Project 100,000: An Analysis

The Gwern article McNamara’s Folly: The Denial of Individual Differences has gone viral many times. I think however he overstates his case. I, probably like many others who read his review or similar articles about Project 100,000, such as on or on Reddit, was under the mistaken impression that McNamara enlisted literal retards (or borderline retards) out of an act of misguided compassion, or that Project 100,000 was an abject failure that exploited the handicapped, but the evidence paints a somewhat more nuanced picture.

I think the examples he gives are cherry picking. He describes a particular ‘moron’, Freddie Hensley, who died:

To emphasize how visible they were, Gregory goes into detail about the one exception: the handsome young Freddie Hensley, who nevertheless was not the sharpest tool in the shed, to the surprise of everyone interacting with him and discovering things like his reflexes being far too slow to shoot a rifle or believing that thunder and lightning were unrelated; Hensley was sent into combat and died

So, yes, plenty of ‘morons’ were put in harm’s way, and some of them died, but this was true of all recruits. Of 2,700,000 American troops sent to Vietnam, 58,000 perished, of which only 5,500 were from Project 100,000. So it’s not like Project 100,000 recruits were singled-out for death. Only the Coast Guard was relatively immune from mass causalities.

The casualty rate of Project 100,000 recruits was comparable or even less than the overall rate. 350,000 Project 100,000 recruits were sent to Vietnam. 5,500 died in service, giving a casualty rate of around 1.6%. Overall, 2,700,000 US troops were sent of Vietnam and 59,000 died, so some elementary math gives a 2.3% casualty rate for the ‘non-morons’:


So Project 100,000 recruits may have actually fared better than average.

Some sources state that only half of Project 100,000 recruits were sent into combat [specifically, 71 percent to the army, 10 percent to the marine corps, 10 percent to the navy, and 9 percent to the air force], in which case the casualty rate would be doubled to 3.2%. However, this is still much less than the 5% casualty rate sustained by the Marine Corps, which had the highest rate of casualties of all the branches, followed by the Army at 2.7%. In other words, project 100,000 recruits were spared an even worse possible fate by not being enlisted in the Marines.

The link above says that Project 100,000 recruits died at a 3x higher rate compared to other GIs (the control group), but I was unable to corroborate this with the math. I think the worse-case calculation shows a 50% higher rate.

This doesn’t necessarily mean they were competent soldiers (the evidence suggests they were not), but they were not sacrificial lambs or cannon fodder as portrayed either.

Project 100,000 recruits were not just selected by IQ, but also included individuals who would have been rejected for other reasons, such as exceeding or failing to meet weight requirements, poor hearing or vision, or having ‘poor moral character’.

Project 100,000 recruits were not retarded or mentally handicapped. So one may see the word moron and be inclined to believe that Project 100,000 recruits were like Bill Murray in Stripes, but this is an exaggeration. During McNamara’s era, the term ‘moron’ was technically defined as someone with an IQ between 50-69. An ‘imbecile’ was defined as someone having an IQ between 20-49, and anyone below that was an ‘idiot’. Moron sounds worse than imbecile, but at the time it was not.

To enlist in the armed forces, an individual must score at least 10% on the AFQT battery, which is like an IQ test. Such scores are percentiles, so a score of 99, which is the highest possible score, implies scoring higher than 99% of test-takers. Category IV enlistees score between 10-30%. Anything below 10% is category V (by comparison, a score of 93-99% is category I), which since 1948 have never been recruited even during Project 100,000. These are the actual morons, not the IV recruits.

Although IV recruits can serve, according to DoD specifications, IV recruits are typically capped at just 4 or so percent of all recruits. However, during shortages, restrictions are typically relaxed, as was the case with Project 100,000 and during WW2. But given that 350,000 Project 100,000 troops served out of 2,700,000, implies that the cap was raised to just 13%, so even without Project 100,000, there still would have been a large number of IV recruits (or ‘morons’) anyway.

However during WW2 a different test was used, the AGCT. Astonishingly (or appallingly, depending on how you view it), 9% and 28% of WW2 recruits fell under categories V and IV respectively, versus 0% and 13% for Project 100,000. Only in 1948 did Congress make it illegal to induct category V enlistees. Thus, the US military has a long history of digging from the bottom of the barrel, so it’s not like what McNamara had done was unprecedented.

The above link also states that Project 100,000 recruits (New Standards men), although they took longer to train, they were not all failures, and were not an overall net-negative:

In combat, their record is less clear. Some performed poorly, some performed well. In a few instances, the failures of New Standards men probably cost their lives and the lives of other Marines. But New Standards Marines did not force leaders to alter their tactics, nor do they seem to have lowered the overall fighting power of Marine units.

In the end, Project 100,000 had almost no impact on theMarine Corps. Project 100,000 did not significantly contribute to the Marine Corps’ disciplinary problems or hamper combat operations in Vietnam. New Standards menplaced a burden on the training system, but this was aburden that the Marine Corps would have had to bear anyway. Given the Marine Corps’ inability to attract better qualified recruits, not long after McNamara’s announcement the Marine Corps would probably have followed the precedent of World War II and Korea, lowering standards to fill its ranks.

Moreover, recruitment shortages forced the Marines to induct more IV enlistees, and even without Project 100,000, standards would have been lowered anyway.

Trying to convert AFQT categories/percentages into IQ scores is somewhat difficult, because the former tends to not work well for extremes of IQ. The AFQT is trying to test for competency for large numbers of people, so this means it’s designed with different objectives in mind and has different ‘norming’ than a conventional IQ test. However, I was able to find a conversion table from 1980 between AFQT and IQ scores:

A correlation of around .8 between the AFQT and the Wechsler Adult Intelligence Test was ascertained, according to a 1974 Air Force Human Resources study.

Gwern writes:

The idea of Project 100,000 was to kill two birds with one stone by drafting recruits who were developmentally disabled, unhealthy, evil, or just too dumb to be conscripted previously: it would provide the warm bodies needed for Vietnam, and use the military to educate the least fortunate and give them a leg up as part of the Great Society’s faith in education to eliminate individual differences and refute the idea that intelligence is real.

According to the table above, category IV recruits have IQs between 72/81-91, which is higher than I would have otherwise guessed. An IQ of 90 is technically considered ‘average’ according to some classifications…a far cry from being developmentally disabled or a moron. And given that the rarity of low IQ falls exponentially as stipulated by the mathematics underlying the Bell Curve, a calculation shows most Project 100,000 recruits likely had IQs in the mid 80s, which is dull or slow but above handicapped, especially given that low ASVB scores were not the sole criterion for induction into Project 100,000, but as discussed above, included other attributes, such as flat feet or poor eyesight, that would have otherwise been disqualifying.


In another study cited by the RAND report, 84 three-man teams from the Army’s active-duty signal battalions were given the task of making a communications system operational. Teams consisting of Category IIIA personnel had a 67 percent chance of succeeding. Those consisting of Category IIIB (who’d ranked in the 31-49 percentile on the aptitude test) had a 47 percent chance. Those with Category IV personnel had only a 29 percent chance.

As the above links shows, although category IV recruits take longer to train and are not as competent at marksmanship and other aspects of soldiering, are not entirely useless or incompetent. The rationale is that during shortages, low-quality troops is still better than not having any. Given that category IV enlistees are still used today, even at double-digit rates during the Iraq war according to the above link, suggests the DoD still has use for them, Project 100,000 not withstanding.

So why was Project 100,000 considered a failure? Because Project 100,000 recruits fared worse in civilian life, after war, compared to a control group of non-veterans. Project 100,000 was supposed to, in theory, serve as sort of launch pad to lift the disadvantaged out of poverty, in which it failed:

Comparisons between Project 100,000 participants and their non-veteran peers showed that, in terms of employment status, educational achievement, and income, non-veterans appeared better off. Veterans were more likely to be unemployed and to have a significantly lower level of education. Income differences ranged from $5,000 [to] $7,000, in favor of non-veterans. Veterans were more likely to have been divorced.

This should not be a surprise though. A study from 1987 showed that Vietnam veterans were more likely to be unemployed, overall, compared to non-veterans due to disability and poor education. So again, similar to war casualties, it’s not like Project 100,000 recruits were that much worse off than others. The inability of Project 100,000 recruits to thrive in civilian life, can probably be explained more by having low IQs than Project 100,000 failing. Yes, in failing to meet impossibly optimistic hopes of lifting the poor out of poverty, Project 100,000 failed, but IQ and unemployment rate and wages are negatively and positively correlated respectively anyway, so having a low IQ is will be an impediment regardless of veteran status.