The professional development“system” for teachers is, by all ac-counts, broken. Despite evi-dence that specific programs canimprove teacher knowledge andpractice and student outcomes,these programs seldom reachreal teachers on a large scale. To

use a shopping metaphor, these research-proven pro-grams, which are often offered by university faculty ornationally recognized providers, are “boutiques” serv-ing a handful of fortunate teachers while leaving manymore to shop at the Wal-Marts of the professional de-velopment world. There, most teachers receive unin-spired and often poor-quality professional develop-ment and related learning opportunities.

Typically, reformers address such perceptions offailure by discovering and celebrating new formatsand content for teacher professional learning. In thepast two decades alone, advocates of continuingteacher education have promoted school-based learn-ing opportunities, such as coaching and lesson study;new topics, in the form of increased focus on subject

470 PHI DELTA KAPPAN Image: Liquid Library

Fixing Teacher Professional DevelopmentA broken system of professionallearning requires decisive action inorder to ensure wise expenditure oflimited resources.

BY HEATHER C. HILL

n Heather C. Hill is an associate professor in the Harvard Gradu-ate School of Education, Cambridge, Massachusetts.

The Leading Edge | Professional Learning

MARCH 2009 471

matter content and, more recently, the analysis of as-sessment and related data; and new delivery mecha-nisms, including content transmitted only online.

However, without a hard look at the actors in the sys-tem and the incentives facing those actors, these fadswill prove the educational equivalent of pouring newwine into old bottles. Instead, policy makers must in-vest in fixing the system writ large in order to have adiscernible impact on teaching and learning.

Participation Doesn’t Mean ResultsFrom the outside, one might not see the U.S. pro-

fessional development system as particularly trou-bled. Nearly every teacher participates in some formof learning every year. Research articles trumpetingthe success of a particular method or program appearpractically monthly, and practitioner magazines burstwith accounts of the phenomenal improvements inteacher knowledge and skills that result.

But buried beneath these often-glowing reports arecolder facts. For instance, from the available evidence,teachers apparently have little use for their learningexperiences: Most teachers engage in only the mini-mum professional learning required by their state ordistrict each year. In 1999-2000, the most recent yearavailable, National Center for Education Statistics(NCES) data showed that just over half of respon-dents to an NCES survey reported spending a day orless in professional development over the past year;only a small minority reported attending four or moredays within the past year (2001). This generally lowrate of participation closely matches many state’s re-licensure requirements, typically 15 days over a five-year period (NASDTEC 2004), suggesting that mostteachers do the bare minimum required under law.

Furthermore, new and exciting forms of profes-sional development guarantee neither high-quality

Without a hard look at the actors in the system and theincentives facing those actors,these fads will prove theeducational equivalent ofpouring new wine into oldbottles.

472 PHI DELTA KAPPAN

delivery nor substantive effects on teachers, teaching,and learning. Consider lesson study, a Japanese modelof site-based, ongoing study of how specific lessons af-fect student learning. Evidence suggests that lessonstudy has been widely implemented across the coun-try. In my own recent survey (2008), 51% of middle

school math teachers reported engaging in lessonstudy during the prior year. However, as with manyeducational innovations, lesson study appears to havebeen quickly subsumed into the standard operatingstructure of U.S. schools. Although the model callsfor an intensive time commitment by teachers, almost60% of those who participated reported spendingeight or fewer hours during the year; only 4% re-ported engaging in over 80 hours. Perhaps as a result,this model did not predict teachers’ gains in mathknowledge over the year in question.

Finally, teachers themselves are lukewarm abouttheir professional development experiences. Whenqueried about the impact of the past three years ofprofessional development experiences, less than aquarter, on average, reported that professional devel-opment affected their instruction (Horizon 2002).Most teachers, in fact, reported that professional de-velopment reinforced their existing practices, and aminority reported no effect at all.

Teachers’ observations tend to match those madeby academics who study “Wal-Mart”-style teacherlearning opportunities. Such studies help shed lighton both the system of professional development andteacher reluctance to participate. For instance, thoughboutique programs aren’t immune from problems,my own observations over the past decade suggestthat at least in mathematics, the professional develop-ment reaching regular teachers through district con-tracts, regional conferences, and similar means can bequite poor. Despite a number of high-quality pro-grams and sessions, others covered math only super-ficially, contained mathematical ambiguities and er-rors, or provided inaccurate information about stu-dent learning.

One reason for this variability is the capacity ofproviders. Although some are highly skilled in the ar-

eas in which they provide training, this isn’t univer-sally the case. In 2005, I surveyed individuals provid-ing mathematics professional development to teach-ers, asking them about their background and prompt-ing them to solve a series of math problems. Most re-spondents said providing mathematics professionaldevelopment was not their only responsibility; otherresponsibilities included teaching, coordinating cur-riculum for districts, and consulting. Roughly halfprovided professional development in subjects otherthan math. Their performance on the math assess-ment varied; in fact, when compared to a large sam-ple of teachers who took the same assessment, roughlyone-sixth of the professional development providersfell below the 50th percentile of the teacher sample.

Even if teachers’ learning opportunities are of mod-erate quality and contain no errors or unproven facts,there is also the problem of transfer. In one recentstudy, we saw teachers taking lessons or activities fromprofessional development into their classrooms, oftento ill effect. In many cases, the activities were im-ported into classrooms without the mathematics theywere meant to represent; in others, the math was pres-ent but distorted (Hill 2008).

Finally, too much professional development canactually decrease instructional coherence. District of-ficials have more than once expressed frustration be-cause professional development advice and supple-mental materials undermine district-adopted curric-ula and instructional approaches.

The four points above — quality of the product,capacity of the providers, transfer, and coherence —suggest that we must reexamine our assumption that“ineffective” professional development is benign.While the evidence for this point is anecdotal, the ex-istence of “malignant” professional developmentwould help to explain why, in many studies, teacherattendance at professional development has not beenassociated with gains in student outcomes.

Fixing What’s BrokenRather than replacing one form of professional de-

velopment with another, we would be wiser to exam-ine what exists and to make it better. The problem isnot that we lack promising programs, formats, or con-tent; it’s that they rarely reach the typical teacher in aform that maintains their integrity and effects. Use aneconomic perspective to consider why this occurs.

First, recognize that continuing teacher educationis big business in the United States. Estimates placeprofessional development spending at between 1%and 6% of district expenditures (Hertert 1997;

Teachers themselves arelukewarm about theirprofessional developmentexperiences.

MARCH 2009 473

Killeen, Monk, and Plecki 2002; Odden et al. 2002;Miles 2003). However, these figures don’t considerprofessional development provided by states or fed-eral spending to support professional learning forteachers. For example, the National Science Founda-tion and U.S. Department of Education Math-Sci-ence Partnerships spent nearly $1.2 billion (NSF2007) on mathematics and science learning for pre-service and inservice teachers between the years 2002and 2007. And these figures are no doubt matched bywhat teachers spend to fund their own educational ex-periences.

Once the professional development system is rec-ognized as a marketplace, tools for imagining im-provements start to emerge. Economists often exam-ine markets from four key perspectives: supply, de-mand, information, and efficiency. All four are usefulin delineating the challenges facing efforts to reformcontinuing teacher education.

The most obviously applicable perspective is sup-ply. Here, “supply” means all available professionaldevelopment opportunities, analogous to the amountof oil on the world market or the number of widgetsproduced by manufacturers. The problem with thesupply side of professional development is that thereis an almost infinite supply. Perhaps because face-to-face interactions between providers and teachers areconsidered paramount, professional development ismostly local. In our 2005 survey, for instance, mostprofessional developers lived in the state in which theyprovided services, often within 50 miles of the districtthat referred them to us. Local demand for providersmeans a large number of small firms (or individuals)enter this business, often without special training orspecific expertise. This, in turn, leads to products thatare low-quality, offering teachers only quick fixes and,in some worst cases, misinformation. The challengeis how to improve the quality, not the quantity, ofwhat is available.

Think about the demand side of the professional

development marketplace. “Demand” means the av-erage consumer’s desire for professional developmentand related programs. Demand has a time component— teachers want or need a certain number of hours— and a content component. On both fronts, teach-ers face only modest inducements to invest in theirown learning. States typically require only two tothree days per year and, because of experience withpoorly designed learning opportunities, teachers havelittle appetite for more, or more intellectually chal-lenging, fare. Why spend a day attending a question-able program when that day could be spent correct-ing papers or planning a unit? These opportunitycosts, coupled with a misguided formal incentivestructure, mean that demand for high-quality profes-sional development is typically weak.

Information about product quality is the glue thatholds markets together. It allows consumers to makewise buying decisions and also informs suppliersabout a product’s sales potential and price. Yet virtu-ally no information exists to help consumers of pro-fessional development. Teachers gamble on whetherprofessional development program A or B will im-prove their ability to connect with students, delivercontent, and enhance learning. The same is true fordistricts, which typically conduct only the barest ofevaluations on their smorgasbord of professional de-velopment offerings. In this situation, consumerscan’t make “rational” choices — ones that improvetheir own practice and their student’s outcomes. Nordo effective programs thrive and ineffective programsfade away, another feature of information-rich mar-kets. Driving ineffective programs and providers fromthe marketplace is particularly critical, given thelargely local nature of this enterprise and the low bar-riers to entry.

Efficiency asks whether teachers have access to theprofessional learning they need. Do teachers who re-quire assistance in teaching early reading enroll inprograms specifically designed around these topics?Do teachers who need more mathematical contentknowledge appear in mathematics-focused learningopportunities? This is a seldom-investigated questionin the United States, for little is known about mostteachers’ capabilities and knowledge. Yet the evidencethat exists is not promising. In my own recent study,teachers whose mathematical knowledge was lowwere not more likely, over the next year, to enroll incontent-focused workshops. This reframes debatesabout whether teachers or districts should “choose”which program to attend. Rather than debating themerits of teachers versus districts, analysts should note

Driving ineffective programs and providers from themarketplace is particularlycritical, given the largely localnature of this enterprise and the low barriers to entry.

474 PHI DELTA KAPPAN

that such decisions are rarely made through an analy-sis of specific teachers’ deficits and needs.

Improving PoliciesState and district policy makers can restructure the

professional development system in four ways.Fix the supply side. Adopt a medical model. In

medicine, the Accreditation Council for ContinuingMedical Education (ACCME) developed and main-tains a set of standards for continuing medical educa-tion. In education, the National Staff DevelopmentCouncil has identified a set of standards for profes-sional development but does not certify or endorseproviders. However, in medicine, there are teeth inthe regulations that govern the supply of providers.Those teeth come in the form of an accreditationprocess with attached incentives for practitioners.

First, medical doctors must complete 20 to 40hours of continuing medical education (CME) eachyear through an ACCME-accredited organization.Second, ACCME exerts moderate control over whoand what can be accredited. For instance, ACCMEaccredits only specific models of continuing medicaleducation, including live interactions, journal read-ing, Internet coursework, writing test items for licen-sure exams, and reviewing for journals. ACCME alsoholds providers accountable for the content theyteach doctors, specifies how it will validate that con-tent, and approves all course material. In particular,the organization requires that all recommendationsmade by providers must be based on evidence – andfurther, that the evidence itself be generated throughexperimental design.

In education, such guidelines would eliminateworkshops or related learning opportunities in whichthe provider “discovers” an idea about content or stu-dent learning or “invents” a new way to teach mate-rials, then passes recommendations along to teachersbased only on his or her own thinking. Instead, con-tent would have to be grounded in empirical study.

A challenge in education is whether the empiricalresearch base in education is wide enough to sustainsuch content restrictions and which organizationswould step forward to help regulate the quality oflearning opportunities. Could those organizations ef-fectively monitor members of their own community?Nevertheless, the need for quality control is urgent,and medicine’s experiences provide one model.

Fix the “demand” side. Increasing the demand forprofessional development depends on improving thequality of such experiences; teachers won’t attend un-less they perceive benefit, and rightly so. However, if

we want teachers to invest in substantial and effectivelearning opportunities, we must also change teachers’“taste” for learning.

The main strategy is to align professional educationincentives, models, and norms with what we wish tosee teachers do. Although there is a strong utopianideal surrounding teacher learning — that everyteacher is a “continual learner” striving to better her-self for the good of students — the reality is thatteachers, like other professionals, respond to the in-centives, norms, and models that surround them.This suggests that states and districts reconsider pro-fessional development requirements that ask only fora certain number of hours within a certain time spanand establish incentives for deeper investment — oreven criteria based on student learning.

Policy makers also should explicitly encourage,both in official policy and rhetoric, teacher invest-ment in the programs that we know work, as well asin more general types of professional developmentthought to improve teaching and learning. Recently,some have argued that content-focused professionaldevelopment based on classroom practice — includ-ing evidence around student learning, the study ofcurriculum materials, and so forth — is most likelyto affect teacher knowledge and performance and stu-dent outcomes (Cohen and Hill 2001; Garet et al.2001; Hill and Ball 2004). Explicitly encouraging in-vestment in these opportunities via policy and educa-tion’s bully pulpit, so to speak, can help.

Improve “information” about program quality.Conduct small-scale but rigorous studies that meas-ure the effectiveness of local and regional professionaldevelopment programs and suppliers. There would becomplications: Suppliers often change the content oftheir activities according to district specifications; lo-cal evaluators lack capacity to engage in rigorous re-search, often preferring to observe and offer “forma-tive” feedback; providers themselves are often loatheto be evaluated; and studies involving the outcomethat matters most — student learning — are difficultto mount. Yet with time and attention, educators cansolve all of these problems.

Another avenue is to observe and rate professionaldevelopment opportunities according to stringentand content-specific guidelines. This would be simi-lar to accreditation, but with a focus on enactmentrather than simply materials and plans. Programs fo-cused on improving teachers’ mathematical knowl-edge, for instance, would have to display dense andaccurate mathematics, opportunities for teachers todelve deeply into mathematical ideas and perform de-

MARCH 2009 475

manding mathematical work, and opportunities forpractice and refinement of key mathematical tasks ofteaching. One can imagine similar criteria for othersubjects.

Whatever occurs, publicize the results. Little isgained from collecting information for only bureau-cratic purposes. Instead, put the information, in di-gestible form, into the hands of those making choicesabout ongoing learning opportunities.

Improve efficiency. Perhaps the most difficultproblem will be ensuring that continuing teacher ed-ucation is suited to and suitable for the educators itserves. Ironically, the U.S. education system has a longhistory of differentiated instruction — placing stu-dents in courses or tracks according to perceived tal-ent and prior achievement. But, teacher learning op-portunities are typically generic, one size fits all. Se-lection into these opportunities occurs almost ran-domly; a highly expert writing teacher is as likely toattend a writing workshop as an unskilled teacher inthis arena. This means professional development dol-lars, in particular, may not be used as efficiently or ef-fectively as they could be.

Professional development will be more effectiveand more efficient if we link specific teachers’ weak-nesses with the learning opportunities most likely toremedy those weaknesses. For instance, teachers scor-ing below a cutoff in math knowledge would be re-quired to attend math-focused coursework. Or,teachers who fail over several years to perfect class-room management routines would be paired withothers who are expert in this arena. This entails amuch more nuanced and intrusive system of teacherevaluation than we now have.

One possibility, now under consideration in Mass-achusetts (2008), is to test all math teachers in low-performing schools on their mathematical knowledge.My own research shows that teachers’ math knowledgecorrelates highly with what we call the “mathematicalquality of instruction” — or the presence/absence ofmath errors, the richness of the math presented to stu-dents, and teachers’ ability to work with student mathproductions and errors (Hill et al. in press). Teachers’math knowledge also predicts gains in student achieve-

ment (Hill, Rowan, and Ball 2005). Using mathknowledge as a proxy for problematic instruction andstudent gains, then, the state would identify teachersin need and require a remediation plan, developed inconsultation with the building principal or mathcoach. While this approach hasn’t been implemented,it provides one possible model for improving the effi-ciency of the system, matching need to resources.

If You Don’t Fix It, Scrap ItProfessional development has led to improvements

in teachers’ knowledge, instruction, and student out-comes. Both a body of research studies and anecdotalevidence suggest such effects. But the question is this:In the wider system, are improvements in teachingand learning worth the expense, in terms both of dol-lars and teachers’ time? Currently, there is no way toanswer this question. We can’t estimate what percent-age of professional development is worthwhile, interms of return on investment, or whether thoselearning opportunities would continue to occur in theabsence of formal requirements to participate in pro-fessional development.

Nonetheless, here’s some advice for the states, dis-tricts, and providers who jointly govern the profes-sional development system: If you don’t fix it, scrap it.We should abandon professional development thatexists only to fulfill state licensure requirements. Fun-ders, whether from government or foundations,should support only proven and highly promisingprograms.

Fixing it would mean moving to a system charac-

We should abandon professional development thatexists only to fulfill statelicensure requirements.

476 PHI DELTA KAPPAN

terized by improvements in many of the elements de-scribed above: high-quality suppliers; teacher “de-mand” for challenging, relevant, and lengthy oppor-tunities to learn; efficient matching of teacher needsto resources and programs; and lots of information tohelp calibrate market forces. Continuing the eco-nomic theme, we might see a transition from a largelylocally driven market to what economists call an oli-gopoly, in which a few providers dominate the mar-ket because of their proven quality and effects. Towork, such programs would have to be nationally scal-able, accessible, and affordable to all districts. Severalprograms, such as Marilyn Burns’ Math Solutions,PBS TeacherLine, and LessonLab’s modules, seemheaded in this direction and offer models from whichto work. Should they be proven effective and becomewidely adopted, they could revolutionize professionaldevelopment in the United States.

Finally, getting to such a system will not be easy. Itwould require major overhauls of nearly every ele-ment that supports the current system, from how dis-tricts and teachers identify and buy professional de-velopment to program evaluation to renegotiation ofunion contracts, which govern teacher work commit-ments and effort. This work would require partner-ships on the part of key components of the system,including districts, professional organizations, highereducation institutions, and providers themselves.But, unless this work is done, we must considerwhether the return on investment is appropriate givendemands on schools and teachers today. If not, it’stime to invest precious resources elsewhere. K

REFERENCES

Cohen, David K., and Heather C. Hill. Learning Policy:When State Education Reform Works. New Haven,Conn.: Yale University Press, 2001.

Garet, Michael S., Andrew C. Porter, Laura Desimone,Beatrice F. Birman, and Kwang Suk Yoon. “What MakesProfessional Development Effective? Results from aNational Sample of Teachers.” American EducationalResearch Journal 38, no. 4 (2001): 915-945.

Hertert, Linda. Investing in Teacher ProfessionalDevelopment: A Look at 16 Districts. Denver, Colo.:Education Commission of the States, 1997.

Hill, Heather C. “The Nature and Effects of MiddleSchool Mathematics Teacher Learning Experiences.”Manuscript. Harvard Graduate School of Education,2008.

Hill, Heather C., and Deborah L. Ball. “LearningMathematics for Teaching: Results from California’sMathematics Professional Development Institutes.”Journal of Research in Mathematics Education 35(2004): 330-351.

Hill, Heather C., Merrie Blunk, CharalambosCharalambous, Jennifer Lewis, Geoffrey C. Phelps,Laurie Sleep, and Deborah L. Ball. “MathematicalKnowledge for Teaching and the Mathematical Quality ofInstruction: An Exploratory Study.” Cognition andInstruction (in press)

Hill, Heather C., Brian Rowan, and Deborah L. Ball.“Effects of Teachers’ Mathematical Knowledge forTeaching on Student Achievement.” AmericanEducational Research Journal 42, no. 2 (2005): 371-406.

Horizon Research. The 2000 National Survey of Scienceand Mathematics Education: Compendium of Tables.Chapel Hill, N.C.: Horizon Research, 2002.

Killeen, Kieran M., David H. Monk, and Margaret L.Plecki. “School District Spending on ProfessionalDevelopment: Insights Available from National Data(1992-98).” Journal of Education Finance 28 (Summer2002): 25-50.

Massachusetts Department of Elementary andSecondary Education. www.doe.mass.edu/lawsregs/603cmr2.html?section=05.

Miles, Karen H. “Rethinking District ProfessionalDevelopment Spending to Support SchoolImprovement: Lessons from Comparative SpendingAnalysis.” In School Finance and Teacher Quality:Exploring the Connections, ed. Margaret L. Plecki andDavid H. Monk. Larchmont, N.Y.: Eye on Education,2003.

National Association of State Directors of TeacherEducation and Certification (NASDTEC).Knowledgebase Table E1: Professional DevelopmentDescription. Whitinsville, Mass.: NASDTEC, 2004.

National Center for Education Statistics (NCES). TeacherPreparation and Professional Development: 2000. NCES2001-088. Washington D.C.: U.S. Department ofEducation, 2001.

National Science Foundation. www.nsf.gov/awardsearch. Downloaded and totaled December 21,2007.

Odden, Allan, Sarah Archibald, Mark Fermanich, and H. Alix Gallagher. “A Cost Framework for ProfessionalDevelopment.” Journal of Education Finance 28(Summer 2002): 51-74.

Copyright NoticePhi Delta Kappa International, Inc., holds copyright to this article, whichmay be reproduced or otherwise used only in accordance with U.S. lawgoverning fair use. MULTIPLE copies, in print and electronic formats, maynot be made or distributed without express permission from Phi DeltaKappa International, Inc. All rights reserved.

Note that photographs, artwork, advertising, and other elements to whichPhi Delta Kappa does not hold copyright may have been removed fromthese pages.

All images included with this document are used with permission andmay not be separated from this editoral content or used for any otherpurpose without the express written permission of the copyright holder.

Please fax permission requests to the attention of KAPPAN PermissionsEditor at 812/339-0018 or e-mail permission requests tokappan@pdkintl.org.

For further information, contact:

Phi Delta Kappa International, Inc.408 N. Union St.Bloomington, Indiana 47405-3800812/339-1156 Phone800/766-1156 Tollfree812/339-0018 Fax

http://www.pdkintl.orgFind more articles using PDK’s Publication Archives Search athttp://www.pdkintl.org/search.htm.

k0903hil.pdfHeather C. Hill, Fixing Teacher Professional Development, Phi DeltaKappan, Vol. 90, No. 07, March 2009, pp. 470-477.

File Name and Bibliographic Information