Latin and the SAT

There is some interesting statistical information for Latin teachers and students in the 2010 College-Bound Seniors Total Group Profile Report by the College Board.  It is accessible here: http://professionals.collegeboard.com/data-reports-research/sat/cb-seniors-2010

The information suggests a little inquiry into how students who take Latin fare on the Reading, Mathematics and Writing tests vis-a-vis students who take other languages and vis-a-vis students taking other subjects (not necessarily excluding foreign languages as well, I suppose).  The base line is the average for all 1.5 million test-takers (Table 1):

Table 17 offers SAT mean scores for students taking different foreign languages, both by language and years pursued (including AP/Honors Courses). 

In the total of the mean scores, Latin (1667) is exceeded only by Chinese (1702).  Latin is 50 to 60 points above the national average on each of the three mean scores.  Spanish is almost dead on the national average.  French - the other subject I teach - is in between.  There are almost 11 times the number of Spanish students as Latin Students and almost three times the number of French students.  As one would have suspected students who have taken an AP/Honors course or have pursued the study of a language at least four years or both, have significantly higher reading, math and writing mean scores than those who have not.  Unfortunately this particular sub-set of data is not broken down by language.  Where Latin stacks up with respect to correlations with other course work is more fully developed in Tables 15 - 18 (the others of which I am not going to cut and paste here).  In these the figures indicate that of all the other course work mentioned (ten for English, five for Math, five for sciences, ten for History, eight for the Arts, six for Computers, as well as the nine other Foreign Languages besides Latin), only those taking Calculus (total mean scores: 1748 for 356,931 students) and Chinese (total mean scores: 1702 for 29,231 students) have higher mean SAT scores than were achieved by the 76,904 takers of Latin courses (total mean scores: 1667). 

Just as one comparison - since courses revolving around the operation of computers seem to be proliferating - here is the same set of mean scores correlated with courses in, e.g., computer literacy, computer programming and the like.  Since these are all around or close to the average scores, there would seem to be little correlation between these academic pursuits and better than average SAT scores.

Table 20 offers the same data for students who took various subject tests.  I break this down into two tables, one for the language tests and the other for English, History, Math and Science tests.

This tends to suggest that for students with sufficient confidence in a subject to sign up for a subject test, among foreign language students Latin students are easily in the top position with respect to their mean scores in Reading and Writing, and second only to Chinese students in Math (by only 3 points).  I note that all the scores on this chart are way above the national averages, though again Spanish students are in the bottom position relative to the others.  The Latin students also outscore in their mean Reading, Math and Writing scores all others testing in English, History, Math and Science, with the exception of Chemistry and Physics students who outscore them in Math.  I suspect it is also noteworthy that the standard deviation on the three main tests for those taking the Latin subject test is the lowest of the whole bunch, indicating that these scores are all clustered more tightly around the numerical average. 

One other interesting comparison is found in the tables on the Subject Tests Score Distributions (tables 21 - 24).  I include here only table 24:

Table 23 shows similar distributions for Chinese, French, German and Hebrew.  It looks like Latin students were the least prepared to perform at the highest level on their particular subject test of any students taking any language subject test, and the distribution of their scores was almost even in each 50th from 800 down to 500.  Nevertheless, as was obvious from Table 20, there is something about this particular group of 2,874 students which is correlated with a set of very high mean SAT scores. 

I really don't know what that something is.  There may be something about Latin that leads those who pursue its study for several years to have a higher than average competency level in reading comprehension, mathematical reasoning and writing skills.  On the other hand, those who have higher than average competency levels in reading comprehension, mathematical reasoning and writing skills may gravitate (or be pushed) toward the study of Latin.  Maybe there is something to both explanations.

The other question - related but perhaps more perplexing - is why the number of Latin students is so relatively low.  Note that the number of students taking the Latin subject test is one percent of the total number of students taking subject tests.  Note that the number of students who reported having taken Latin (76,094) is about five percent of the total SAT test takers (1,547,990).  In my small high school (about 600 students fairly evenly distributed in grades 9 - 12), this year (2010-2011) there have been two Latin IV students and one Latin III student, zero French IV students and three French III students - i.e., about one percent in upper level Latin and in upper level French.  There are about 30 French and 30 Latin students altogether over all four grade levels.  There are more than twice as many Spanish students. 

One would think that if high SAT scores really mattered there would be a lot more students taking Latin.  Either hardly anybody has been paying attention to this batch of statistics (and they seem to have been pointing the same direction for years - check out Bolchazy-Carducci's [less than wholly candid] Latin advantage page), or perhaps only about five percent of the population actually has enough competency in reading comprehension, mathematical reasoning and writing skills to take Latin.