals much longer than 250 words (owing to decades of product liability lawsuits, as the judge should know).

Why? Because portable power saws can be insanely dangerous. When a relative of mine was a teenager, for example, his chainsaw hit a nail buried in a tree and bucked back into his face.

If an applicant can’t make sense of a 250-word text about power saw maintenance, he might well wind up on lifetime disability.

Let me explain how the Fire Department of New York hiring process worked in 1999 and 2002. (For 2007, even before the Bush Administration sued, the written test was dumbed down to increase diversity.)

A lengthy, intensive system produced quality firefighters with high esprit de corps. (John Derbyshire once told me that in his Long Island neighborhood, blue-collar women consider FDNY guys the most desirable catches as husbands; they tend to be more stable than the similarly well-paid NYPD.)

All applicants took the intensive 50+ page written test. (Study guides were provided beforehand.)

Why start with a written test? They provide a cheap and fair way for the city to drop the deadwood early in the hiring process. Also, rank-order hiring based on test scores speeds up the training of the smart guys who would make good supervisors later in their careers.

For applicants of middling intelligence, those who study firefighting hardest do best. This weeds out those who are either lazy or not committed to firefighting as a career. Moreover, encouraging applicants to study on their own before taking the test gets the winners through the expensive Fire Academy faster and more reliably.

Everybody who attained the passing score (which was a large majority of test-takers) was invited back for a more expensive physical performance test (sort of like the one that Kevin James failed in his attempt to become a New Jersey state trooper in the opening scene of 2009's movie "Paul Blart, Mall Cop").

Scores on the written and physical tests were then averaged and the best performers were called in for medical, psychological, and background checks. The survivors were invited to enroll in rank order at New York City’s 27-acre Fire Academy on Randall’s Island.

In 1999, 90 percent of whites, 77 percent of Hispanics, and 60 percent of blacks scored high enough on the written test to qualify for the physical test. If you use Excel’s Normdist function, you’ll see that the white-black gap is a (very standard) one standard deviation. The pseudonymous statistical analysis La Griffe du Lion has dubbed this one standard deviation white-black gap the Fundamental Constant of Sociology. The white-Hispanic gap was 0.54 standard deviations, which is about normal, too.

In 2002, the passing score was lowered from 85% to 70%, presumably to get around the Equal Employment Opportunity Commission’s notoriously innumerate Four-Fifths Rule. This says that burden of proof is on the employer to disprove illegal discrimination if the lowest scoring group doesn’t pass at a rate at least 80 percent as good as the highest group. That incentivizes employers to lower standards.

Sure enough, the white passing rate to 97 percent, the Hispanic rate to 93 percent, and the black rate to 85 percent. (Hey, EEOC, 85 percent divided by 97 percent is over Four-Fifths!) Yet the racial gaps in standard deviations were only a little narrower (white-black 0.85 and white-Hispanic 0.45). Bizarrely, Judge Garaufis ruled this wasn't good enough because there was still a disparity in outcomes, even though it was less than one-fifth.

Obviously, a test that only three percent of whites flunked is not a very hard test. Still, the good news about the FDNY was that it continued to hire in rank order, so that it got the cream of the crop. In contrast, Chicago, in its efforts to avoid being charged with Disparate Impact, has made the fireman’s hiring test so easy that 96 percent of whites pass; and then hiring is done by lottery. (Or, as cynics have suggested, hiring might be done in order of the applicant’s number of dead relatives who voted for Mayor Daley in the last election.)

Disparate Impact theory is a cover story for corruption, incompetence, and innocent people burning to death.

No wonder the Bush Administration was for it.

"This is a particularly smarmy post. You give an example of hiring for a fire department. What is the relationship between fire department hiring and the broader economy?"

Firefighter cases like Ricci, Vulcan, and Lewis (to name three from the last 12 months), in fact, are merely the tip of the iceberg. There are de facto quotas everywhere, but white firemen are more likely to sue when victimized by them because they have explicit civil service protections that are supposed to make this illegal, and because they are brave.

I think George Mason can set an example for us all by administering the NCAA basketball officials exam to all of their students and selecting those with the 12 highest scores for its varsity team. Let us know when you implement this policy so we can add you to our schedule.

"Won't a test that captures intelligence predict better firefighters, on average?"

Is it really "intelligent" to run into a burning building?

Being cool in dangerous situations is probably more predictive of being a better firefighter...although for those of equal bravery/coolness, I do suspect differential IQ is slightly more predictive of less injuries/deaths.

Actually the correlation between these tests and actual job performance is nowhere as solid as Steve Sailer implies. In simplified terms, what happens is that a relatively narrow range of test scores or IQs ends up in a particular job; and the correlation between actual job performance and test scores turns out to be pretty lousy. So what the statisticians do is they correct for the limited distribution. In essence they imagine they also have morons and genius grade people in that job, and that is how the correlation is solidified.

Most of these tests are objective no doubt, but they fail to measure whatever it is that is needed for the job.

Further, nobody commented on the manufacturers' specs. I provided. The breathing apparatus alarm that is mentioned in the firefighters' test sample question works as per that particular manufacturer based on detection of half-tank full, quarter-tank full, etc. It no more tells you how much time you have than the fuel gauge in your car tells you how many more miles you have. You need the equivalent of a mileage figure.

So, if, as in the spec I quoted, the alarm comes at quarter tank, then if a quarter of the 30 minute tank gives you 6 minutes of breathable air, then a quarter of a 45 minute tank gives you 45/30 * 6 = 9 minutes of air, and a quarter of a 60 minute tank gives you 60/30 * 6 = 12 minutes of air.

You would think that a quarter of a 30 minute tank would give you 7+ minutes; presumably they're building in some safety margin. Notice the way they've done it, the larger tank has a larger safety margin, i.e., on the 30 minute tank they allow 6 of the available 7.5 minutes, i.e., a margin of 1.5 minutes; on the 60 minute tank they allow 12 of the available 15 minutes, i.e., a margin of 3 minutes.

But the point is, IMO, it is far more meaningful to tell/teach that the SCBA apparatus gives an alarm when three-quarters exhausted, and given a safety margin, you have somewhat less than one-quarter of the rated time left of breathable air.

By the way, according to the Chicago Sun-Times:

"The ‘95 exam was drafted by an African American with an eye toward diversifying the Fire Department."

That would be Dr. James Outtz. You can see his picture here:

That's the reality of today, not segregated lunch counters.

Why not just hire some black firefighters to write the test.
Then let everyone take the test.
Anyone want to take bets on the outcome?
What would be the excuse then?

"Why not just hire some black firefighters to write the test."

That's largely what is done with fire and police tests. Consulting firms intensely interview higher ranking blacks about what should be on the test. For example, in the Ricci case, the #2 man in the New Haven fire department, who is black, was willing to testify that he had had much input into the test and found it reasonable, but he wasn't allowed to influence the politicians who threw the test out because of disparate impact.

In the Chicago case the Supreme Court ruled upon this week, the test was written by a black psychologist.

None of this matters except for legal / political CYA. Good tests always come up with about a one standard deviation difference between whites and blacks.

The legal problem is that you aren't supposed to reference that overwhelmingly documented fact. For example, in the Chicago case, the district judge rejected the testimony of an expert hired by the city who testified that his review of 13 meta-analyses of cognitive testing showed that cognitive tests correlate with job performance. The judge argued that the city had to prove cognitive testing would work in Chicago in 1995.

There is considerable academic literature on the correlations between IQ tests and job performance. Likewise the AFQT has been evaluated repeatedly. I can't say the same for the civil service tests, but as previously noted they were introduced to deal with problems of nepotism and favoritism; any proposal to get rid of them should at least contain within it a way to address those problems as well as the civil service tests did.
More generally, your questions obviously don't reflect a desire for answers (which are easy enough to find) - they're just a rhetorical gimmick to make it sound like this is an area with lots of unknowns rather than an area where we have a hundred years of data and research.

'Whether or not blacks score one deviation less on tests really doesn't say anything except about standardized testing.'

It doesn't say anything *to you* because you already know what you want to believe. But the gap is extraordinarily difficult to explain without assuming some difference in average intellectual capability between races.

Good tests always come up with about a one standard deviation difference between whites and blacks.

One more thought, this gives the game away. If the performance of the average black fire fighter is one deviation difference from the performance of the average white one, then you have a justified disparate impact. If, though, you cannot prove that then the test is not justified even if it was designed with the help of black fire fighters, even if it has questions you *think* determine a quality fire fighter.

If (1) there is no demonstrable connection between test results and performance as a firefighter and (2) there is a large difference in black and white performance on the test, then (3) the test would appear to be racially biased.

Does this seem irrational?


First, you're not talking about a biased test; you're talking about biased testgivers.

Second, smart people do everything better than dumb people, so there will always be a connection between intelligence test results and performance as a firefighter, even if you claim it cannot be demonstrated to your satisfaction.

Third, blacks aren't as smart as whites, on average, so (given a sufficient sample size) they will always do worse than whites.

So what?

To avoid losing another disparate impact lawsuit, in 2006 Chicago gave a pass-fail test to tens of thousands of firefighting job applicants that 96% of white test-takers passed it. Firefighters were selected randomly from all those who passed the test.

Would you like your child's life to be in the hands of somebody who scored at the fifth percentile of white guys who'd like to be a fireman?

Would you like your child's life to be in the hands of somebody who scored at the fifth percentile of white guys who'd like to be a fireman?

More importantly, is handing your child to the rookie who scored in the first percentile rather than the five year vetern who scored in the fifth an act premised on a false assumption?

@Alex, you say that "You would be hard pressed to find any obvious racial bias."

Part of the New Haven firefighters case was that study books had been created (some made from previous tests) and passed down in (white) firefighting families.

While this question is purely a math question, you could imagine one that asked e.g. for a tank rated at 60 minutes, how much exit time would you have? The answer being 12 is based on somewhat arbitrary mathematics of safety margins and probabilities (why not 11 or 13 minutes?) ... and if it was written down in those study manuals that only previous (white) firefighters had access to, you have a problem.

And even if this is not the case: what if the exact question itself had been on a previous test?

Comments for this post are closed