Equity and Excellence in the College Board Advanced Placement Program
by William Lichten - January 16, 2007
This paper uses college standards to evaluate the advanced placement (AP) program. The College Board’s claim that a score of 3 “qualifies” disagrees with the facts of college acceptance. The pass rate has dropped from 51 percent in 1998 to 39 percent in 2006. More telling is the incremental pass rate of 29 percent, which reflects the changes over the 1998-2006 period. By objective measure, the expansion of AP courses into inner city schools has failed: African American and Mexican American (AP Spanish excepted) minorities have an incremental pass rate near 10 percent. These shortcomings, which contradict the claim that AP is “for everyone,” call for a reform of AP admissions policy.
Equity, Excellence And Scaling Up
Advanced Placement began focused on excellence for bright prep school students headed for Ivy League colleges. Today, in search of equity, it has scaled up by reaching out to underserved groupsespecially racial minoritiesand expanded to one of the largest programs in American educational history.
This paper follows up an earlier, critical assessment of the AP program (Lichten, 2000). Based on college acceptance, the earlier paper showed a growing failure rate of the program, especially among minorities; indicated that a score of 3 no longer qualified; argued that the goal of 10 AP courses in every high school was unrealistic; found government mandates and earmarks for AP to be dubious; and the need for quality, not quantity of examination papers.
The SAT and AP Programs Compared
Both the AP and SAT initially served an elite, Ivy League population, but now reach a much wider group. The SAT presented two problems: the first was a downward drift in SAT scores over the years. In successful response, (Dorans, 2002) recentered SAT tests to 500, with a standard deviation of 100. The second difficulty is the black-white test gap, which is nearly a century old and remains unsolved (Shuey, 1966; Osborne & McGurk, 1982; Jencks & Phillips, 1998). How this gap affects the SAT and AP score distributions is shown in Figures 1A and 1B.
Figure 1A. SAT writing score distributions for the Nation and for African Americans. The vertical line shows the prediction of AP from SAT scores. Source: College Board 2006F, College Board 2006G.
Figure 1B. AP test score distributions for the Nation and African Americans. Source: College Board. (2006H)
TRENDS IN COLLEGE ACCEPTANCE OF ADVANCED PLACEMENT CREDIT
The average AP cut score has moved up about half a point between 1998 and 2006. Thus, acceptance scores of non-selective colleges have switched from 3 to a mixture of 3 and 4; selective colleges from a mixture of 3 and 4 to largely 4; and highly selective colleges from 4 to 4 and 5. (At Stanford, even a 5 often is not accepted. For details, see College Board, 2006D.) In 1998, 55 percent of AP exams with a 3, and overall, about half of AP exams passed (Lichten, 2000) The pass rate for exams with a 3 sank in 2006 to 30 percent and the overall value has dropped to 39 percent. Two factors account for the decrease of the passing rate. One arises from colleges raising their cut scores. The other comes from program expansion (Lichten, 2000).1 As AP grew, average scores dropped, with mean grades in 1986 of 3.10; 3.00 in 1996; and 2.89 in 2006 (College Board, 2006B).2 Figure 2 shows one reason for this decline. The fastest growing score was 1 (the score for a blank exam booklet). (College Board, 2006B)
Fig. 2. Growth of number of AP test scores. The fastest growing score is a 1. (Sources, Table 1 and College Board, 2006B).
The number of exams taken has grown exponentially, tripling every decade (Figure 2, see also Lichten, 2000, Table 6), but the number of passed exams lags (Lichten, 2000). One outcome of this and earlier studies is that a 3 on the AP exam clearly does not qualify for advanced placement (Calculus BC still is an exception), a result that is well known to teachers and college faculty (Haswell, 2003, College Board, 2006D). Table 1 compares College Boards claimed qualification scale with one based on the 2006 sample of representative colleges. The claims for passed exams exceed the data by almost 100 percent.
Table 1. AP scales compared. The College Board claims approximately twice as many students passing as in the present data.
Married in haste, we may repent at leisure.
Recently, the College Board began to heed outside criticisms of AP quality (Bleske-Rechek, Lubinski & Benbow, 2004; Dougherty, Mellor, & Jian, 2005; Geiser & Santelices, 2004- for a rebuttal see Camara & Michaelides, 2005; Klopfenstein, 2003; Klopfenstein & Thomas, 2006; Lichten, 2000- for a rebuttal, see Camara, Dorans, Morgan, & Myford, 2000; Lichten, 2001; National Research Council, 2002; Sadler, 2006). The College Board (2006E) plans to update its science curriculum. Trevor Packer, the AP Director, conceded, Theres a lot of fake AP out there, in recognition of the problem of courses improperly labeled as AP. The College Board will require prior approval of high school AP course syllabi (College Board, 2006C). It and ACT have recognized the need to improve college preparatory curriculum and teaching (ACT, 2004, College Board Pre-AP).
Put your best foot forward. Students do not send all their AP test scores to colleges. A failing grade is hardly a winner for college admission.3 From 2005 to 2006 the total number of AP tests taken grew from 2,105,800 to 2,312,600, an increase of 206,800 (College Board, 2006F). But the number of test scores students forwarded to their college grew from 1,210,100 to 1,241,800, an increase of 31,700. By the measure of scores forwarded by students, only 15 percent of the growth in number of examinations was useful. An estimate of the 2005-2006 growth of the number of passing exams is 47,000, 23 percent of the total growth. Both the number of submitted scores and percent passing show that AP has reached diminished returns even earlier than predicted (Lichten, 2000).
The Failure of AP in inner cities
Despite the warning that the AP program in non-selective, inner city schools would have high failure rates (Lichten, 2000, Lichten and Wainer, 2000) the College Board went full steam ahead. Its optimistic slogan, aimed at minorities and unsupported by facts, was that AP isnt just for top students headed for college. AP offers something for everyone (College Board, 2004). The most common score in 2006 for African Americans was a 1 (Fig. 1A); the 2006 pass rate (calculated from table 1, right hand column) was 16 percent. Likewise, for Mexican Americans the pass rate was 28 percent; for non-Spanish courses, the pass rate was 16 percent. In view of these figures, the something offered was the experience of failing. To gauge the value added by the 1998-2006 growth, the incremental percentages of passing exams was 11 percent for African Americans, 20 percent for Mexican Americans, and 12 percent for Mexican Americans without Spanish credit. As judged by these results, the failure resulting from the 1998-2006 expansion of AP into the inner cities was nearly total. Thus, the College Boards claim in its adsthat it stands for equal educational opportunityis a chimera. A failing program is hardly an opportunity.
The College Board (2006A, p. 10). has acknowledged the serious problem of minority AP performance. However, its response is to advocate more AP test expansion, which would make the problem go from bad to worse. This paper has shown that a growing majority of all AP tests failnot just those taken by minority members. The present practice of rushing unprepared high school students into college level courses accomplishes little or nothing positive (National Research Council, 2002)4 and can be harmful (Bleske-Rechek, Lubinski & Benbow, 2004)
WHY AP CANT BE FOR EVERYONE
Introductory college courses are at the level of the average student entering college. Less well-prepared students require remedial instruction. Only a small minority of students are advanced enough to do college-level work in high school. This common sense argument is supported by College Board studies connecting PSAT scores to AP success (Camara & Millsap, 1998, Ewing, Camara & Millsap 2006, see also Lichten, 2001). Similar results were found for ACT scores and college success (ACT, 2004). What the majority of college-bound students need in high school is college preparation, not college imitation. Algebra II is more important than calculus to this group (ACT 2004; Adelman, 1999; Adelman, 2006; Bleske-Rechek, Lubinski, & Benbow, 2004; Education Trust, 1999; Lichten, 2001).
Camara and coworkers (Camara & Millsap, 1998, Ewing, Camara & Millsap 2006, see also Lichten, 2001, Lichten and Wainer, 2000) have shown that the best known predictor of a persons AP score is the PSAT (or SAT) score. The Spearman Brown formula (Anastasia, 1997, p. 96) leads to a further result, which holds for AP, SAT, ACT, or other large scale data: the average performance on the AP test for a large group of people (such as occurs in AP ethnic data) can be predicted without statistical error from that groups average SAT score.
To illustrate, figures 1A and 1B show the AP pass rate for African Americans. Figure 1A contains a vertical line, which cuts the nations SAT data at the 39 percent pass percentages for national AP exams. The predicted pass rate for African-Americans is at 15 percent, in exact agreement with the data. Thus, the high AP failure rate for African Americans was predicted from data furnished by the College Board (Camara & Millsap, 1998, Ewing, Camara and Millsap, 2006).
Unfortunately, College Board management seems to have ignored its own data in recent policy decisions. Even worse is the contrast between College Board research, which is often excellent, and AP advertising, which this paper has shown to be misleading and irresponsible.
SUMMARY AND COMMENTS
With burgeoning growth of AP came slipshod teaching, careless administration, curriculum stasis, and an escalating failure rate. Reforms underway could solve some of these problems; moreover, rather than the present anarchy, it would be important to control admission to AP. Selective colleges are careful to admit the most promising applicants. It makes no sense that the AP program, which is based on courses from these colleges, has no standards for admission.
The PSAT, SAT, or ACT tests are widely used, are often the most effective predictors of success in the AP program, and could be used to guide admission to it.5 If limits were set to achieve the College Boards claimed success rate of ca. 75 percent (Table 1), a million tests would still be taken.
The Sorcerers Apprentice
The sorcerers apprentice used his masters spell to haul water but couldnt find the magic word to stop the flood (Wikipedia, 2006). Advanced Placement is the magic phrase which loosed a deluge of test booklets, some good and most bad. What we now need is the command to rescue us from our folly.
1. The authors values for the decrease in PSAT-V scores are incorrect (Lichten, 2000). For corrections, see Camara, Dorans, Morgan, & Myford (2001). The small samples of colleges used in this paper and preceding papers are subject to sampling errors of a few percentage points.
2. These declines conflict with the Boards claim that the average AP performance has not slipped (College Board, 2006A, p. 6). The reason for this discrepancy is that the College Board argument is based only on the multiple-choice section of the AP tests; the final grade includes both multiple choice and written parts. The College Board fails to mention this important omission. It follows that writingan essential hallmark of the quality of the AP testsis slipping (Haswell, 2003).
3. Because students send only somepresumably the best possibleof their exam results to the colleges, 43 percent of submitted exam scores qualify, compared to 39 percent of all taken exams.
4. College Boards programs of Vertical Teaming and Pre-AP classes might be useful for those students with the ability to take AP.
5. A recent report recommended an examination at the end of 10th grade, which, if passed at a high level, would admit students to courses such as AP and the International Baccalaureate (National Center on Education and the Economy, 2006).
ACT. (2004). Crisis at the core. Iowa City: ACT. Retrieved January 14, 2007 from http://www.act.org/path/policy/reports/cisis,html
Adelman, C. (1999). Answers in the toolbox. Washington: U.S. Department of Education.
Adelman, C. (2006) The toolbox revisited. Washington: U.S. Department of Education. Retrieved January 14, 2007 from www.ed.gov/rschstat/research/pubs/toolboxrevisit/index.html
Anastasi, A. (1997). Psychological Testing (7th ed.). Upper Saddle River, NJ:Prentice-Hall.
Bleske-Rechek, A., Lubinski, D.,& Benbow, C. P. (2004). Meeting the educational needs of special populations. Psychological Science, 15(4), 217-224.
Camara, W., Dorans, N, Morgan, R. & Myford, C. (2000, August 1). Advanced Placement: Access not exclusion. Educational Policy Analysis Archives, 8(40). Retrieved January 14, 2007 from http://epaa.asu.edu/epaa/v8n40.html.
Camara, W. J. & Michaelides, M. (2005). AP use in admissions: A response to Geiser and Santelices. Retrieved January 14, 2007 from http://www.collegeboard.com/research/pdf/051425Geiser_050406.pdf.
Camara, W. J., & Millsap, R. (1998). Using the PSAT/NMSQT and course grades in predicting success in the Advanced Placement Program. Research Report 98-4. New York: College Board.
College Board. (2004). Get With the Program. Retrieved January 14, 2007 from http://apcentral.collegeboard.com/apc/public/repository/apminoritybroeng2_10396.pdf.
College Board. (2006A). Advanced Placement report to the nation. New York: College Board. Retrieved January 14, 2007 from http://apcentral.collegeboard.com/documentlibrary.
College Board. (2006B). AP Grade distributions-all subjects 1987-2006. Retrieved January 14, 2007 from http://apcentral.collegeboard.com/apc/public/repository/ap06_grade_dist_all.pdf
College Board. (2006D). AP Credit Policy. Retrieved January 14, 2007 from http://apps.collegeboard.com/apcreditpolicy/index.jsp
College Board. (2006E). National Science Foundation awards $1.8 million to College Board to redesign AP® science courses. Retrieved January 14, 2007 from http://www.collegeboard.com/press/releases/51572.html.
College Board. (2006F). College report of AP examinations 2005-2006. Retrieved January 14, 2007 from http://apcentral.collegeboard.com/apc/public/repository/ap06_college_rpt_state.pdf
College Board. (2006G). Percentile Ranks. Retrieved January 14, 2007 from http://www.collegeboard.com/prod_downloads/highered/ra/sat/CR_M_%20W_PercentileRanksGenderEthnicGroups.pdf.
College Board. (2006H). AP: Program Summary Report. Retrieved January 14, 2007 from http://www.collegeboard.com/student/testing/ap/exgrd_sum/2006.html
College Board. (2006I). College Report of AP Examinations 2005-2006. Retrieved January 14, 2007 from http://apcentral.collegeboard.com/apc/public/repository/ap06_college_rpt_state.pdf
College Board Pre-AP Program. (n.d.). Retrieved January 14, 2007 from http://apcentral.collegeboard.com/apc/public/preap/index.html.
Congreve, W. (1693). The Old Bachelor. London: P. Buck.
Dorans, N. (2002). The recentering of SAT scales and its effects on score distributions and score interpretations. RR-02-04. New York: College Board. Retrieved January 14, 2007 from http://www.collegeboard.com/research/abstract/20701.html?pdf=/research/pdf/200211_20702.pdf&store=993941
Dougherty, C., Mellor, L., & Jian, S. (2006). The relationship between Advanced Placement and college graduation. Austin, TX: National Center for Educational Accountability. Retrieved January 14, 2007 from http://www.just4kids.org/en/files/Publication-the_Relationship_between_Advanced_Placement_and_College_Graduation-02-09-06.pdf
Education Trust (1999). Thinking K-1, vol. 3, Issue 1, Spring, 3-23.
Ewing, M., Camara, W.J., & Millsap, R.E. (2006). The relationship between PSAT/NMSQT scores and AP examination grades: A follow-up study. Research Report 2006-1. New York: College Board.
Geiser, S. & Santelices, V. (2004). The role of Advanced Placement and honors courses in college admissions. University of California at Berkeley: Center for Studies in Higher Education. Retrieved January 14, 2007 from http://cshe.berkeley.edu/publications/docs/ROP.Geiser.4.04.pdf
Haswell, R. (2003). Raising the cut-off score on the College Board Advanced Placement examinations in composition and literature. Retrieved January 14, 2007 from http://comppile.tamucc.edu/raisingAP.htm,
Jencks, C., Phillips, M. (1998). The Black-White test gap. Washington: Brookings.
Klopfenstein, K. (2003, Fall). Recommendations for maintaining the quality of Advanced Placement programs. American Secondary Education, 32(1), 39-48.
Klopfenstein, K. & Thomas, M. K. (2006). The link between Advanced Placement experience and early college success. Unpublished manuscript.
Lichten, W. (2000). Whither Advanced Placement? Education Policy Analysis Archives, 8(29). Retrieved January 14, 2007 from http://epaa.asu.edu/epaa/v8n29.html.
Lichten, W. (2001, April). Where should we go with Advanced Placement? Paper presented at the meeting of the American Physical Society. Retrieved January 14, 2007 from http://aps.org/units/fed/newsletters/fall2001/lichten.cfm
National Research Council. (2002). Learning and understanding: Improving advanced study of mathematics and science in U.S. high schools. Committee on Programs for Advanced Study of Mathematics and Science in American High Schools. Washington: National Academy Press.
Osborne, R.T., & McGurk, F.C.J. (1982). The testing of Negro intelligence (Vol. 2). Athens, GA: The Foundation for Human Understanding.
Sadler, P. (2006). High school AP courses do not predict college success in science. Retrieved January 14, 2007 from http://www.fas.harvard.edu/home/news_and_events/releases/ap_02172006.html
Shuey, A. M. (1966). The testing of Negro intelligence. New York: Social Science Press
Wikipedia. (2006). The Sorcerers Apprentice. Retrieved January 14, 2007 from http://en.wikipedia.org/wiki/Sorcerer's_apprentice