[an error occurred while processing this directive]

The Relation between High School Study of Foreign Languages and ACT English and Mathematics Performance

Scott A. Olsen and Lionel K. Brown

THIS study sought to determine whether English and mathematics performance levels of students who have studied a foreign language in high school differ from those of students who have not. In prior research that controlled for variations in student ability, the English and mathematics performance levels of students who had studied a foreign language tended to be higher than those of students who had not (Wiley; Eddy; Bastian; Timpe; Skelton). Our study controlled for student ability by using three categories of high school class rank. Furthermore, to establish a clear reflection of differences related to foreign language study, we classified students by the amount of high school English and mathematics course work they had completed.

We collected our data from 17,451 students who applied for admission to Northeast Missouri State University from 1981 through 1985. The study did not include students who indicated that English was not the language they spoke at home or students with incomplete ACT scores or incomplete ACT student profile data. We also screened the data to detect, and to exclude from the study, student responses that seemed invalid, falsified, or highly improbable (e.g., 32 semesters of foreign language study, ACT scores of 1). While the data on which this study is based are taken from a wide variety of high school students with diverse interests and ability levels, those students do not constitute a cross section of all high school students. Rather, they are more representative of a midwestern population of graduating high school seniors (mostly from Missouri, Iowa, and Illinois) who considered attending a four-year, state-supported university and who chose to take the ACT. Obviously, these data do not represent students who did not graduate, who applied to colleges with characteristics that differed from those of Northeast Missouri State University (e.g., smaller, less selective, urban-based), who chose vocational or technical schools, who did not continue their formal education after high school, or who never took the ACT for whatever reason.

The foreign language variable of primary interest in this study was defined by student responses on the ACT student profile. Specifically, students who had completed at least one semester of course work in any combination of foreign languages were classified as FLYES (Foreign language yes). Similarly, students who had completed no foreign language courses in high school were classified as FLNO (Foreign language no). There were 9,179 FLYES students and 8,272 FLNO students.

To make a clearer comparison of FLYES and FLNO students, we attempted to control for the general level of student ability, the amount of English course work taken, and the amount of mathematics course work taken. First, to indicate general ability, we divided students into three ranks: those graduating in the upper (first) quartile of their high school class, those graduating in the second quartile, and those graduating in the lower half of their high school class. Because students apply for college admission on a self-selective basis, there were only 2,840 students in the lower half of their class, compared with 6,546 in the second quartile and 8,065 in the first quartile. Second, since we assumed that the content and difficulty of high school English and mathematics course work vary greatly, we used simple dichotomies for classification purposes. We place students with more than six semesters of high school English course work in the “more” category ( n = 11,298) and students with six or fewer semesters in the “less” category ( n = 6,153). Similarly, for mathematics course work we defined the “more” category ( n = 11,483) as more than four semesters of course work in mathematics and the “less” category ( n = 5,968) as four or fewer semesters.

To evaluate English performance, we used the English usage test of the ACT assessment battery (ACTE), which measures five elements of standard written English: punctuation (17%), grammar (17%), sentence structure (26%), diction and style (23%), and logical and organization (17%). The test gives an overview of the students' ability to use the language rather than to recall specific rules. The measures of mathematics performance was the mathematics usage subtest score of the ACT assessment battery (ACTM). This test contains questions on arithmetic and algebraic operations (10%), arithmetic and algebraic reasoning (35%), geometry (20%), intermediate algebra (20%), number and numeration concepts (10%), and advanced topics (5%).

The analysis methodology compared ACTE and ACTM means for FLYES and FLNO students across the various classification combinations of class rank (CR), English course work (ENGL), and mathematics course work (MATH). That is, at the most detailed level of analysis, FLYES and FLNO students could be compared on twelve (3*2*2) possible combinations of ability: three categories of class rank (CR-first quartile, CR-second quartile, and CR-lower half), two categories of English course work completed (ENGL-more and ENGL-less), and two categories of mathematics course work completed (MATH-more and MATH-less). For example, one of these twelve possible combinations would compare FLYES and FLNO students who were in the second quartile of class rank and who had “more” English course work but “less” mathematics course work; another would compare FLYES and FLNO students who were in the lower half of their class and who had “less” English course work and “less” mathematics course work; and so forth. The primary statistical methodology we used to compare means was a full-model, four-factor, unbalanced analysis of variance with a 0.05 criterion for significance (for details of this methodology, see Brown).

The study found that FLYES students had significantly higher mean scores on both the ACTE and the ACTM than FLNO students did. For English performance, FLYES students had a mean ACTE of 18.63, while FLNO students had a mean of 16.95. For mathematics performance, the mean ACTM score for FLYES students was 17.37, whereas for FLNO students the mean was 15.85.

Figure 1 illustrate the differences in ACTE means of FLYES and FLNO students for each level of ability as indicated by class rank. As expected, class rank accounted for the greatest variation in scores, regardless of the foreign language classification. But more important, figure 1 shows that at each level of class rank the FLYES student had significantly higher ACTE scores than the FLNO students did.

In comparing ACTE means of FLYES and FLNO students, we also took into account the amount of course work in English and mathematics, and more course work in either area was related to higher ACTE scores. As figure 1 indicates, students who had completed more English and/or more mathematics course work tended to have slightly higher ACTE means, and students who had completed less English and/or less mathematics course work tended to have slightly lower ACTE means. However, the influence of English and mathematics course work on ACTE means did not change our general conclusion about the effects of foreign language study. Thus we found that regardless of ability level as indicated by class rank, regardless of the amount of English course work taken, and regardless of the amount of mathematics course work taken, students who had studied a foreign language performed significantly higher on the ACT than did those who had not studied a foreign language.

Figure 2 displays the mean ACTM scores for FLYES and FLNO students at each class rank. Again, as with the ACTE scores, class rank accounted for much of the variation in ACTM scores, but within each level of class rank those who had studied a foreign language had significantly higher ACTM means than did those who had not.

Comparisons of ACTM means for FLYES and FLNO students are not uniformly interpretable across all twelve combinations of class rank, English course work, and mathematics course work. Specifically, there were virtually no differences in FLYES and FLNO ACTM means for students in the following three classifications: CR-lower half, ENGL-more, MATH-les, CR-lower half, ENGL-less, MATH-more; and CR-second quartile, ENGL-less, MATH-more. In the remaining nine comparisons at this level of detail, however, FLYES students had higher ACTM means that FLNO students did, with differences ranging from 0.524 to 1.226 on the ACT scale. Thus the more general differences between FLYES and FLNO students seen in figure 2 for the lower two classifications of class rank depended on the combination of English and mathematics course work the students had completed. Except for students in the three categories just noted, FLYES students as a whole had higher mathematics performance than FLNO students did.

As an indicator of student academic ability, class rank was a highly significant source of variance throughout this study. It was an important control because of the widespread assumption that since brighter students tend to select foreign languages as electives, higher scores related to language study are really no more than a function of higher ability. This study found that students who had completed a foreign language course tended to have higher ACTE and ACTM scores regardless of their class rank (ability level).

The amounts of previous course work in English and mathematics were also useful controls in this study. Naturally, the ACTE means of students who took more English were significantly higher than those of students who took less, as were the ACTM scores of students who took more mathematics classes. However, students who took more mathematics classes also scored significantly higher on ACTE. Conversely, English course work was not a significant source of variance in ACTM scores, whereas foreign language course work was. Further and more detailed study of interrelations among parts of the curriculum and learning processes might reveal, as suggested by Jarvis, that the mental processing skills required to do mathematics problems are also developed by language processing and vice versa.

The findings of this study are consistent with those of earlier studies, even with the additional controls for course work that the present study offered. Skelton reports in a somewhat similar study that college freshmen who had studied a foreign language scored significantly better on each of six achievement tests. Wiley's study, which controlled for ability by grouping according to both high school cumulative grade point average (GPA) and ACT English scores, found that college students of Latin, French, German, and Spanish who had studied foreign languages in high school all had significantly higher mean college GPAs than did students who had not studied foreign languages in high school. Masciantonio's study of elementary students in Indianapolis also shows significant differences in nine achievement measures, including three in mathematics, after students had taken Latin during only one school year.

One study did not consider how many semesters of foreign language students had in high school, nor did it consider which languages or combination of languages were studied. Given the results of this study, the next logical step is to investigate the possible differential influence these two variables might have on English and mathematics performance.

In conclusion, we should point out that our results are not of the cause-and-effect variety. Other variables—student motivation, for example—may influence the relation of foreign language study to performance in English and mathematics. However, this study does provide evidence in support of those who argue, as does Whitehead, that foreign language instruction should be considered a basic part of the curriculum offerings, not a frill with value to a limited few.

Scott A. Olsen is Associate Professor of Education at Northeast Missouri State University. Lionel K. Brown is Instructor of Spanish and German in the Gallatin R-V High School, and Administrative Assistant in the Gallatin R-V School District, Gallatin, Missouri.

Works Cited

Bastian, Terry R. “An Investigation into the Effects of Second Language Learning on Achievement in English.” English Language Arts Skills in Instruction. Urbana: Clearinghouse on Reading and Communication Skills, 1979. ERIC ED 189 646.

Brown, Lionel K. “The Relationship between Student Achievement as Measured on the ACT English and Mathematics Subtests and Foreign Language Study, Length of Study, and Languages Studied.” Thesis 171. Northeast Missouri State U, 1989.

Eddy, Peter A. The Effect of Foreign Language Study in High School on Verbal Ability as Measured by the Scholastic Aptitude Test—Verbal. Washington: Center for Applied Linguistics, 1981. ERIC ED 196 312.

Jarvis, Gilbert A. “The Value of Second-Language Learning.” Learning a Second Language. Ed. F.M. Grittner. Chicago: U of Chicago P, 1980. 26–43.

Masciantonio, Rudolph. “Tangible Benefits of the Study of Latin: A Review of Research.” Foreign Language Annals 10 (1977): 375–82.

Skelton, Robert B. “High-School Foreign Language Study and Freshman Performance.” School and Society 85 (1957): 203–05.

Timpe, Eugene F. “The Effect of Foreign Language Study on ACT Scores.” ADFL Bulletin 11.2 (1979): 10–11. [Show Article]

Whitehead, Kenneth D. “Foreign Language Study: One of the Basics.” American Education 19.2 (1983): 5–14.

Wiley, Patricia D. “High School Foreign Language Study and College Academic Performance.” Classical Outlook 62.2 (1985): 33–36.

Figure 1:
ACTE mean scores for FLYES and FLNO students by class rank

Figure 2:
ACTM means scores for FLYES and FLNO students by class rank

© 1992 by the Association of Departments of Foreign Languages. All Rights Reserved.

[an error occurred while processing this directive]