becker: teaching economics to undergraduates teaching ... · becker: teaching economics to...

Journal of Economic LiteratureVol. XXXV (September 1997), pp. 1347–1373

Becker: Teaching Economics to Undergraduates

Teaching Economics toUndergraduates

WILLIAM E. BECKERIndiana University

Parts of this draft were prepared though the support of the Universities of Adelaide, Mel-bourne, and South Australia, provided while Becker was on leave from Indiana Universityduring the fall 1995. Constructive criticism on earlier drafts was provided by Robin Bartlett,William Baumol, Suzanne Becker, Stephen Buckles, George Bredon, Paul Grimes, RobertHighsmith, Hirschel Kasper, Peter Kennedy, Mark Maier, Joan Middendorf, Michael Salemi,Phillip Saunders, John Siegfried, Rosalind Springsteen, Myra Strober, and Michael Watts. T.Aldrich Finegan, Maier, and Siegfried contributed helpful data leads. Sue Becker and JulieMarker provided research assistance. Finally, special thanks are owed to W. Lee Hansen,William Walstad, and the anonymous referees who helped me to focus on the key issues.

UNFLATTERING CRITIQUES of col-lege teaching appear regularly in

books with titles such as Profscam, Im-postors in the Temple, and The Fall ofthe Ivory Tower. Martin Anderson’s(1992) Impostors in the Temple singlesout economists as particularly contemp-tuous of teaching. Yet, a notable numberof economists are concerned about theteaching of undergraduate economics,and as this article shows, contributegreatly to its development and evalu-ation. At the same time, many econo-mists are not keeping up with educa-tional changes in their discipline andtheir institutions.

Economics is now a discipline domi-nated by teaching jobs. Over the past 20years, approximately 70 percent of allnew jobs advertised in the AEA JobOpenings for Economists (JOE), havebeen in academe,1 with an increasing

proportion of those jobs at graduate de-gree granting institutions (Figure 1). Asthe higher education industry movedfrom one dominated by four-year liberalarts colleges to graduate degree-grant-ing institutions, all faculties to some de-gree may have lost sight of their obliga-tion to teach undergraduates. Ineconomics, however, the effect mayhave been more pronounced. Unlikeother social scientists, economists mayhave had the luxury to ignore under-graduates as enrollments in economicsrose in the 1980s. But now faced withdeclining numbers, academic econo-mists must give more consideration tohow they teach.

This article provides a summary ofthe evidence on teaching economics toundergraduates. It describes whateconomists are doing in classrooms, anddiscusses the consequences of their fail-

1347

1 The AEA initiated Job Openings for Econo-mists (JOE) in 1975. Although Elton Hinshaw(1975, p. 489) stated that participation was goodfrom the start, variations in the 1975 and 1976data may reflect a start-up phenomenon that in-

volves more than simple random sampling error.The 1977 through 1996 mean number of new jobsper year in academe is 1,192, while it is 489 fornonacademic jobs.

ure to do more. It reviews what the re-search on teaching economics has to of-fer classroom teachers and considers al-ternative measures of educationaloutputs. Extensive consideration isgiven to nonlecture teaching methodsthat are more prevalent in other disci-plines.

I. The Teaching Environment

The founders of the American Eco-nomic Association demonstrated an in-terest in the teaching of economicsprinciples but not all of its subsequentleaders have shared a commitment tothat mission. As colleges expanded intograduate education, economists lostsight of the importance of undergradu-ate courses and the way they are taught.In contrast with other disciplines thathave moved to a broad teaching reper-toire, economics continues to be taughtby the lecture method in all under-graduate courses.

A. Institutional Change

Between 1970 and 1994 the propor-tion of baccalaureate degree-grantingcolleges in the Carnegie Classification(1994) decreased from 53.4 percent to45.4 percent of the total number of bac-calaureate and advanced-degree-grant-ing institutions. The proportion of eachof the three Carnegie types of post-bac-calaureate institutions increased, withthe research category rising from 6.8percent to 8.9 percent of the total andthe proportions of doctorate and mas-ter’s degree granting institutions in-creasing as well (Table 1). Consistentwith these trends in higher education,but more dramatic, is the shift in thetype of institutions advertizing in JOE.In the last 20 years, the ratio of jobopenings in “four-year colleges” tothose classified as “universities withgraduate programs” plummeted fromabout 8 in 10, to 4 in 10 (Figure 1).Academic jobs in economics are dispro-

Rat

io o

f 4-y

r to

Gra

duat

e

Figure 1. Ratio of College to University Jobs in Economics

1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996

.94

.87

.80

.73

.66

.59

.52

.45

.00

1348 Journal of Economic Literature, Vol. XXXV (September 1997)

portionately opening in graduate de-gree-granting institutions.

The “upward” movement of a depart-ment and institution on the degree lad-der or a mere name change such asfrom the College of Saint Thomas tothe University of Saint Thomas, carrieswith it an implied shift from a liberalarts orientation to a graduate or profes-sional school orientation. Faculty mem-bers face real or perceived change to aresearch emphasis, even though theirmake-ups may have changed little,2 andthe bulk of their work continues to bethe teaching of undergraduates. Policymakers see this movement as a Malthu-sian cycle in which competition forfunds to build their research prestigeleads institutions to use their own re-sources to subsidize research (RogerGeiger and Irwin Feller 1995). Between1970 and 1994, for example, the propor-tion of research and development ex-

penditures financed by own-institutionresources increased from 10.41 percentto 18.21 percent while external stateand federal support fell (Table 2).Changes in accounting procedures overthis period preclude tracking internalfunds to establish the source of reallo-cation, but the dichotomization ofhigher education into “graduate educa-tion and research” versus “undergradu-ate education,”3 by Rothschild andWhite (1993) for example, suggests that

TABLE 1CLASSIFICATION OF POSTSECONDARY EDUCATIONAL INSTITUTIONS

(NUMBER OF INSTITUTIONS/PROPORTION OF SUBTOTAL)

1970 1976 1987 1994

Research 92 98 104 1256.81% 7.20% 7.54% 8.92%

Doctorate 81 86 109 1116.0% 6.32% 7.90% 7.92%

Masters’ 456 594 595 52933.78% 43.64% 43.12% 37.73%

BS/BA 721 583 572 63753.41% 42.84% 41.45% 45.44%

Subtotal 1,350 1,361 1,380 1,402Associate 1,063 1,146 1,367 1,471Total 2,413 2,507 2,747 2,873

Source: Carnegie Foundation for the Advancement of Teaching (Various Editions).

2 Ronald Ehrenberg, Kasper, and Daniel Rees(1991) found that the rate of retention of facultymembers between 1972 and 1989 averaged about85 percent for assistant professors, 93 percent forassociate professors, and 92 percent for professors.

3 Michael Rothschild and Lawrence White(1993) dismiss the idea that undergraduate educa-tion subsidizes graduate education and researchwith the argument that an industry with joint pro-duction at some institutions and single productionat others would not be sustainable; that we ob-serve this industry implies that there are no subsi-dies. Although the logic (if A, then B; thus, not Bimplies not A) is valid, Rothschild and White’spremise (A) is a compound event: if institutions ofhigher education faced the same regulations, andif they all produced the same undergraduate prod-uct, and if students and their parents had accurateinformation, and if firms were free to enter, . . .,and if undergraduate education subsidizes gradu-ate education, then undergraduate institutions andjoint graduate and undergraduate institutionswould not coexist. Existence of the different typesof producers implies only that at least one of themany premises is wrong.

Becker: Teaching Economics to Undergraduates 1349

undergraduate education was not on thereceiving end. In the quest for prestigeit seems unlikely that faculties in eco-nomics disregard assertions, such asthat found in White (1995), that there isan inverse relationship between theranking of graduate economics depart-ments and the amount of attentionthose departments give to the develop-ment and assessment of teaching.

Economics faculties at some liberalarts colleges now perceive a need todemonstrate they behave like facultieswith graduate programs, with an un-precedented two articles in the Journalof Economic Education providing rank-ings of these colleges by number andplacement of publications (HowardBodenhorn forthcoming; James Hartleyand Michael Robinson forthcoming). Itappears that younger liberal arts facultymembers in particular are attempting toestablish publication records like thoseat the better graduate schools, reflect-ing a change in emphasis within thesedepartments (Bodenhorn forthcoming).Although they are still called liberalarts, the mission of many of these col-leges has changed with the introduction

of business schools, law schools, andsmall graduate programs (David Brene-man 1994, p.12).

Finally, as traditional four-year col-leges established graduate programs,and faculties at the remaining liberalarts colleges tried to emulate graduatefaculties, the number of two-year com-munity colleges classified by theCarnegie Foundation expanded from1,063 institutions in 1970 to 1,471 in1994 (Table 1). But community collegesdo not embrace or are not embraced bythe majority of economists associatedwith the AEA, as indicated by theAEA’s elimination of this classificationfrom JOE in 1982. The growth in com-munity colleges implies a growth in thedemand for college teaching of econom-ics that is being ignored by the AEA.

B. Economists and Their Organizations

The first AEA standing Committeeon Economic Education was appointedin 1953. Its successors have been active,with sessions at the AEA annual meet-ings and papers appearing regularly inthe American Economic Review Pro-ceedings. To avoid direct AEA sponsor-

TABLE 2R&D EXPENDITURES AT UNIVERSITIES AND COLLEGES, BY SOURCE OF FUNDS: SELECTED FISCAL YEARS

(DOLLARS IN MILLIONS/PROPORTION OF TOTAL)

FiscalYear Total

FederalGovernment

State and Local

GovernmentPrivate

IndustryInstitutional

Self-FundingOther

Sources

1970 2,335 1,64770.54%

2199.38%

612.61%

24310.41%

1657.07%

1976 3,729 2,51267.36%

3649.76%

1233.30%

44611.96%

2857.64%

1987 12,153 7,34360.42%

1,0238.42%

7906.50%

2,16817.84%

8286.81%

1994 21,081 12,66160.06%

1,5627.41%

1,4306.78%

3,83818.21%

1,5907.54%

Source: National Science Foundation. Survey of Scientific and Engineering Expenditures at Universities andColleges. Fiscal Year 1994. Table B−1.


ship in the delivery of programs, theAEA Committee on Economic Educa-tion linked with the National (formerlyJoint) Council on Economic Education(a nonprofit organization receiving sup-port from all sectors of the economy).In addition to implementing pre-collegeprograms, in the 1960s the AEA com-mittee and NCEE developed a nation-ally normed test of micro and macro-economics principles: The Test ofUnderstanding of College Economics(TUCE), now in its third edition (Saun-ders 1991). The Journal of EconomicEducation was established to provide achannel of communication for researchwork and related publishable materialin economic education, primarily at theundergraduate level. In 1973, the AEAcommittee and NCEE initiated ateacher training program for graduatestudents and new instructors, with the1992–1994 offering involving 236 par-ticipants from 180 colleges and univer-sities (Salemi, Saunders, and Walstad1996). These on-going activities, as wellas timely but fixed-term initiatives, havebeen central to the program planning ofthe AEA Committee on Economic Edu-cation.

The various ventures of the AEA intoeducation have not always been metwith applause. Those opposed to AEAeducational activities point to the AEACertificate of Incorporation (1989) ob-jectives, which call for “the encourage-ment of economic research” but includeno reference to the encouragement ofeconomic education. George Stigler, asa past AEA president and member ofthe AEA Executive Committee, wasmost strident in promoting the viewthat teaching is akin to pontification,and thus contrary to “the encourage-ment of perfect freedom of economicdiscussion,” another AEA objective. Un-like the publishing of journals, pro-grams aimed at teaching were seen as

an endorsement of the ideas expressedby participants.

In contrast to the AEA, the AmericanHistory Association’s annual meetingprogram states “The Program Commit-tee has asked commentators in all ses-sions to address the implications of thepapers being given not only for researchbut also for teaching” (AHA 110th An-nual Meeting, Atlanta, January 4–7,1996, p. 18). A dedicated section in theprogram identified 27 sessions andevents that dealt with the teaching ofhistory, including a separate receptionfor two-year college faculty. On thequantitative side, the American Statis-tics Association has a long record ofconcern for teaching but yet PresidentLynne Billard (1996, p. 9), after de-scribing a report she had done on therole of statistical societies in education,encouraged more ASA involvement ineducation:

It is doubtful that many of us would disputethe need for an association to become in-volved in such activities though we might wellhave differing views as to how these can bebest effected . . . I draw our attention to twomore general issues, ones in which we as anassociation, could and should perhaps be-come involved, specifically, the quality of in-struction, and statistics as part of a core cur-riculum.

Scholars in other disciplines that ex-perienced enrollment declines in the1980s are doing likewise. Economistsare noticeably absent from groupsformed to advance college teaching.For instance, of more than 1,000 indi-viduals listed in the Directory of Coop-erative Learning Practitioners inHigher Education only a dozen or soare economists. Typically only three orfour economists are among the 200 at-tendees at the International Society forExploring Teaching Alternatives. Onlytwo of the 100 scholars involved in theprestigious Harvard Assessment Semi-


TABLE 3TEACHING METHODS, EXAMPLES, ASSIGNMENTS, AND CLASS SIZE

Median (and Mean) Responses by University Type

ResearchDoctorateGranting Masters

LiberalArts

AssociateInstitutions

IntroductoryLecture Time 83(78)% 83(68)% 83(73)% 83(72)% 83(73)%

Guest Lectures 0( 3)% 0( 3)% 0( 3)% 0( 2)% 0( 3)%

Class Time Use of Chalkboarda 83(60)% 83(65)% 83(68)% 83(57)% 83(65)%

Overheadb 6(28)% 0(17)% 6(25)% 6(17)% 22(30)%

Computer labs 0( 3)% 0( 5)% 0( 6)% 0( 4)% 0( 4)%

Coop learning 0( 8)% 6(15)% 6(16)% 0(11)% 6(19)%

Examples from Literature 0( 9)% 6( 8)% 6(11)% 6( 5)% 6(15)%

Sports 6(11)% 6(11)% 6(11)% 6(11)% 6(21)%

Assigned Materials Textbooks 83(78)% 83(75)% 83(77)% 83(74)% 83(73)%

Workbooks 6(36)% 6(31)% 22(35)% 22(32)% 22(37)%

Class notesc 6(30)% 6(33)% 6(30)% 6(26)% 22(37)%

Problem setsc 50(48)% 22(37)% 22(38)% 22(34)% 22(29)%

Popular press 22(25)% 22(33)% 6(28)% 22(24)% 22(30)%

Academic pubs. 0( 3)% 0( 3)% 0( 5)% 0( 3)% 0( 4)%

Class Size 100(162) 30(31) 40(45) 45(67) 30(30)

Numberd 100 96 120 77 60

TheoryLecture Time 83(74)% 83(66)% 83(73)% 83(74)% 83(83)%

Guest Lectures 0( 2)% 0( 3)% 0( 2)% 0( 2)% 6( 6)%


Overheadb 6(16)% 0(15)% 6(16)% 0( 7)% 6( 6)%

Computer labs 0( 4)% 0(12)% 0( 6)% 0( 4)% 0( 0)%

Coop learning 0(10)% 6(20)% 6(13)% 0( 9)% 22(22)%

Examples from Literature 0( 7)% 0( 6)% 6(11)% 6( 4)% 6( 6)%

Sports 6( 8)% 6( 8)% 6( 8)% 6( 9)% 22(22)%


Workbooks 6(25)% 6(27)% 6(22)% 6(24)% 0( 0)%

Class notesc 14(31)% 6(32)% 6(24)% 6(22)% 83(83)%

Problem setsc 50(52)% 22(37)% 22(39)% 22(39)% 6( 6)%

Popular press 6(25)% 22(33)% 6(24)% 6(18)% 50(50)%

Academic pubs. 6(12)% 6(16)% 6(12)% 0(10)% 6( 6)%

Class Size 40(50) 20(20) 25(24) 30(30) 5( 5)

Numberd 103 70 98 59 2

TABLE 3 (Cont.)TEACHING METHODS, EXAMPLES, ASSIGNMENTS, AND CLASS SIZE

Median (and Mean) Responses by University Type

ResearchDoctorateGranting Masters

LiberalArts

AssociateInstitutions

Statistics/EconometricsLecture Time 83(79)% 83(73)% 83(71)% 83(67)% 83(83)%

Guest Lectures 0( 0)% 0( 2)% 0( 1)% 0( 0)% 0( 2)%


Overheadb 6(23)% 0(10)% 6(20)% 0( 7)% 0( 2)%

Computer lab 22(34)% 22(29)% 22(34)% 22(30)% 6(11)%

Coop learning 0(15)% 6(23)% 6(22)% 0(10)% 22(35)%

Examples from Literature 0( 2)% 0( 3)% 0( 7)% 0( 5)% 6( 4)%

Sports 6( 7)% 6(10)% 6( 9)% 6( 9)% 6(11)%


Workbooks 0(16)% 0(18)% 0(15)% 0(22)% 50(44)%

Class notesc 6(31)% 6(33)% 22(34)% 22(33)% 83(83)%

Problem setsc 83(64)% 50(46)% 50(51)% 50(50)% 6(18)%

Popular press 0( 9)% 0(11)% 0(10)% 0( 9)% 0( 7)%

Academic pubs. 0(19)% 0( 9)% 0( 7)% 3(14)% 0( 0)%

Class Size 30(37) 19(20) 25(25) 22(23) 25(22)

Numberd 52 35 61 37 4

Upper DivisionLecture Time 83(74)% 50(61)% 83(68)% 83(68)% 83(83)%

Guest Lectures 0( 4)% 6( 6)% 6( 6)% 6( 5)% 3( 3)%


Overheadb 6(19)% 0(10)% 6(17)% 0(11)% 3( 3)%

Computer labs 0( 7)% 0( 8)% 0( 8)% 0( 7)% 0( 0)%

Coop learning 0(11)% 6(20)% 6(18)% 6(19)% 25(25)%

Examples from Literature 0( 7)% 6( 6)% 6( 9)% 6( 7)% 28(28)%

Sports 6( 7)% 6( 8)% 6( 8)% 6( 8)% 36(36)%


Workbooks 0(16)% 0(10)% 0(20)% 0(11)% 42(42)%

Class notesc 6(31)% 6(30)% 6(26)% 6(23)% 42(42)%

Problem setsc 50(43)% 6(28)% 22(33)% 22(33)% 3( 3)%

Popular press 22(28)% 22(34)% 22(29)% 6(26)% 53(53)%

Academic pubs. 22(41)% 22(33)% 6(25)% 22(31)% 3( 3)%

Class Size 30(35) 15(18) 20(20) 20(26) 15(14)

Numberd 159 87 108 85 4

a. Written during class only b. Prepared acetates onlyc. Instructor prepared d. Number providing class size information

nars taught economics (both in Har-vard’s Kennedy School of Government;neither were faculty members in theDepartment of Economics, and onlyone had a Ph.D. in economics). Schol-ars from the arts, education, humani-ties, other social sciences, and naturaland earth sciences from many institu-tions participated in the Harvard As-sessment Seminars.4 As much of therest of higher education implementsnew approaches to teaching, traditionaleconomists may be stuck in the rut ofdoing to undergraduates what their in-structors did to them. In response stu-dents may be voting with their feetwhen they abandon economics.5

C. Teaching Practices

The manner in which economics hasbeen and continues to be taught to un-dergraduates is documented in nationalsurveys (Becker and Watts 1996; Sieg-fried et al. 1996; Cynthia Benzing andPaul Christ 1997). The 625 respondentsto the Becker and Watts survey are typi-cal, showing the representative U.S. un-dergraduate economics teacher to be amale (83 percent) Caucasian (89 per-cent), with a Ph.D. degree (86 percent).At research universities the teachingload averages two courses per semester,whereas it is three courses per semesterat doctoral, master’s, and liberal artsinstitutions, and five courses per semes-

ter at associate degree-granting institu-tions.

Becker and Watts’ survey of teachingmethods in four types of undergraduatecourses (introductory, intermediate the-ory, statistics and econometrics, and up-per-division field courses), for fiveabridged Carnegie classifications is pre-sented in Table 3. The picture of aneconomist lecturing to a class, while hewrites on the chalkboard and assignsreading from a textbook, appears accu-rate for all courses and all institutions,with only a minor caveat for upper-divi-sion courses at doctorate institutions,where the median instructor spends lessthan 83 percent of class time lecturing.It is also only in these upper-divisioncourses at doctorate, master’s, and lib-eral arts institutions that guest lecturesare used to any degree. Stigler’s view ofan academic economist as sole preacheris supported by survey data.

Cooperative learning techniques inwhich students work together in theclassroom are noticeably absent in alleconomics courses at research universi-ties where the largest classes tend to beemployed. Curiously, while much iswritten on the use of computers in theteaching of economics, and conferencesessions devoted to education technol-ogy are well attended, computer labsare used notably only in the teaching ofstatistics and econometrics. Whetherthis lack of newer instructional methodsis primarily because of low instructordemand or inadequate facilities sup-plied by the institutions is unknown.Regardless of the reason, however, it isconsistent with a passive learning envi-ronment that does not engage students.In contrast, class discussion and otherforms of active learning, and not exten-sive lecturing, are now the most promi-nent forms of instruction used acrossthe rest of higher education (Linda Saxet al. 1996, p. 13).

4 According to Siegfried et al. (1991, p. 207),economists give scant attention to placing theirdiscipline within the broader liberal arts curricu-lum. Grimes and James Niss (1991) state that busi-ness faculty want to reduce required economicscourses because the courses are not integratedwith the business curriculum. Then for whom areeconomics courses offered and to what curriculumare they tied?

5 If resources are not tied to enrollments to somedegree, there is no incentive for an academic unit to careabout students. Even if resources are tied to enroll-ments, however, there is a principal-agent problem. Afaculty member can decrease his or her work load byweeding students out of a course. This principal-agentproblem is addressed by monitoring withdrawals.


II. Number of Majors in Economics

National data on degrees granted areavailable from the government but indi-vidual course enrollments and numbersof majors are not. The proportion ofbaccalaureate degrees awarded in eco-nomics fell in the 1950s, from approxi-mately 3.5 percent to 1.8 percent (asread from the left-hand vertical axis,Figure 2). This proportion rose in the1960s only to fall again in the early1970s, hitting a low of 1.5 percent of alldegrees granted in 1975. It resumed itsupward climb reaching 2.4 percent inthe late 1980s but has been fallingsince. Robert Margo and Siegfried(1996) note that after excluding thepre-1955 period, there is no trend inthe relative number of economics de-grees awarded. They speculate that self-equilibrating mechanisms induce themean share of economics degrees to asteady state 2.2 percent of all under-

graduate degrees awarded. This allegedstability, however, may be the conse-quence of a floor effect. As the ratio ofeconomics to total bachelor’s degreesapproaches zero or some unknown smallvalue, it must turn up or lead to re-trenchments and closings of depart-ments of economics.

A pure noise explanation of the inter-temporal behavior of economics majorsaside, an explanation of the recenttrend in economics majors is often tiedto the number of degrees awarded inbusiness. In the late 1980s, when theproportion of economics degrees was ata relative high, the proportion of busi-ness degrees was also at a high of about25 percent of all bachelor’s degreesgranted but the proportion of bachelor’sdegrees awarded in the other social sci-ences was at a low of 7 percent (as readfrom the right-hand axis in Figure 2).

The perceived association of thenumber of degrees awarded in business

Eco

nom

ics

1950

3.5

3

2.5

2

1.5

1

0.5

0

Bus

ines

s an

d O

ther

Soc

ial S

cien

ces

28

24

20

16

12

8

4

0

Business

Economics

Other Soc. Sci.

Figure 2. Percentage of All Bachelor’s Degrees Awarded by Field, 1950-1994

Source: Willis and Pieper (1996) and National Center for Education Statistics (1995 and 1996).

1955 1960 1965 1970 1975 1980 1985 1990 1995


and the other social sciences is now thefocus of attention. Detailed micro datafrom the University of North Carolinasupports the idea that much of theswing in economics majors since 1980reflects an increase and subsequent de-cline in the popularity of undergraduatebusiness studies (Salemi and CarlieEubanks 1996).6 When demand forbusiness degrees increased in the 1970sand early 1980s and business schoolsraised grade point and course require-ments, many excluded students electedeconomics as a second best major. Thedecline in student demand for businessin the 1990s brought a reduction inbusiness school entrance requirementsand a decrease in those electing eco-nomics as a major. Although the othersocial sciences gained students, no onediscipline stands out as a particular re-cipient (Rachel Willis and Paul Pieper1996).

To some extent choice of major andeven elective courses are based on ratesof return. Unfortunately, time series donot exist on lifetime earning prospectsassociated with alternative courses of

study. We do have a snapshot of earn-ings of 215,000 persons who reportedhaving a bachelor’s or more advanceddegree in the 1990 decennial census(Daniel Hecker 1995, p. 4). In the 25–34 year age group, for example, menholding bachelor’s degrees in econom-ics have median earnings 3 percenthigher than the overall average; by ages35–44 it is 14 percent higher. Forwomen in the 25–34 year age group,holders of economics bachelor’s degreesearn 13 percent more than the average;by ages 35–44 they earn 53 percentmore than the average. The medianearnings of both men and women eco-nomics bachelor’s degree holders arehigher than bachelor’s degree holdersin general business and the other socialsciences. Unless there were dramaticchanges in the distribution of salariesover time, for which data are unavail-able, the relative high return to an eco-nomics major cannot explain the down-turn in the number majoring ineconomics. Furthermore, other thanmaking economics majors more attrac-tive to prospective employers little canbe done about external labor marketforces that determine compensationpackages.

III. Changing the Teaching/LearningEnvironment to Attract Students

The influence of teaching skills onenrollments in economics is dismissedwith the observation that faculty turn-over is slow (Willis and Pieper 1996):those teaching in 1994 are much thesame as those who taught in 1990. Thisargument, however, assumes that fac-ulty members do not change their waysin accordance with incentives. Yet,when confronted with falling enroll-ments and declining resources, facul-ties do consider assigning highergrades, expecting less rigor, and provid-

6 Salemi and Eubanks (1996) do not differenti-ate among economics degrees offered by differentunits in universities. Siegfried et al. (1991, p. 198)state that “administratively departments of eco-nomics are divided among colleges of arts and sci-ences (65 percent), schools of business (30 per-cent), and a few other administrative units.” Thenumber of bachelor degrees awarded in businesseconomics is small, hitting a high of 3,779 degreesin the 1991/92 school year, and then falling to3,454 degrees in 1993/94, roughly its 1987/88level. If there is little difference in what is re-quired of students to earn a degree in economicsand business economics (Siegfried et al. 1991),then what students do for a major cannot be anexplanation of the difference in trends. I devoteno more time to differences in administrativeunits, although an interesting line of inquiry wouldbe an exploration of effects of having multipleunits within the same university offering compet-ing degrees in economics. For example, the workof David Brasfield et al. (1996) suggests that eco-nomics departments that do not face competitionfrom a business program at the same institutionare at more risk for losing majors.


ing more options to students. Bettermeans of instruction should also be con-sidered.

A. Grades

Economics is among the lower-grad-ing departments; thus, economists haveroom to inflate grades to buy students.But will higher grades attract studentsto economics? The answer is a qualifiedyes. Simulations show that startingfrom the position of a relatively lower-grading department, inflating gradesin the introductory course to resemblethose in higher-grading departmentswill lead to students taking one ormore courses beyond the introduc-tory course, provided other depart-ments do not respond in kind (Rich-ard Sabot and John Wakeman-Linn1991).

Grades seem particularly importantto the persistence of female students(Gordon Anderson, Dwayne Benjamin,and Melvyn Fuss 1994; Karen Dynanand Cecilia Rouse forthcoming). JaneHorvath, Barbara Beaudin, and SheilaWright (1992) estimate that the prob-ability of persisting from the first intro-ductory course to the second is 16 per-cent higher for male students with a Bgrade than for women with a B grade.This result is consistent with conven-tional wisdom within education that fe-male students are more self-criticalthan their male counterparts and thatwomen’s satisfaction with the academicexperience is more closely tied togrades than it is for men.7 When com-pared to men, women require strongercues and more reassurance to persist inthe study of economics.

Attempting to attract and retain stu-dents with grades is an individual in-structor and institutional decision thatgoes beyond the scope of this paper. Itis worth noting, however, that gradedistributions can be altered without in-creasing mean measured grade infla-tion—for example, increasing the pro-portion of Bs, by assigning this highergrade to those who previously wouldhave been given a B− or C+, and at thesame time increasing the proportion ofDs, by assigning lower grades to thosewho previously would have been given aC− or D+, may encourage those wewant to continue while discouraging theothers with little or no change in theaverage grade.

B. Requirements

An often heard assertion is that in-creasing the mathematics requirementin economics produced a dramaticchange in the mix of students takingcourses and majoring in economics(Kasper et al. 1991; Kasper 1996). Ifthere was a move to more mathematicalrigor in economics relative to the re-quirements in other related disciplines,then the number of students in econom-ics may have fallen with the increasedmathematics requirements.

Information available from the Col-lege Board annual surveys of the admis-sion criteria of colleges and universitiesshows that mathematics requirementshave been increasing generally. From1986 to 1992, for example, the percent-age of the 1,562 post-secondary institu-tions that had high school mathematicsrequirements or recommendations foradmission increased from 69 percent to83 percent, and the average number ofhigh school units of mathematics re-quired or recommended rose from 2.6to 2.7. Furthermore, student mathemat-ics skills, as measured by either themean ACT or SAT, were unchanged in

7 Richard Light (1990, p. 20) acknowledges thatfor women satisfaction with a specific course ismore closely tied to grades than it is for men, butthen reports that women’s overall satisfaction withthe entire college experience is less tied to gradesthan is the case for men.


the late 1980s and early 1990s whenthe number of economics majors roseand then fell. A steep decrease in meanACT math score occurred between 1970and 1983, when it fell from a highof 20 in 1967 to a low of 16.9 in1983 (National Center for EducationalStatistics Digest of Education Statistics,1995, Table 130). A similar plunge is evi-dent in SAT scores between 1970 and1984. Yet, between 1970 and 1983 theshare of economics majors fell and thenrose (Figure 2).

Although students may be able tohandle the mathematics of economics,it may be that they do not see its rele-vance. The liberal arts education is tobe a broadening experience but the in-corporation of mathematics into eco-nomics tends to be a narrowing experi-ence that focuses on the technical skillsof the theorist (Bradley Bateman1992).8 On the other hand, some stu-dents may be attracted to economics asa place to apply their mathematicsskills. For example, minors and electivecourses in economics have been createdand are succeeding in appealing tothe mathematical aptitudes of engineer-ing and science students, bringingsignificant increases in both theoryand field course enrollments and nodecrease in the number of traditionaleconomics majors (Brasfield et al. 1996;Gregory Gelles and Walter Johnson1994). The effect of changes in thelevel of mathematical rigor expectedof economics students may be schoolspecific. On balance, however, incre-mental changes in requirements cannotbe expected to produce dramaticchanges in majors or course enrollments.

C. Class Size

Research universities appear to fi-nance their relatively low teaching loads(averaging two courses per semester)through larger class sizes (Table 3). Re-spondents from research universitiesteach larger classes across the fourtypes of courses, with mean class size inintroductory courses of 162 students,compared with average class sizes of 45and 67 in master’s and liberal arts insti-tutions and 30 at both doctoral and as-sociate degree-granting institutions.Siegfried et al. (1996, p. 189), on theother hand, report “Average class sizeacross both macro and micro is aboutthirty in two-year colleges, thirty-five inliberal arts colleges, forty-five in com-prehensive (master’s) universities, andaround sixty in research and doctoral in-stitutions.” Either respondents to theBecker and Watts survey exaggerateclass sizes, or the Siegfried et al. dataset does not include many large classesor misses many students in largeclasses. The latter conclusion seemslikely because the Siegfried et al. dataset, assembled as part of the norming ofthe TUCE III, required volunteeringinstructors to incur higher costs of ad-ministering and managing the term-longdata collection process in largerclasses.9 The cost of completing the

8 Becker (1990) and Strober, Allen Cook, andKasi Fuller (forthcoming) argue that the symboliclogic used in economics may hinder some studentswith good backgrounds in high school mathemat-ics because the conventions used in algebra andgeometry appear contrary to those in economics.

9 Class sizes in Siegfried et al. (1996) were cal-culated as the mean number of pretests and post-tests obtained in the norming of TUCE III. Re-sults were stratified by consolidating the Carnegieresearch and doctoral universities into one group,with the remaining three groups (master’s, liberalarts, and associate) kept separate. Becker andWatts used the less aggregated Carnegie catego-ries of “research” and “doctoral” level institutions.But combining the research and doctoral respon-dents of the Becker and Watts sample gives amean introductory class size of 98 students, whichis still above the 60 in the “research and doctor-ate” category in the TUCE III data. Even at theliberal arts institutions, the Becker and Wattsmean introductory class size of 67 far exceeds themean class size of 35 students for four-year col-leges in the TUCE III data set.


Becker and Watts survey was the samefor teachers of large and small classes.Thus, their finding of large class sizes atboth research universities and liberalarts colleges is credible.

Class size is important because it af-fords instructors the opportunity to trydifferent teaching methods. As WilbertMcKeachie (1990, p. 190) stated in a re-view of the class-size literature,

It seems plausible that the effect of class sizeon learning depends on what the teacherdoes . . . in larger classes, faculty memberstypically require less written work and spendmore time lecturing and less in discussion . . .Lecture tends to be at least equal to, andoften more effective than, discussion for im-mediate recall of factual knowledge on acourse examination, but discussion tends tobe superior for long-term retention.

Although Benzing and Christ (1997) re-port that a large proportion of econo-mists responding to their survey believestudents learn more in small classes,economists show no tendencies to selectteaching strategies based on course typeor class size. If we expect students tolearn more in one type of class versusanother, we have to vary instructionalmethods to fit the class.

D. Instruction

The “Report of the Commission onGraduate Education in Economics”calls for bringing “real-world issues intothe classroom” (Anne Krueger et al.1991). Yet, the commission’s assumedinstructional mode is lecture. It con-cludes that the lack of emphasis on ex-position skills is at least in part “a judg-ment that the appropriate style ofprofessional communication is some-thing (graduate) students can figure outfor themselves by watching their teach-ers” (p. 1049). The report of a commit-tee of economics faculty members fromprestigious liberal arts colleges in theUnited States (Kasper et al. 1991) em-

phasizes the need for graduate studentsto have breadth in content coverage tobecome good teachers of undergradu-ates but gave no attention to the needfor breadth and training in teachingmethods.

In contrast to these reports of econo-mists, education psychologists, and in-struction specialists appear united onone important influence on student per-formance and desire to continue a sub-ject’s study: the need for active studentinvolvement with classmates in thelearning process (David Johnson, RogerJohnson, and Karl Smith 1991; Alexan-der Astin 1993). Although evidencefrom pre-posttest analyses is lacking,for reasons to be discussed, survey re-sults indicate the importance of aca-demic support groups especially when itcomes to persistence.10

(1) Learning with Group Activities.Both alumni and senior class respon-dents to the Harvard Assessment Semi-nars survey said the key to academicsuccess is closely linking the academicwork to another person or group (Light1992). Respondents had two recommen-dations:

First, and unanimously recom-mended, freshman should take smallclasses that require in-class interaction.The student-teacher ratio is the mostimportant constraint in the teachingand learning process because two-waycommunication becomes increasinglydifficult as class size increases (SherwinRosen 1987). In small classes, each stu-dent has an opportunity to get to knowand communicate with the teacher.Equally important, according to educa-tion psychologists, is for each student tohave an opportunity to get to know sev-

10 When negative evidence on the effect of co-operative learning strategies is found, as seen inAstin (1993, p. 197), the findings are dismissed aspossibly the results of poorly designed group proj-ects and are designated as issues for future study.


eral others who share at least one thingin common: the course.

Second, the majority of respondentsto the Harvard surveys recommendedstudents form small study groups. Theyidentified the need for both in- andout-of-class study groups, with activitiesthat permit students to work togetheron structured learning tasks—what isknown as cooperative learning. Out-of-class study group activities can takemany forms. For large enrollment intro-ductory classes, students can be dividedinto groups of five or six, with eachgroup responsible for finding its owntime and place to meet. Some instruc-tors even arrange to meet or have anassistant meet with these groups of indi-viduals, although resource constraintsmake this option difficult to implementexcept for small classes.11 For in-classgroups in larger classes, instructors canuse the “think, pair, share” activity inwhich students are required to thinkabout a question posed by the instruc-tor; compare, discuss, and refine theiranswers with neighboring students; andshare their team’s responses with theclass if asked to by the instructor. Thethink, pair, share activity requires onlya few minutes, but provides a break instandard lectures by involving studentsand it allows instructors to assessquickly, and on a regular basis, whatstudents are learning. Diane Keenan

and Maier (1995) describe many coop-erative learning activities that can beused in the teaching of economics.

Technology can complement tradi-tional classroom activities. But unlesscomputers and other high-tech systemscan replace the instructor and supportpersonnel their introduction alwaysadds to the cost of instruction. Replac-ing the instructor with technology thatadapts to the needs of individual learn-ers is a dream that no instructional de-veloper has come close to achieving.Light (1990, p. 9), however, claims thatstudents who work in small groups,even when interacting with high-techequipment, learn more than studentswho work alone. Research shows nogreat breakthrough in the use of tech-nology in the educational process.Physical capital may be a poor substi-tute for human capital in education.12

There is a free-rider problem inher-ent in cooperative learning that Bartlett(1995) solves with an innovative gradingmethod that works in smaller classeswhere students get to know each other.Students are given the option of form-ing groups or not. Bartlett uses a ran-dom-number process to choose onemember of the group to take the examor make the presentation, and the gradeearned by the student is assigned to allmembers of the group. This creates in-centives for all members of the group tobe sure that everyone is well preparedbecause no one knows in advance whowill represent the group. KathleenMcKinney and Mary Graham-Buxton

11 National survey results show that in general“group-oriented teaching methods are most likelyto be used by younger faculty” (Sax et al. 1996, p.13). Whether older professors or younger graduatestudent assistants do better at encouraging smallgroup formation in economics classes is unknown.Early work in economic education suggests thatthere is no difference in their ability to producestudent learning. Watts and Gerald Lynch (1989),however, tell us that undergraduate students ofnonnative-English speaking graduate student in-structors do worse on final exams than students ofinstructors whose first language is English. Unfor-tunately, all of this work with exam scores suffersfrom the modeling, estimation, and testing prob-lems discussed in this article.

12 Rosen (1987) observed that no major innova-tion in education has occurred since the printingpress made inexpensive book publication possible.That higher education is a major contributor totechnological advances elsewhere in the economybut yet cannot capitalize on this technology inteaching is ironic. The Internet and software suchas WinEcon may lead to changes in educationalpractices but it is too early to say anything aboutthe effectiveness and efficiency of these computertools in teaching economics.


(1993) describe an alternative solutionto the free-rider problem in a large en-rollment introductory sociology course.Each student is required to complete awork-sheet demonstrating preparationbefore being permitted to participate inthe group activity. Students must par-ticipate to get credit toward the coursegrade.

(2) Learning with Individual Activi-ties. Cooperative classroom learning ex-periences can require large blocks oftime and not all students like to get en-gaged in group activities. To transmitfactual information and demonstratemathematical and graphical construc-tions, the lecture and blackboard maybe the desired delivery mode. Coopera-tive learning advocates (such as D.Johnson, R. Johnson, and K. Smith1991, pp. 38–40) seem to ignore that aslong as decisions outside the classroomare made on an individually competitivebasis, students may choose to performindividually in the learning environmentas well. Instructors in economics cannotrely on cooperative learning alone; theyneed to be able to employ alternativeteaching strategies—strategies that cankeep students actively involved, withboth practice and feedback. In this re-gard, the conclusion of Siegfried andRendigs Fels (1979) still holds: “Differ-ent students learn economics in differ-ent ways. The best teaching strategyprovides alternative learning methods”(p. 953). Many of these alternativemethods are described in Becker andWatts (1995).

Based on observations of one hun-dred chemistry lectures, Alex Johnstoneand Frederick Percival (1976) reportthat students have a noticeable behaviorchange (a lapse in attention) about 10to 18 minutes into a lecture, with lapsesbecoming more frequent as time passes.Thus they recommend a varied ap-proach, periodically involving students

in an activity closely tied to the lecturematerial. After discussing a theoreticalpoint, for example, an instructor can askstudents to construct a corollary, thinkof applications, or offer an examplequestion.

Another method that forces studentsinto action, and also provides continu-ous daily feedback, is the “minute pa-per.” Typically assigned in the last min-utes of each class when maintainingstudent attention is difficult,13 each stu-dent is required to respond to two ques-tions: (1) What was the most importantthing you learned in class today? (2)What question is unanswered? Althoughanswers may or may not be graded,these written responses can be used tomonitor comprehension as well as at-tendance and focus the discussion inthe following class period. The one-minute paper provides an incentive forattendance and participation, in con-trast with a passive attendance man-date.

Variations on the minute paper arenumerous and can lead to longer writ-ing assignments. The use of computersto provide quick feedback in large in-troductory economics courses datesback to the 1960s. Computerized ex-perimental economics projects andcomputer simulations are also success-ful in getting large numbers of studentsinvolved on an individualized basis withlittle instructional time required afterthe initial investment (Arlington Wil-liams and James Walker 1993).

Contrary to the notion that studentsprefer courses where they can work intheir own way and at their own pacewith relatively few quizzes, short pa-

13 The idea of monitoring the process of educa-tion from the start of a course to its termination,and monitoring students from the day they enterto the day they graduate, fits well with the indus-trial model of total quality management. JohnChizmar (1994) and Hansen (1993) discuss andapply TQM in the teaching of economics.


pers, or other means of feedback, “thelarge majority of students are sure theylearn significantly more when coursesare highly structured, with relativelymany quizzes and assignments to handin” (Light 1990, p. 5), even if studentsdislike such frequent evaluation at thetime (p. 32). Students stress that regu-lar evaluation during a well-organizedcourse enables them to plan their workmore than a few days in advance. Theyalso like quick turnaround from the in-structor, and prefer structured assess-ment routines in which they have theability to revise and make changes be-fore receiving a final grade.

At all types of institutions and in alltypes of courses, instructors prepareclass notes for students (Table 3). In-structor-supplied graphs may be benefi-cial for students who do not know howto draw graphs, but students who drawgraphs accurately on their own performbetter on the end-of-period economicsexam than those who do not. ElchananCohn and Sharon Cohn (1994) specu-late that those better at drawing who dotheir own graphs have an advantage be-cause they listen, draw graphs, and re-view their own work, whereas the oth-ers only listen and review theinstructor’s work. Periodically collect-ing and reviewing students’ classroomnotebooks is an assessment device thatencourages students to take this class-room activity seriously and provides in-formation on what students are attend-ing to in class.

Note taking and graph drawing arestepping stones to activities that requirestudent involvement. If nothing else,they force students to attend class, as-suming class notes are not availableelsewhere. Instructor-supplied classnotes and handouts are treated by stu-dents as substitutes for attendance (E.Cohn, S. Cohn, and James Bradley1995). Contrary to earlier studies, re-

cent studies (Garey Durden and LarryEllis 1995; David Romer 1993; KangPark and Peter Kerr 1990) show thatclass attendance does matter for variousforms of academic achievement that gobeyond what a student can cram forright before the final exam. Attendancemay not matter when achievement ismeasured by multiple-choice questions,but it does matter when written com-munication skills are assessed. If classesinvolve only lectures on textbook-typecontent, then end-of-course multiple-choice tests can be mastered by stu-dents who do not attend classes. If stu-dents are actively engaged indiscussion, in projects, and have peri-odic assessment, then attendance andparticipation are critical.

E. Educating Economists

There is only one national programdesigned specifically to help economistsmove from the lecture mode to alterna-tive teaching methods: the TeacherTraining Programs periodically con-ducted by the AEA Committee on Eco-nomic Education. With only a few hun-dred economists participating inoccasional program offerings, however,it is going to take a long time to ac-quaint a noticeable fraction of the AEAmembers with alternative teachingmethods via this program. Even for in-dividual graduate departments runningtheir own programs as offshoots, the at-titude appears to be that there are only“some incidental benefits for the gradu-ate instructors’ current (undergraduate)students” (White 1995, p. 83).

How the graduate student teacher’scurrent students could be viewed asgetting only “incidental benefits” is pe-culiar, and may simply be more evi-dence that undergraduate education ineconomics is unsatisfactory because ofthe poor training and the wrong signalsgiven to graduate students. There is no


credible test score evidence in the eco-nomic education literature on the opti-mal way to teach graduate students andfaculty members to be teachers. Thebenefit of programs aimed at an instruc-tor’s current students, however, seemspreferable to programs aimed at futurestudents. Furthermore, there is likelyno better way to document futureteaching potential than to provide evi-dence of current teaching practices andstudent outcomes.

IV. Measuring the Outcome of theTeaching and Learning Process

Although the above section calls forthe use of varied and active teachingand learning methods, specific methodsbetter than lecturing and writing on thechalkboard for teaching economics havenot been identified with traditional testscore comparisons between control andexperimental groups. This is not to saythat the lecture method is the best wayto teach economics. Quite the contrary,the problem is with the tests and otherinstruments used for the assessment ofteaching/learning experiments.

A. Test Scores

Fels set the stage for quantitativeanalysis of teaching techniques with his1969 presidential address to the South-ern Economic Association, “Hard Re-search on a Soft Subject: HypothesisTesting in Economic Education.” Heexpressed a need for a production func-tion analysis of teaching techniques,where a test score or change in testscores is related to student, instructor,and environment or institutional vari-ables (Siegfried and Fels 1979).

Much of the published research onteaching college economics uses perfor-mance on the Test of Understanding ofCollege Economics as the outcome of acourse of study. Throughout its three

editions, the TUCE has been a two-part, multiple-choice test of micro andmacroeconomics. It is intended to mea-sure more than recall, with over two-thirds of the 33 questions on each ofthe micro and macro-tests designed toassess student aptitude in applying eco-nomics to solving problems as found inintroductory two-semester micro andmacroeconomics courses of study(Saunders 1991).

In a common production functionstudy, the TUCE is used to test eachstudent’s knowledge of economics atthe beginning and end of a course. Achange score for each student is calcu-lated as the difference between his orher pretest (beginning) and posttest(end) scores. The TUCE scores, or thechange scores, are assumed to be pro-duced by human capital inputs (SAT orACT scores, initial knowledge of eco-nomics or some other subject, gradepoints, previous courses of study), utili-zation measures (time spent by studentor teacher in given activities), and tech-nology, environment, or mode of deliv-ery (lectures, group work, computeruse). Of all the variations considered byresearchers since 1968, the only consis-tently significant and meaningful ex-planatory variables of post-TUCEscores are pre-aptitude measures suchas pre-TUCE and SAT/ACT scores. Thepolicy implications could not be moreclear: to produce students who arehighly knowledgeable in economics,start with those who already have a highaptitude.14

14 Given the importance of pre-course aptitudemeasures, and the need to tailor instruction to theindividual student, it is curious that faculty mem-bers at many colleges and universities have al-lowed registrars to block their access to studentrecords for instructional purposes. As Nan Max-well and Jane Lopus (1994) report, students areless than accurate in providing information abouttheir backgrounds. Thus, using student self-re-ported data in regressions will always involve


This input-output approach has fourproblems. First, production functionsare only one part of a student’s decisionmaking system. Observed productioninputs are not exogenous but are deter-mined within this system. Second, dataloss and the resulting prospect for sam-ple selection bias in the standard pre-posttest design are substantial, with 20to 40 percent of those who took the pre-test no longer enrolled at the time ofthe posttest. Third, from probabilityand statistical theory, we know that fail-ure to reject the null hypothesis doesnot imply its acceptance. That an ex-perimental teaching method shows nostatistically significant improvementover the lecture does not imply that it isnot better. Finally, and possibly mostimportant, education is a multi-productoutput that cannot be reflected in a sin-gle multiple-choice test score. Theseproblems with the production functionmind set are being addressed by eco-nomic education researchers.

(1) Student Decision Making Frame-work. There are models that providemicroeconomic theoretical rationalesfor why researchers fail to find consis-tent evidence of the superiority of oneteaching technique over another in theproduction of learning in economics.Becker (1982) constructs a model inwhich a student maximizes the utility ofdifferent forms of knowledge, currentconsumption, and expected future in-come. This utility maximization is sub-ject to a time constraint, the productionrelationships that enable the student toacquire knowledge and consumption,and the manner in which the differentforms of knowledge are measured andenter into future income. The shadow

prices of knowledge reflect opportunitycosts generated by the time constraint,production functions, and uncertain fu-ture income.

The Becker model shows that im-proved teaching technology that enablesstudents to more efficiently convertstudy time in economics classes intoeconomic knowledge need not result inany change in student desire for moreeconomics knowledge.15 The time sav-ings from the more efficient pedagogyin the economics course may be in-vested by the student in the acquisitionof knowledge in other subjects or maybe used for market work or leisure. Theshadow prices of the different forms ofknowledge in equilibrium with the mar-ginal utility of each form of knowledge,leisure, and future income determinestudent choices. It is not just a produc-tion function relationship that gives riseto a certain mix of inputs being com-bined to produce a given output. Ob-served student study times, learning inother subjects, student opinions on sat-isfaction with a given course of study,and similar student input measures thatare used to explain knowledge andlearning of economics are, in fact, en-dogenous.

The model makes explicit how re-wards for knowledge in economics rela-tive to other subject knowledge enter astudent’s decision making. It also callsattention to the accuracy with whichstudent knowledge is measured and thebehavioral implications of the student’sposition in the distribution of thatknowledge. Researchers working withthe TUCE III data find, for example,that students do not take the test seri-

problems of errors in variables. Salemi andGeorge Tauchen (1987) discuss other forms of er-rors in variables problems encountered in the esti-mation of standard single-equation learning mod-els.

15 The word “knowledge” is used here to repre-sent a stock measure of student achievement; itcan be replaced with any educational outcomeproduced by the student with various forms ofstudy time and technology, as measured at a singlepoint in time.


ously when it does not count in thecourse grade. Yet, many educators con-tinue to overlook the effect of incen-tives on measured student performance.

In assessing the influence of incen-tives, it must be recognized that themarket for new graduates need notplace a value on student learning in thesame way it values the final attainmentof knowledge. The unquestioned use ofvalue-added (change score) models ineconomic education research ignoresthe fact that students, teachers, and em-ployers do not place the same value onpre- and post-course knowledge. Thebeliefs of an instructor, a test designcommittee, or an entire faculty aboutthe importance of certain forms ofknowledge and intellectual skills are notalways consistent with what studentsdesire, and what employers desire andpay for the bundle of knowledge andskills embodied in the graduate whostudies economics versus some othersubject.

(2) Data Loss and Sample Selection.A sizable proportion of students whoenroll in economics courses sub-sequently withdraw, never completingthe end of course evaluation or finalexam. A classic study of the differencebetween pretest and posttest scores toassess “value added” in these coursesexcludes all who dropped. The processthat determines which students quit be-tween the pretest and the posttest islikely related to the process that deter-mines test scores. Becker, John Powers,and Saunders (1996) provide probitmodel estimates showing, all else equal,that individual students with higherpre-TUCE scores are more prone topersist with the course than those withlower scores.

The timing of withdrawal is related tomany of the same variables that deter-mine test scores (G. Anderson, Ben-jamin, and Fuss 1994). For example,

taking high school calculus and econom-ics contributed greatly to a student’s de-sire to complete the entire two-semes-ter college economics course. However,more experienced students were morelikely to drop sooner; they did notstick around if they saw “the handwrit-ing on the wall.” Consistent with theseresults, Stratford Douglas and JosephSulock (1995) conclude that prior expe-rience with economics, accounting, andmathematics, as well as class atten-dance, all increase the probability ofa student completing an economicscourse. They also show how correctionfor self-selection out of the courseinfluenced the production functionrelationship between the standardinput measures and the course gradesof those who stayed, even thoughthe course drop rate was only 12percent. Becker and Walstad (1990)reveal yet another source of selec-tion bias when test scores are tobe explained; if test administration isvoluntary, teachers who observe thattheir average class score is low onthe pretest may not administer the post-test.

Missing observations on key explana-tory variables can also devastate a largedata set. Data on students are obtainedtypically from the students themselveseven though students err greatly in thedata they self-report (Maxwell andLopus 1994). Using the TUCE III mi-croeconomics data set, Becker, Powers,and Saunders (1996) show that studiesincluding only those students who pro-vide data on themselves and take boththe pretest and posttest are suspect inassessing the contribution of class sizein student learning, unless explicit ac-count is taken of the sample selectionproblems caused by the large amountsof missing data on excluded students.Test score results gleaned from theTUCE III data set are sensitive to the


model specification and data lost be-cause of the variables employed. Be-cause there is no unique way to undothe censoring that is associated withmissing student data, any conclusiondrawn only from students who providedata must be viewed with skepticism.

In the absence of perfect randomizedexperiments, selection problems atsome point in the sampling process canalways be identified. But should wecare if we cannot teach a subject to theuninterested and unwilling? We are al-ways going to be teaching to self-se-lected individuals, so why should ourexperiments not reflect the actual con-ditions under which we work? Whyworry about what does not apply?16 Onthe other hand, if building enrollment isimportant, then the previously uninter-ested students are the ones that mustbe attracted. We need to understandthe selection process in choosing andpersisting in courses, as well as in meas-uring learning.

(3) Testing for the Learning Effect ofInstructional Variables. Recent eco-nomics of education literature advancesthe notion that instructional variablesare unimportant in explaining studentlearning (Eric Hanushek 1991).17 In

concluding that certain instructionalvariables are insignificant in explainingstudent test scores, researchers acceptthe null hypothesis of no average effectin the populations. Statisticians cringeat the idea of “accepting the null hy-pothesis.” The null hypothesis of nolearning effect can never be acceptedfor there is always another hypothesizedvalue, in the direction of the alternativehypothesis, that cannot be rejected withthe same sample data. The Type II er-ror inherent in accepting the null hy-pothesis is well known but largely ig-nored by researchers in education andeconomics alike.

The power of the test (ability to re-ject the null hypothesis) can always beraised by increasing the sample size.Thus, if statistical significance is thecriterion for a successful instructionalmethod, then ever larger sample sizeswill “deliver the goods.” Statistical sig-nificance of an instructional methodmight be demonstrated with a suffi-ciently large sample, but the differencein change scores will likely be trivial onmultiple-choice tests of 25 to 40 items(the number of questions typically re-quired to demonstrate a valid and reli-able test that able students can com-plete in a 50 to 75 minute period).18

Differences of only a few correct an-swers in pretest and posttest compari-sons of control and experimental group

16 James Heckman and Jeffrey Smith (1995) callattention to the difficulty in constructing a coun-terfactual situation for an alternative instructionalmethod when participation is voluntary or ran-domly assigned. Without a counterfactual situation(i.e., what would have happened if these samepeople were in the control group), it is impossibleto do assessment.

17 David Card and Alan Krueger (1996) report aconsistency across studies showing the importanceof school quality on a student’s subsequent earn-ings. They recognize that tests can be adminis-tered easily at any time in the education processand thus provide a cheap tool for monitoring pro-grams. In recognition of time lag for measuringearnings effects, they recommend the use of drop-out rates as an alternative to test scores for imme-diate and ongoing program assessment. After all,unless students finish their programs, they cannotenjoy the potential economic benefits.

18 Replication and aggregation across studies isone way to achieve larger sample sizes, but JessicaUtts (1991) demonstrates the folly of such meta-analyses. Recently, for example, Larry Hedges,Richard Lane, and Rob Greenwald (1994a, 1994b)use a meta-analysis involving an aggregation ofp-values to cast doubt on Hanushek’s assertion re-garding the relevance of expenditure on instruc-tional methods in generating test scores. A case-by-case review of their presentation of Hanushek’sdata, however, suggests that the focal point ofmuch discussion in education, the teacher/pupilratio (or class size), is irrelevant in explaining stu-dent performance when measured by test scores.It appears to be the aggregation method employedthat is producing the results of Hedges et al.


results are the rule, not the exception,even after adjusting for sample selec-tion.

In summary, the use of the educa-tional production functions with testscores as the only output measure is toonarrow. Pre-and posttest, single-equa-tion specifications, with potentially en-dogenous regressors, simply may not beable to capture the differences that weare trying to produce with diverseteaching methods. Adjustments for sam-ple selection problems are needed buteven after these adjustments with largesamples, failure to reject the null hy-pothesis of no instructional effect maypoint more to deficiencies in the multi-ple-choice test outcome measure or ap-plication of the classical experimentaldesign than to the failure of the alterna-tive instructional method under scru-tiny.

B. Other Student Outcomes

An economics major typically beginswith two introductory courses that em-phasize the principles of micro andmacroeconomics courses (taken primar-ily by students who are fulfilling re-quirements for other majors). Econom-ics majors then take two intermediatecourses in micro and macroeconomics, acourse in statistics/econometrics, andsome field courses (Siegfried et al.1991). One set of proficiencies that stu-dents should gain in the completion ofthe economics major includes gainingaccess to current economic knowledge,displaying command of that knowledge,displaying the ability to draw on thatknowledge, using economic knowledgeto explore issues, and creating newknowledge (Hansen 1986).

Hansen’s proficiencies demonstrate abelief, likely shared by most econo-mists, that courses in economics con-tribute more to student developmentthan what is measured by paper and

pencil tests. Although there have beenattempts to turn these proficiencies intomeasurable outcomes other than testscores (James McCoy, Don Chamber-lain, and Rob Seay 1994), multiple-choice, short-answer, and essay ques-tions form the basis for grades, with themajority of instructors in introductorycourses putting a 50 percent weight onmultiple-choice test scores in assigninggrades (Becker and Watts 1996). Termpapers and short papers are seldom as-signed and used in course grading. Lackof structured writing assignments thatlead to larger papers is unfortunategiven survey results at the University ofWisconsin showing that jobs held byformer undergraduate economics ma-jors require well-developed written andoral communication skills, some sensefor quantitative relationships, and abil-ity to complete large projects and workwith others (Hansen 1993, p. 264). Theneed for communication skills and in-teraction with others is also made clearin the Harvard Assessment Seminarssurvey results that “of all the skills stu-dents say they want to strengthen, writ-ing is mentioned three times more thanany other” (Light 1992, p. 8).

Because of resource constraints, in-structors of large enrollment introduc-tory courses claim they have little re-course but to use multiple-choice testsfor the majority of grading. They cannotengage large numbers of students in bigwriting projects. Curiously, these largeclasses are also detrimental to studentpersistence: the larger the initial classsize, the more likely a student will with-draw from an introductory economicscourse (Becker, Powers, and Saunders1996). Obviously, those who cannotcomplete introductory courses are notgoing to take more courses in econom-ics. In small upper-division economicscourses instructors argue against allo-cating time to foster written and oral


communication skills because of a beliefthat these skills are not emphasized onthe standardized exams used for admis-sion to professional and graduateschools.

Training to succeed in standardizedmultiple-choice tests requires a differ-ent kind of college environment than isthe case with other cognitive and af-fective outcomes (Astin 1993, p. 221).Performance on standardized tests in-creases in highly competitive environ-ments in which test-taking skills are em-phasized. But this does not say thatengaging students in writing, editing,and revision will not benefit them onthe verbal portion of the GRE as well asassist them in later large-scale writingprojects. The question we should ad-dress is: How do our tests and other ac-tivities relate to what students shouldbe able to do as a result of studying eco-nomics?

In measuring student learning in eco-nomics, instructors often distinguish be-tween classroom multiple-choice (fixedresponse), and essay or short-answer(construct-response) questions. But dothese different forms of time-con-strained testing capture different di-mensions of student performance or arethey measuring essentially the samething? Other than possible gender dif-ferences, research suggests little differ-ence between what is measured by fixedand construct-response tests, especiallywhen the cost of valid construction andreliable scoring is considered (BenjaminGreene 1997; Walstad and Becker 1994;Keith Lumsden and Alex Scott 1987).19

But this research along with faculty de-

bates about the merits of multiple-choice versus essay exams misses thebroader spectrum of educational out-comes students are to achieve and theway in which they are achieved. Re-search needs to move beyond simplemeasures of knowledge to considerwhat leads to student persistence dur-ing a course, into another course, orinto a major, and what skills studentsneed for future performance in theworkplace.

C. Student Evaluations of Teaching

Instructor success is typically judgednot by the student learning outcomesbut by student evaluation question-naires administered at the end ofcourses. Because student interest andmotivation are important for academicperformance and persistence, their per-ceptions about the effectiveness of thecourse and the quality of instruction areworthy of consideration.

Instructional units often design theirown systems for student evaluations ofinstructors. Psychometricians and in-structional specialists tell us, however,that only properly prepared studentevaluation instruments and proceduresare valid and reliable measures ofteaching (McKeachie 1990, pp. 195–96). The two most widely acknowledgedstudent-rating systems, for which na-tional norms exist, are the Student In-structional Report (SIR), from the Edu-cational Testing Service, and theInstructional Development and Effec-

19 Common wisdom in education is that womendo not do as well as men on more quantitativematerial assessed by time-constrained, multiple-choice tests, although within economic educationdifferences in course performance are believed todepend on whether the posttest or change score isused as the dependent variable (Siegfried and Fels1979). By the end of the college career, however,

a gap between men and women is still evident. Ofthe 5,815 men and 2,164 women electing to takethe GRE Subject Test in Economics between 1989and 1993, men averaged 651 and women averaged603 on this time-constrained, multiple-choice test,according to Mary Hirschfeld, Robert Moore, andEleanor Brown (1995). Even after adjusting forGPA, SAT scores, and math background, the male-female gap persists. Female students may lackconfidence when it comes to test taking in eco-nomics for they are more likely than males to dropa course before taking an exam (Greene 1997).


tiveness Assessment system (IDEA),from Kansas State University’s Centerfor Faculty Evaluation and Develop-ment. These two systems are designedto provide information on student per-ception of teaching effectiveness of theinstructor, the educational value of thecourse, and how much students believethey learned.

Dennis Aigner and Frederick Thum’s(1986) study of student evaluation ofeconomics classes at the University ofSouthern California lends support tothe educational psychologists’ argumentthat students appreciate an instructor’sclarity in exposition, ability to get stu-dents actively involved in class discus-sions, and willingness to accept studentideas. Instructor-student interactionvariables that can be controlled by theinstructor greatly influence the ratingsgiven by students.

Critics cite studies from a variety ofdisciplines, such as that of PhilipAbrami, Sylvia d’Apollonia, and PeterCohen (1990), and those in economics,such as Edward Gramlich and GlenGreenlee (1993), and the Aigner andThum study, to support claims thatteachers can buy higher student ratingsby giving higher grades. If this is true,however, the effect is generally re-garded as slight. For example, RichardStratton, Steven Myers, and RandallKing (1994) suggest that introducingstudent evaluations leads to an initialincrease in grades of about 11 percent,but that this increase is followed by sta-bilization and then possible decrease,after adjusting for student changes inlearning.

Students are known to treat the disci-plines differently in awarding courseevaluations; yet, differences in averagediscipline scores are often overlookedin making comparisons. Economics isone of the disciplines that is consis-tently at the bottom of both course and

instructor effectiveness scales for the 45disciplines in the SIR and IDEA na-tional data base (William Cashin 1990).We do not know why disciplines arerated differently by students, althoughfor economics it may be that studentevaluations cannot fully reflect the con-tribution of individual teacher skills tostudent learning of economics (Wattsand William Bosshardt 1991). The useof student evaluations as a measure ofteaching must take into account that astudent’s willingness to provide anevaluation is endogenous to the teach-ing-learning process (Becker, Powers,and Saunders 1996). Because 20 to 40percent of students do not completeevaluations, sample selection problemsmust be considered in analyzing and in-terpreting student evaluations of in-structors, courses, and teaching meth-ods.20 Although student evaluationscannot be disregarded, they should notbe viewed as the single best measure ofteaching; other outcome measures(course withdrawal rates, cognitivelearning, continuation into other eco-nomics courses, and career success)must be considered.

V. Conclusion

The thrust of my thesis could bewrong: The movement of liberal artscolleges to master’s and doctorate-granting institutions may not haveworked against their undergraduateprograms. Teaching methods may notinfluence enrollments and majors ineconomics. The current mix of teachingmethods, with an almost exclusive em-phasis on lecture, chalkboard, and text-book, and paper and pencil exams may

20 Over time a poor teacher’s student evaluationscores may rise, with fewer withdrawals, becauseinformed students learn to avoid this teacher inthe first place. This dynamic form of selection isdifficult to spot without years of data in which stu-dents are free to choose instructors.


be the optimal way to teach economics.Teaching methods that get students ac-tively involved in the learning processmay be a waste of time. It may be thatthose single-equation pre-posttest stud-ies, from which some have concludedthat instruction methods are irrelevant,are correct. But as long as students areable to vote with their feet, and thereare attractive alternative courses andmajors, academic economists are foolishto ignore the cost of a Type II error inaccepting the status quo.

To increase enrollments in introduc-tory courses, enhance persistencewithin these courses, and augment thenumber of majors, a move to smallerclass sizes is supported by the litera-ture. Small classes, in which studentsare actively engaged in the learningprocess, through discussion and small-group activities, encourage persistenceand appear to be conducive to learning,especially if combined with rapid feed-back and positive reinforcement. Unlikelectures for which class notes can be re-produced, student involvement in class-room activities requires attendance.Even for large enrollment introductorycourses, where small sections may notbe possible, teaching methods otherthan lecture can be employed. From anefficiency standpoint some may arguethat the actual size of the classroom isless important than what is done to en-gage students, but without knowing thevalue of the educational outputs effi-ciency is difficult to define. The criticalpoint for either small or large classes,however, is student participation. Toengage students, instructors must giveup the comfortable position of 10 to 14weeks of lecture, with graphs and equa-tions following in an orderly progres-sion that students are expected to sim-ply reproduce on exams.

The granting of higher grades maycontribute to persistence especially for

women, if only in the short run, forsuch a strategy is dependent on the re-sponse of other units within an institu-tion and maybe even among institu-tions. There is little empirical evidencethat lowering mathematics require-ments will increase enrollments. Stu-dents with stronger mathematics back-grounds, however, are more likely topersist in the study of economics.

Instead of thinking only of valueadded with a pre-posttest mind set, orend-of-term student evaluations of in-structors, economists must considerother outcome measures that directlyaffect both their students’ persistenceand their own survival as academicians.The numerous research questions thatarise once we move beyond a simplepre-posttest analysis have been ad-dressed by Becker et al. (1991), as wellas others, so I deliberately avoid repeat-ing them here. To assist practitioners intheir search for better teaching meth-ods, economic education researchersmust consider the sample selectionproblems inherent in pre-post assess-ment, and adequately model the stu-dent and instructor decision makingprocess. In the absence of these ad-vances in research, however, it may be acostly mistake to conclude that instruc-tional methods do not matter simply be-cause they have not been shown to in-fluence multiple-choice test scores insingle-equation regressions with en-dogenous regressors.

REFERENCES

ABRAMI, PHILIP C.; D’APOLLONIA, SYLVIA ANDCOHEN, PETER A. “Validity of Student Ratingsof Instructors: What We Know and What WeDo Not,” J. Educational Psychology, June 1990,82(2), pp. 219–31.

AIGNER, DENNIS J. AND THUM, FREDERICK D.“On Student Evaluation of Teaching Ability,” J.Econ. Educ., Fall 1986, 17(4), pp. 243–65.

ANDERSON, GORDON; BENJAMIN, DWAYNE ANDFUSS, MELVYN. “The Determinants of Successin University Introductory Economics Courses,”J. Econ. Educ., Spring 1994, 25(2), pp. 99–119.


ANDERSON, MARTIN. Impostors in the temple.New York: Simon and Schuster, 1992.

ASTIN, ALEXANDER W. What matters in college?San Francisco: Jossey-Bass, 1993.

BARTLETT, ROBIN L. “A Flip of the Coin—A Rollof the Die: An Answer to the Free-Rider Prob-lem in Economic Instruction,” J. Econ. Educ.,Spring 1995, 26(2), pp. 131–39.

BATEMAN, BRADLEY W. “The Education ofEconomists: A Different Perspective,” J. Econ.Lit., Sept. 1992, 30(3), pp. 1491–95.

BECKER, WILLIAM E. “The Educational Processand Student Achievement Given Uncertainty inMeasurement,” Amer. Econ. Rev., Mar. 1982,72(1), pp. 229–36.

–––––. “The Use of Mathematics and Statistics inThe Teaching of Economics,” The principles ofeconomics course: A handbook for instructors.Eds.: PHILLIP SAUNDERS AND WILLIAM B.WALSTAD. New York: McGraw-Hill, 1990, pp.180–90.

BECKER, WILLIAM E.; POWERS, JOHN AND SAUN-DERS, PHILLIP. “Problems of Missing StudentData and The Importance of Class Size inStudent Achievement.” Working paper, Apr.1997.

BECKER, WILLIAM E. AND WALSTAD, WILLIAMB. “Data Loss from Pretest to Posttest as a Sam-ple Selection Problem,” Rev. Econ. Statist.,Feb. 1990, 72(1), pp. 184–88.

BECKER, WILLIAM E. AND WATTS, MICHAEL.“Teaching Tools: Teaching Methods in Under-graduate Economics,” Econ. Inquiry, Oct. 1995,33(4), pp. 692–700.

–––––. “Chalk and Talk: A National Survey onTeaching Undergraduate Economics,” Amer.Econ. Rev., May 1996, 86(2), pp. 448–53.

BECKER, WILLIAM E. ET AL. “An Agenda for Re-search on Economic Education in Colleges andUniversities,” Amer. Econ. Rev., May 1991,81(2), pp. 26–31.

BENZING, CYNTHIA AND CHRIST, PAUL. “A Sur-vey of Teaching Methods Among EconomicsFaculty,” J. Econ. Educ., Spring 1997, 28(2) pp.182–88.

BILLARD, LYNNE. “President’s Corner: Our Rolein Education,” Amstat News, Mar. 1996, p. 9.

BODENHORN, HOWARD. “Teachers, and ScholarsToo: Economic Scholarship at Elite Liberal ArtsColleges,” J. Econ. Educ., forthcoming.

BRASFIELD, DAVID ET AL. “Why Have SomeSchools Not Experienced a Decrease in thePercentage of Students Majoring in Economics?“J. Econ. Educ., Fall 1996, 27(4), pp. 362–70.”

BRENEMAN, DAVID W. Liberal arts colleges:Thriving, surviving, or endangered? Washing-ton, DC: Brookings Institution, 1994.

CARD, DAVID AND KRUEGER, ALAN. “The Eco-nomic Return to School Quality,” Assessingeducational practices: The contribution of eco-nomics. Eds.: WILLIAM E. BECKER AND WIL-LIAM J. BAUMOL. Cambridge: MIT Press, 1996,pp. 161–82.

CARNEGIE FOUNDATION FOR THE ADVANCE-

MENT OF TEACHING. A classification of institu-tions of higher education. 1994 ed. Princeton,NJ: Princeton U. Press.

CASHIN, WILLIAM. “Students Do Rate DifferentAcademic Fields Differently,” Student ratingsof instruction: Issues for improving practice.Eds.: MICHAEL THEALL AND JENNIFERFRANKLIN. New Directions for Teaching andLearning, no. 43 (Fall 1990). San Francisco:Jossey-Bass, 1990, pp. 113–21.

CHIZMAR, JOHN F. “Total Quality Management(TQM) of Teaching and Learning,” J. Econ.Educ., Spring 1994, 25(2), pp. 179–90.

COHN, ELCHANAN AND COHN, SHARON. “Graphsand Learning in Principles of Economics,”Amer. Econ. Rev., May 1994, 84(2), pp. 197–200.

COHN, ELCHANAN; COHN, SHARON ANDBRADLEY, JAMES JR. “Notetaking, WorkingMemory, and Learning in Principles of Eco-nomics,” J. Econ. Educ., Fall 1995, 26(4), pp.291–307.

COLLEGE BOARD. Annual survey of colleges:Summary statistics. New York: College En-trance Examination Board, 1992–93, 1993–94.

DOUGLAS, STRATFORD AND SULOCK, JOSEPH.“Estimating Educational Production Functionswith Correction for Drops,” J. Econ. Educ.,Spring 1995, 26(2), pp. 101–12.

DURDEN, GAREY C. AND ELLIS, LARRY V. “TheEffects of Attendance on Student Learning inPrinciples of Economics,” Amer. Econ. Rev.,May 1995, 85(2), pp. 343–46.

DYNAN, KAREN AND ROUSE, CECILIA. “The Un-derrepresentation of Women in Economics: AStudy of Undergraduate Economics Students,”J. Econ. Educ., forthcoming.

EHRENBERG, RONALD G.; KASPER, HIRSCHELAND REES, DANIEL I. “Faculty Turnover atAmerican Colleges and Universities: Analyses ofAAUP Data,” Econ. Educ. Rev., 1991, 10(2), pp.99–110.

GEIGER, ROGER AND FELLER, IRWIN. “The Dis-persion of Academic Research in the 1980s,” J.Higher Education, May/June 1995, 66(3), pp.336–60.

GELLES, GREGORY M. AND JOHNSON, WALTERD. “An Overlooked Opportunity: Instituting aFormal Economics Minor,” J. Econ. Educ.,Summer 1994, 25(3), pp. 271–74.

GRAMLICH, EDWARD M. AND GREENLEE, GLENA. “Measuring Teaching Performance,” J. Econ.Educ., Winter 1993, 24(1), pp. 3–13.

GREENE, BENJAMIN. “Verbal Abilities, Gender,and the Introductory Economics Course: ANew Look at an Old Assumption,” J. Econ.Educ., Winter 1997, 28(1), pp. 13–30.

GRIMES, PAUL W. AND NISS, JAMES F. “Eco-nomic Understanding and Student Success in aBusiness Curriculum,” J. Education for Busi-ness, May/June 1991, 66(5), pp. 309–13.

HANSEN, W. LEE. “What Knowledge Is MostWorth Knowing—For Economics Majors?”Amer. Econ. Rev., May 1986, 76(2), pp. 149–52.

———. “Bringing Total Quality Improvement into


the College Classroom,” Higher Education.Apr. 1993, 25(3), pp. 259–79.

HANUSHEK, ERIC. “When School Finance ‘Re-form’ May Not Be a Good Policy,” HarvardJournal of Legislation, 1991, 28, pp. 423–56.

HARTLEY, JAMES AND ROBINSON, MICHAEL.“Economic Research at National Liberal ArtsColleges: School Rankings,” J. Econ. Educ.,forthcoming.

HECKER, DANIEL E. “Earnings of College Gradu-ates, 1993,” Mon. Lab. Rev., Dec. 1995,118(12), pp. 3–17.

HECKMAN, JAMES J. AND SMITH, JEFFREY A. “As-sessing the Case for Social Experiments,” J.Econ. Perspectives, Spring 1995, 9(2), pp. 85–110.

HEDGES, LARRY V.; LANE, RICHARD ANDGREENWALD, ROB. “Does Money Matter? AMeta-Analysis of Studies of the Effects of Dif-ferential School Inputs on Student Outcomes,”Educational Researcher, Apr. 1994a, 23(3), pp.5–14.

–––––. “Money Does Matter Somewhat: A Replyto Hanushek,” Educational Researcher, May1994b, 23(4), pp. 9–10.

HINSHAW, ELTON. “Report of the Director: JobOpenings for Economists,” Amer. Econ. Rev.,May 1975, 65(2), pp. 488–89.

HIRSCHFELD, MARY; MOORE, ROBERT L. ANDBROWN, ELEANOR. “Exploring the Gender Gapon the GRE Subject Test in Economics,” J.Econ. Educ., Winter 1995, 26(1), pp. 3–15.

HORVATH, JANE; BEAUDIN, BARBARA Q. ANDWRIGHT, SHEILA P. “Persisting in the Intro-ductory Economics Course: An Exploration ofGender Differences,” J. Econ. Educ., Spring1992, 23(2), pp. 101–08.

JOHNSON, DAVID W.; JOHNSON, ROGER T. ANDSMITH, KARL A. Cooperative learning: Increas-ing college faculty instructional productivity.ASHE-ERIC Higher Education Report No4.Washington, DC: School of Education andHuman Development, George Washington U.,1991.

JOHNSTONE, ALEX H. AND PERCIVAL,FREDERICK. “Attention Breaks in Lecture”Education in Chemistry, 1976, 13, pp. 49–50.

KASPER, HIRSCHEL. “The Undergraduate RevoltAgainst Economics.” Paper presented at theASSA meetings, San Francisco, 1996.

KASPER, HIRSCHEL ET AL. “The Education ofEconomists,” J. Econ. Lit., Sept. 1991, 29(3),pp. 1088–1109.

KEENAN, DIANE AND MAIER, MARK H. Econom-ics live: Learning economics the collaborativeway. New York: McGraw-Hill, 1995.

KRUEGER, ANNE O. ET AL. “Report of the Com-mission on Graduate Education in Economics,”J. Econ. Lit., Sept. 1991, 29(3), pp. 1035–53.

LIGHT, RICHARD J. The Harvard assessment semi-nars. Cambridge, MA: Harvard U. GraduateSchool of Education and Kennedy School ofGovernment, First Report 1990.

–––––. The Harvard assessment seminars. Cam-

bridge, MA: Harvard U. Graduate School ofEducation and Kennedy School of Government,Second Report 1992.

LUMSDEN, KEITH G. AND SCOTT, ALEX. “TheEconomics Student Reexamined: Male-FemaleDifferences in Comprehension,” J. Econ. Educ.,Fall 1987, 18(4), pp. 365–75.

MARGO, ROBERT A. AND SIEGFRIED, JOHN J.“Long-Run Trends in Economics Bachelor’sDegrees,” J. Econ. Educ., Fall 1996, 27(4), pp.326–36.

MAXWELL, NAN L. AND LOPUS, JANE S. “TheLake Wobegon Effect in Student Self-ReportedData,” Amer. Econ. Rev., May 1994, 84(2), pp.201–05.

MCCOY, JAMES P.; CHAMBERLAIN, DON ANDSEAY, ROB. “The Status and Perceptions of Uni-versity Outcomes Assessment in Economics,” J.Econ. Educ., Fall 1994, 25(4), pp. 358–66.

MCKEACHIE, WILBERT J. “Research on CollegeTeaching: The Historical Background,” J. Edu-cational Psychology, June 1990, 82(2), pp. 189–200.

MCKINNEY, KATHLEEN AND GRAHAM-BUXTON,MARY. “The Use of Collaborative LearningGroups in the Large Class: Is It Possible?”Teaching Sociology, Oct. 1993, 21(4), pp. 403–08.

NATIONAL CENTER FOR EDUCATION STATISTICS.Digest of education statistics. U.S. Departmentof Education, Office of Educational Researchand Improvement, various years.

PARK, KANG H. AND KERR, PETER M. “Determi-nants of Academic Performance: A MultinomialLogit Approach,” J. Econ. Educ., Spring 1990,21(2), pp. 101–11.

ROMER, DAVID. “Do Students Go to Class?Should They?” J. Econ. Perspectives, Summer1993, 7(3), pp. 167–74.

ROSEN, SHERWIN. “Some Economics of Teach-ing,” J. Lab. Econ., Oct. 1987, 5(4) Part 1, pp.561–75.

ROTHSCHILD, MICHAEL AND WHITE, LAWRENCEJ. “The University in the Marketplace: Some In-sights and Some Puzzles,” in Studies of supplyand demand in higher education. Eds.: CHAR-LES T. CLOTFELTER AND MICHAEL ROTH-SCHILD. Chicago: U. of Chicago Press, 1993,pp. 11–37.

SABOT, RICHARD H. AND WAKEMAN-LINN, JOHN.“Grade Inflation and Course Choice,” J. Econ.Perspectives, Winter 1991, 5(1), pp. 159–70.

SALEMI, MICHAEL K. AND EUBANKS, CARLIE.“Accounting for the Rise and Fall in the Num-ber of Economics Majors with the Discouraged-Business-Major Hypothesis,” J. Econ. Educ.,Fall 1996, 27(4), pp. 350–61.

SALEMI, MICHAEL K.; SAUNDERS, PHILLIP ANDWALSTAD, WILLIAM B. “Teacher Training Pro-grams in Economics: Past, Present, and Fu-ture,” Amer. Econ. Rev., May 1996, 86(2), pp.460–64.

SALEMI, MICHAEL K. AND TAUCHEN, GEORGE E.“Simultaneous Nonlinear Learning Models,” in


Econometric modeling in economic educationresearch. Eds: WILLIAM E. BECKER ANDWILLIAM B. WALSTAD. Norwell, MA: KluwerAcademic, Nijhoff Publishing, 1987, pp. 207–23.

SAUNDERS, PHILLIP. “The Third Edition of theTest of Understanding in College Economics,”J. Econ. Educ., Summer 1991, 22(3), pp. 255–72.

SAX, LINDA J. ET AL. The American collegeteacher: National norms for the 1995–96 HERIfaculty survey. Los Angeles: Higher EducationResearch Institute, UCLA, 1996.

SIEGFRIED, JOHN J. AND FELS, RENDIGS. “Re-search on Teaching College Economics: A Sur-vey,” J. Econ. Lit., Sept. 1979, 17(3), pp. 923–69.

SIEGFRIED, JOHN J. ET AL. “The Status and Pros-pects of the Economics Major,” J. Econ. Educ.,Summer 1991, 22(3), pp. 197–224.

SIEGFRIED, JOHN J. ET AL. “Teaching Tools:How Is Introductory Economics Taught inAmerica?” Economic Inquiry, Jan. 1996, 34(1),pp. 182–92.

STRATTON, RICHARD W.; MYERS, STEVEN C. ANDKING, RANDALL K. “Faculty Behavior, Grades,and Student Evaluations,” J. Econ. Educ., Win-ter 1994, 25(1), pp. 5–15.

STROBER, MYRA; COOK, ALLEN AND FULLER,KASI. “Making and Correcting Errors in Eco-

nomic Analyses: An Examination of Vide-otapes,” J. Econ. Educ., forthcoming.

UTTS, JESSICA. “Replication and Meta-Analysis,”Statistical Science, 1991, (4), pp. 363–78.

WALSTAD, WILLIAM B. AND BECKER, WILLIAME. “Achievement Differences on Multiple-Choice and Essay Tests in Economics,” Amer.Econ. Rev., May 1994, 84(2), pp. 193–96.

WATTS, MICHAEL AND BOSSHARDT, WILLIAM.“How Instructors Make a Difference: PanelData Estimates from Principles of EconomicCourses,” Rev. Econ. Statist., May 1991, 72(2),pp. 336–40.

WATTS, MICHAEL AND LYNCH, GERALD J. “ThePrinciples Courses Revisited,” Amer. Econ.Rev., May 1989, 79(2), pp. 236–41.

WHITE, LAWRENCE J. “Efforts by Departments ofEconomics to Assess Teaching Effectiveness:Results of an Informal Survey,” J. Econ. Educ.,Winter 1995, 26(1), pp. 81–85.

WILLIAMS, ARLINGTON W. AND WALKER, JAMESM. “Computerized Laboratory Exercises forMicroeconomics Education: Three Applica-tions Motivated by Experimental Economics,”J. Econ. Educ., Fall 1993, 24(4), pp. 291– 315.

WILLIS, RACHEL A. AND PIEPER, PAUL J. “TheEconomics Major: A Cross-Sectional View,” J.Econ. Educ., Fall 1996, 27(4), pp. 337–49.


becker: teaching economics to undergraduates teaching ... · becker: teaching economics to...

Documents