Merit-Ranked Class Composition

The post on Why There Are So Few Women at West Point raised our interest in a topic: What would the composition of West Point look like if candidates were admitted and attended based on rank order of candidate scores? Given our previous findings, they would almost certainly look very different than they do now. We asserted, for example, off the ‘back of the napkin’, that the number of women in the classes would decrease significantly. Is that true? Does it affect other groups?

Background

RAND explored this topic earlier in depth with a large dataset over the 1990s-early 2000s. RAND went into detail on the Whole Candidate Scoring (WCS) methodology and tied it to later success in the Army. They also found, among other things, that the candidates who were qualified/offered to attend and did not attend had better WCSs than the cadets who did attend.

An excerpt of their findings is below:

West Point emphasizes and scores three important qualities for each candidate: academic ability, leadership potential, and physical aptitude. Scores on these traits are combined to make up the WCS. An academic ability score is calculated by combining a candidate’s ACT or SAT scores (whichever is the highest in terms of percentile standing) and high school rank,1 making up 60 percent of the WCS. Leadership potential is measured by the community leader score (CLS), calculated by taking the arithmetic average of the AAS, EAS, and FAS. CLS makes up 30 percent of the WCS. The final component is the physical aptitude exam score, which makes up the last 10 percent of a candidate’s WCS. In addition to the WCS, candidates are evaluated on three essays, and every candidate needs a nomination from an eligible sponsor for consideration. Next, we describe each application component.

…

The applicants who declined an admission offer from the USMA have higher application scores than the applicants who accepted…

…

It is possible for a candidate with a higher WCS to be rejected while a candidate with a lower WCS is offered admission to the USMA. This is due to the fact that the WCS, while a central component of the application, is not the only element used in selecting students. Essays, class composition goals, and other factors that we do not observe in the data influence admission decisions.

The opportunity cost of this is not small, as RAND’s table shows. Note the high number of applicants who declined admissions – about a quarter of the overall “offered” pool (that is 4.3K / (12.3K + 4.3K).

But what if those “class composition goals” and other unobserved factors were absent, and cadets truly attended on a rank-ordered basis? We investigate with our class data from 2017-2019.

Unfortunately, we don’t have the formulae the Admissions Committee uses when determining the WCS, or else we’d use those.  However, from RAND’s paper, we do have the breakdown of 60/30/10 Academic / Community Leadership / PAE scoring rubric, as well as SAT and ACT scores, a list of sports, and the PAE results.

Methodology

We started with filtering on the list of cadets noted as “applicant qualified.” It’s unclear how cadets were determined to be “qualified”, but we used that as a starting point. This introduces the caveat that there may have been many candidates who were “qualified,” but didn’t take the step of officially becoming so, so we could be missing viable candidates in the analysis.

We then built a “mini-WCS” assuming that the Academic score counted for 75% of the “mini-wcs”, and the Athletic Activities score and PAE each counted for 12.5%. Standardized test scores received a 75% weight on the academic portion, the class rank received a 25% weight, and we did our best to approximate the point scoring for the Athletic Activities score, although we didn’t have the full range of information that RAND noted that West Point uses – so we stuck with sports and letter counts.

This method of course implies that womens’ and mens’ sports convey the same information about the physical potential of the participants. We will disregard the obvious differences in physical capacity between male and females for this study–for instance, we treat a varsity letter in womens’ softball the same as a varsity letter in mens’ football, although the physical abilities of the players in each sport are very different–although an institution truly focused on winning wars would not.

Our qualified candidate average AAS was 582, where RAND’s was 606, for classes a couple decades earlier – so we feel that we’re in the ballpark there.

We did not include Extracurriculars or faculty appraisals in the analysis since we didn’t have a good rank system to measure them with. Further, we gave points only for qualifying on the PAE. It’s reasonable to point out that we aren’t doing what West Point does, but it’s equally reasonable for us to point out that in the absence of that information, we have to make our own assumptions and model.

Then, we compiled the scores; rank-ordered the cadets; and sorted two sets of cadets per class. The first set is the Rank-Ordered set; that is, what the class would have looked like absent any “class-composition goals.” Then we compared that to the cadets who actually attended in that class. We looked primarily at the categories that West Point promotes in its literature about diversity, gender equality, etc.

Findings

What did we find?

The above chart shows the R(anked) and A(ctual) percentages of each class. A positive delta means the group was over-represented by that much compared to the ranked-method class number of candidates. So, for example, Males are under-represented by 1% compared with their absolute number of Ranked class admittees.

The big takeaways are:

• The Average Class Composition, arrived at through WCS rankings, would be:
  • 82% Male, 18% female
  • 10% Asian, 3% Black, 7% Hispanic, 3% Native / Other, and 77% White
• The latest actual (excluding classes 2020 and ’19) class composition is:
  • 81% male, 19% female
  • 8% Asian, 15% Black, 10% Hispanic, 3% Native/Other, 64% White
• Male and Female: Females are over-represented in Actuals by about 4% of their total, but that’s not near the 15-25% we’d estimated in the last post using only SAT rankings.
  • We found that Athletic scores and the Class Rankings outweigh the aptitude test gap and brought females up near parity in the WCS scoring. (though let us not forget that women compete in womens’ athletics and scoring scales, and men in mens’. If they competed within the same standards, we’d no doubt see large differences.) Specifically, we found that women scored in the 90th percentiles in their high school classes, where men were on average in the 86th percentile. Men scored higher in the Athletic Activities Score (AAS) with an average of 82nd percentile to the female 80th percentile, but as the Class Rank counted for 15% of our overall WCS where the AAS counted for only 12.5%, it helped women more. Whether heavily weighting high school class rank, which doesn’t indicate aptitude as well as standardized testing and is more heavily indicative of skill in conformity and box-checking, identifies better officers is very questionable.
  • Further, this close matching of our methodology with the actual female/male attendance rates seems to validate our WCS ranking methodology, although we can’t say with certainty that class composition goals were the only determinant in the female over-representation.  As a corollary of the above, given the logic of the WCS, the percentage of female cadets is roughly correct. But the 4% over-representation implies that the ceiling of qualified female attendance has been reached at ~20%, barring a dramatic change in the number or quality of the applicant pool.
• Race: “Class composition” goals are very apparent here. USMA over-placed Black cadets by 350%+ and Hispanic and Native cadets by 50%+, while penalizing White candidates by 17% and Asian candidates by 24%. We’ve examined the relative performance of the groups, including academically and in separations, in previous posts. Clearly, diversity goals and Army Sports are more important to West Point than having the best qualified officers.

Summary

We have shown before that the Academy’s admissions are not objectively measuring applicant quality, but are instead targeting various class composition goals to support Diversity, Sports, and Photo-Ops. The demographics of the Admissions page “Meet our Cadets”, depicting 50% female, 50% non-Asian minority–when the overwhelming actual attending majority is white and male–is not accidental.

This post gives further, quantified color to exactly how much this is happening, and which groups are benefiting, and which groups are being penalized.

It’s important to note that this is not behavior without a cost. Taxpayers pay for the Academy, and need the Army to win wars and defend the Constitution. Soldiers depend on the leaders that come out of it. The public ought to be demanding the best it knows it can can get, and not settling for politically-driven experiments or speculation.