"Statistics, Lies and Damned Lies" on Military Recruitment

December 18th, 2005 1:51 PM
What follows is the lede from a Baltimore Sun article today (Sunday) about Army recruitment as reprinted in The Day in New London, Connecticut. It demonstrates, again, the truth of Mark Twain’s dictum, “There are lies, damned lies, and statistics.”

Washington — The Army met its recruiting goal for November by again accepting a high percentage of recruits who scored in the lowest category on the military's aptitude tests, Pentagon officials said Thursday, raising renewed concerns that the quality of the all-volunteer force will suffer.
The Army exceeded its 5,600 recruit goal by 256 for November, while the Army Reserve brought in 1,454 recruits, exceeding its target by 112. To do so, they accepted a “double digit” percentage of recruits who scored between 16 and 30 out of a possible 99 on the military's aptitude test, said officials who requested anonymity.
The article does note, later on, “Still, Army officials continue to say that at the end of the recruiting year, next Sept. 30, the total percentage of Category IV soldiers will be no more than 4 percent.”
Source: http://www.theday.com/eng/web/news/re.aspx?re=127EAF9A-F3BB-4D46-93F7-AD1D0D7C8F1A

What can students of statistics as used in the mainstream media, conclude from this article? There are two errors here. Both are errors of omission. Both are fatal to the article’s premise.

The first is that these are statistics from a single month. The article doesn’t say how this number of recruits compares to a full year of Army recruiting. Give or take a percentage point, it would be roughly 8.3% (1/12th) of the year’s total. Those who follow unemployment statistics, which are reported monthly, know that they go up and down for a variety of reasons. The important statistic is not whether the number has bumped up or down in a given month, but where it stands for the year to date, and for the whole year at its end.

The second point of omission can be deduced from the article, though it would undercut the article’s thesis to state it – so the reporter does not state it. How do the Army recruits compare with the intelligence of the young (mostly) male population from which its volunteers come?

The bottom Category IV from which the Army deliberately takes a limited percentage, is NOT the bottom category among those tested. More importantly, how does this compare to the average ability of all young men in American society?

The Army policy is not to take any young men who do not have their high school diploma or GED (General Equivalency Diploma). The only exception to this rule are a limited number of recruits who pass appropriate exams and, with help, can complete their GED before going into basic training on “delayed entry.” The reporter could have gotten this information with a single click on the Internet. I did. Here’s the source for anyone interested in the complete story:

Source: http://www.armytimes.com/story.php?f=1-292925-1115623.php

Since dropout rates from high schools range as high as 50%, depending on jurisdiction, if this fact had been included, it would be clear that the Army is still recruiting young men who are more able intellectually than their general age group.

So, the conclusion is clear. This article used an accurate statistic (from a single month) to slam the Army. And by leaving out the context, it gives the impression that the Army volunteers are less intelligent than their peers, rather than more so, which is the truth.