Assessing Misconceptions

8/7/2019 Assessing Misconceptions

1/24


2/24

ImplementingConceptual ChangeTeaching in PrimaryScienceDaniel C. NealeDeborah SmithVirginiaG. JohnsonUniversity of Delaware

The ElementarySchoolJournalVolume 91, Number 2o 1990 by The University of Chicago. All rights reserved.0013-5984/91/9102-0002$01.00

AbstractA studywasmade of the extentto which 8 teach-ers in gradesK-3 were able to implementa 2-week conceptual change science unit in theirclassrooms ollowinga 4-week summer nstitute.The effectsof the instituteprogram,which wasdesigned to help teachersdevelop the subject-matter and pedagogical knowledge needed toteach a unit on lightandshadows,were assessedby analyzing 1)videotapesof lessonstaughtbe-fore and after the program,(2) interviewsthatmeasured students'understandingof light andshadows phenomena before and after the unitwas taught,and (3) teachers'writtenevaluationsof their units. 1 year later,teachers were inter-viewed abouttheircontinuinguseof theunitandconceptualchangeteaching strategiesgenerally.Results,which indicated hatthe 8 teachersweregenerallysuccessfulin implementinga concep-tual changeunit of theirown on lightand shad-ows and in changingstudents'conceptions,arediscussed in relationto case methods and cog-nitive-apprenticeshipmodels of training.Interest in the use of teaching strategies thatfacilitate children's conceptual change hasbeen growing (Eylon & Linn, 1988; Porter& Brophy, 1988). Such strategies focus onstudents' prior conceptions of the subjectmatter under study and seek to provide theconditions under which these preconcep-tions may be elicited and challenged so thatstudents can construct more general, pow-erful, or "correct" conceptions (Anderson &Smith, 1985; Posner, Strike, Hewson, &Gertzog, 1982; Vosniadou & Brewer, 1987).For example, in a science unit on lightand shadows, a boy's conception that shad-ows are projected from the front of the body(DeVries, n.d.; Piaget, 1930) may be elicitedby asking him to predict where his ownshadow will fall as he turns his body in the


3/24

110 THEELEMENTARY CHOOLJOURNALsun. This preconception is challenged as theboy faces sideways to the sun while anotherstudent traces the shadow; to his surprise,the shadow appears at the boy's side, notat his front. Through making predictions,recording observations, giving explana-tions, and discussing with his peers, the boymay revise his conceptions toward the sci-entific view that shadows are produced notby some emanation from the body but byhis body as an object that is blocking raysfrom a light source.The present study explored the extent towhich such conceptual change teachingstrategies could be used successfully by pri-mary teachers in science. Eight teachers (K-3) who had participated in a 4-week sum-mer institute were studied as they imple-mented a conceptual change science unit intheir own classrooms. The extent of theirsuccess was examined on the basis of video-tapes of lessons, teachers' written evalua-tions, interviews with teachers, and subject-matter interviews conducted with childrenbefore and after the unit was taught.Conceptual change teaching strategiesare of particular interest because of wide-spread evidence that students bring theirown ideas to the study of school science,and that these ideas are powerful deter-minants of what students gain from instruc-tion (Driver & Erickson, 1983; Eaton, An-derson, & Smith, 1984; Jones, 1988).Furthermore, conventional science instruc-tion often fails to address or to change mis-conceptions about physical phenomenathat students bring with them to class (Ea-ton et al., 1984; Jones, 1988). Even goodstudents who do well on course exams oftencontinue to hold conceptions that are at var-iance with the scientific theories that theyhave studied (Carey, 1986; Champagne,Klopfer, & Anderson, 1980; Clement, 1983;Minstrell, 1984).A difficult problem for those interestedin conceptual change teaching is how to im-plement such teaching strategies in class-rooms. How can conceptual change teach-ing be designed to fit the realities of schools?

How can teachers be helped to develop theknowledge and skill needed for conceptualchange teaching? And what school condi-tions encourage the implementation ofthese strategies?Some guidance in answering these ques-tions can be obtained from the general lit-erature on school improvement and edu-cational innovation. One lesson that hasbeen learned is that the introduction of anynew teaching practice is a complex andproblematic process (Neale, Bailey, & Ross,1981). Factors associated with successfulimplementation include such things as sup-port from district and school administrators,acceptance by teachers, clear guidelines foruse, timely training in skills needed for im-plementation, ongoing support from peersor coaches, and modifications of organiza-tional constraints (see, e.g., Fullan, Bennett,& Rolheiser-Bennett, 1989; Fullan & Pom-fret, 1977; Griffin, 1983; Hall & Hord,1987). However, even the successful imple-mentation of new practices may be shortlived unless special institutional supportsand incentives remain (Hord &Huling-Aus-tin, 1986; Lieberman, 1986; Stallings & Kra-savage, 1986). Thus, efforts to implementconceptual change teaching strategies mustprovide the kinds of substantial supports toteachers that have been recognized asneeded for the implementation of new prac-tices generally.In addition to the general problems ofimplementing new practices, efforts to in-troduce conceptual change teaching in sci-ence must also overcome some special dif-ficulties. First, such teaching calls for athorough understanding of subject-matterknowledge, including knowledge of chil-dren's likely preconceptions and of repre-sentations of subject matter that studentscan grasp. Second, teachers must know howto identify students' misconceptions andknow how to challenge misconceptions byproviding discrepant events. Such processesplace unusually heavy cognitive demandson teachers because unexpected events arefrequent and call for rapid teacher decisions.NOVEMBER1990


4/24

SCIENCETEACHING 111The teacher must simultaneously be con-cerned with subject matter, students' think-ing, lesson structure, and classroom socialbehavior (Anderson & Smith, 1985; Neale,1985; Smith, 1989). Third, in conceptualchange teaching students are encouraged totest their predictions and explanations withexperiments and observations, then to rep-resent and discuss the results. Lessons ofteninvolve small-group activities and open-ended discussion. These are unfamiliar les-son structures for most teachers, who oftenencounter management problems whenthey use such activities (Anderson & Bar-ufaldi, 1980; Smith & Sendelbach, 1982;Tobin & Fraser, 1987).Some promising attempts to implementconceptual change teaching have been re-ported. After disappointing results withconventional teacher training techniques,Anderson and Smith (1983b) found thatjunior high teachers followed conceptualchange strategies if provided with revisedteachers' guides and overhead transparen-cies to use in eliciting specific student mis-conceptions (see also Smith & Anderson,1984). A promising approach in mathe-matics (Carpenter, Fennema, Peterson,Chiang, & Loef, 1988) employed a summercourse that provided teachers with exten-sive information about the development ofchildren's mathematical concepts. Teachersthen used this pedagogical-content knowl-edge (Shulman, 1986) to alter their teachingso that more emphasis was placed on con-ceptual understanding and children's ownproblem solving.We have been exploring ways to en-courage the use of conceptual change teach-ing strategies by primary science teachers.In the process, we have developed proce-dures that are based on a systematic anal-ysis of the teacher knowledge needed to im-plement conceptual change teaching andthat utilize a cognitive-apprenticeship ap-proach to the acquisition of expertise.Knowledge Underlying TeacherExpertiseWith others who view teaching expertise asa complex cognitive process (Clark & Pe-

terson, 1986), we believe that the design ofteacher training activities should begin withan analysis of the knowledge needed forsuccessful teaching. Although research onteacher cognition remains limited, recent re-search provides useful guidelines for spec-ifying such knowledge.Following Leinhardt and Greeno (1986),we believe that expert teaching requireswell-developed procedural knowledge oflesson structures, including activity sche-mata, rules, routines, and information sche-mata. Furthermore, teachers need appro-priate subject-matter knowledge that theycan use with purpose and flexibility withinlesson structures (Leinhardt & Smith, 1985).Such subject-matter knowledge includessubstantive knowledge, pedagogical-con-tent knowledge, and knowledge of curric-ulum in relation to the subject matter beingtaught (Shulman, 1986). Finally, teachers'beliefs about subject matter and how itshould be learned play an important role(Anderson & Smith, 1985). Thus we de-signed our work with teachers on the basisof a systematic analysis of both pedagogicalknowledge (i.e., lesson structures) and con-tent knowledge (including pedagogical-content knowledge and beliefs) needed byan expert who was using conceptual changestrategies in grades K-3 science classes(Neale, 1985).Cognitive ApprenticeshipIf an analysis of teacher knowledge is takenas a point of departure, then teacher train-ing may be understood as a process throughwhich teachers may construct the knowl-edge underlying expertise. Not only mustteachers acquire various elements of knowl-edge, but they must also combine these ele-ments in flexible, articulated structures thatthey can access with little effort (Berliner,1986). Even though they may be expert inusing conventional teaching strategies,when implementing new strategies teachersmay resemble novices and may have to gothrough a number of stages to acquire fullexpertise (Hall & Hord, 1987).


5/24

112 THEELEMENTARY CHOOLJOURNALA promising way to assist novices in ac-quiring expertise is through the, process ofcognitive apprenticeship (Collins, Brown, &Newman, 1989). As in the historic appren-

tice relationship, a master models expertperformance and assists as novices under-take a series of tasks, graded in difficulty,that require increasing expertise. Thegraded tasks and the assistance representscaffolding that is gradually removed as ap-prentices acquire expertise (Gott, 1988).In the present case, modeling was pro-vided in the form of a demonstration uniton light and shadows, videotapes of an ex-pert teacher, and the use of conceptualchange strategies by an instructor duringtraining. Scaffolding was provided by lim-iting the focus of training to a restricted sub-ject-matter domain, by using small groupsand trial teaching, and by providing coach-ing as teachers planned and taught a con-ceptual change unit in their own class-rooms.It should be noted that, from the per-spective of cognitive apprenticeship, themodeling and scaffolding provided werejust a small step toward expertise. Othersteps that either have been or could betaken by teachers include: (a) reteaching theunit on light and shadows, (b) teachingother prepared science units, (c) using newteaching strategies in other subjects, and (d)developing new conceptual change units.Method

Demonstration UnitPrior to the design of activities to beused with teachers, a demonstration unitwas developed to exemplify the principlesof conceptual change teaching in primaryscience classrooms. Based on a survey ofrelated research, the following characteris-tics of conceptual change teaching wereidentified (see Smith & Johns, 1985).1. Instructions directed oward he con-tradiction of significant, generalschemes or mental representationsthatchildrenhave aboutnaturalphe-

nomenaand towardthe developmentof more powerful,scientificallyaccu-ratemodels.2. Instructional pisodesare linkedcon-ceptuallyto prior essons and to chil-dren's nformalexperiencesduringre-views, presentations, activities,demonstrations, discussions, andsummaries.3. Instructionelicits children'sconcep-tions of naturalphenomena throughtheir predictions and explanationsaboutphenomena.4. Activitiesareprovided n which chil-dren test their predictions,discovercontradictoryevidence, and contrastalternativeexplanationsand concep-tions, includingappropriate cientificmodels. Materialsand activities aresimple and focus on immediate re-sults,which can be variedby the chil-dren.5. Childrenrepresent heirown thinkingin severalmodes (e.g.,writing,speak-ing, drawing, graphing), and theteacherchecks theirunderstanding fthese representations.Childrensharethe results with the teacherand eachother and are encouragedto debateeach other.6. The teacherhelpschildren ummarizeexperiences,highlightsand contrastsalternativeviews, and requestsexpla-nations that examine the evidenceavailableto supporteach. Resultsarethen related to priorand future les-sons, and applicationsin children'slives are generatedand tested.

These characteristics were used to revise anearlier Elementary Science Study unit onlight and shadows (Morrison, 1968). The re-vised unit was designed to address commonmisconceptions that children have aboutlight and shadows (see Table 1) and pro-vided activities and discussions so that chil-dren could construct the following scientificcontent:Light travels in straight lines in all di-rections from the source.Shadows are places where light hasbeen prevented from traveling (e.g., bybeing absorbed, reflected, or refracted).For example, children sometimes thinkthat a person's shadow is a concrete entity

NOVEMBER 1990


6/24

SCIENCETEACHING 113TABLE. Targeted Pedagogical Content Knowledge of Children's Misconceptionsabout Light and Shadows

Source Children's MisconceptionsGuesne (1985) 1. Light is the same as its sources, for example, a light bulb, thesun.Anderson & Smith(1983a); Guesne(1985) 2. Light is bright, static "stuff" that stays in one place or hangs inthe air.DeVries (n.d.);Guesne (1985) 3. Light illuminates or brightens.Piaget & Garcia(1974) 4. Light travels in a straight tunnel from its source; as thedistance between the source and the surface is increased,children expect the diameter of the light on the surface todecrease.DeVries (n.d.); Piaget

(1930) 5. A shadow is a concrete or tangible thing.Piaget (1930) 6. A shadow is a projection or emanation from the body.DeVries (n.d.); Piaget(1930) 7. A shadow is pushed out or forced out by the light.DeVries (n.d.) 8. A shadow's location depends on where the person or object islooking or facing or leaning, or the proximity of the object tothe location.DeVries (n.d.);Guesne (1985) 9. A shadow is a reflection of the object.Inhelder & Piaget(1958); Siegler(1981) 10. The size of the shadow depends only on the size of the object.DeVries (n.d.); Piaget(1930) 11. Two objects will each make a separate shadow even if one isin the shadow of the other; a shadow can be present at nightin the darkness.

that comes out of the "front" of the body(where the face is). In the unit, childrenwent outside to trace their shadows on thesidewalks. They first faced away from thesun, so that the shadow did come out the"front" of their bodies. Then, they faced toone side, so that their ideas that shadowswere projections from the body would becontradicted by the location of their shad-ows. When children returned inside, theclass discussed why the shadow's locationdid not move when they turned their bod-ies. In a subsequent lesson, they moveddolls in various positions around a light,predicted where the shadows would be,then turned the light on and recorded wherethe shadows actually appeared. These trac-ings were also presented and discussed inthe whole-group meeting.

Analysis of ExpertiseThe 13-day unit was taught by an expertteacher to several classes of children ingrades K-3 in a college of education labo-ratory classroom. On one such occasion, ex-tensive records were made of the planningand teaching of the unit, including pre- andposttests of children's thinking, audio re-cordings of planning sessions, videotapes ofeach class session, and recordings of stim-ulated recall interviews with the teacher (fordetails, see Neale, 1985; Smith & Johns,1985).These data were used to describe anddocument the subject-matter and pedagog-ical knowledge required for successfulteaching of the unit. Subject-matter knowl-edge was specified in categories suggestedby Shulman (1986), including substantive,


7/24

114 THEELEMENTARY CHOOLJOURNALsyntactic, and pedagogical-content knowl-edge (Smith & Neale, 1989). Pedagogicalknowledge was analyzed, following Lein-hardt and Greeno (1986), in terms of activ-ity schemata, rules and routines, and infor-mation schemata (Neale, Smith, & Wier,1987). The results of these analyses may beconsidered as an example of case knowl-edge (Shulman, 1986), which was then in-corporated into the program with teachers.The purpose of such case knowledge is tocommunicate principled knowledge aboutteaching, in this instance propositionalknowledge of the principles of conceptualchange teaching, as well as proceduralknowledge for their implementation.

SubjectsTen teachers from grades K-3 (two kin-dergarten, six first grade and two thirdgrade) were recruited to participate in a fed-erally funded institute on primary scienceinstruction. Following announcements thatwere mailed to 47 schools, 10 teachers com-pleted formal application procedures,which included a recommendation from theprincipal or supervisor. Applicants were ob-served while teaching to screen out thosewho might be weak in teaching skills gen-erally and therefore unlikely to benefit fromthe institute. Two teachers had been in-volved with the laboratory classroom in thepreceding academic year. All were experi-enced teachers (5-27 years), and all exceptone were female. For confidentiality, allsubjects were given pseudonyms in this ar-ticle. They were teaching in nine schools inthree districts.Teachers' classes ranged in size from21-28. All classes were heterogeneous incomposition. Ethnic composition rangedfrom 25%-42% minority. Ability levelsranged from 14-90 (Normal Curve Equiv-alent) on the reading section of the Cali-fornia Test of Basic Skills. (CTBS scoreswere not available for the kindergartenclasses.)

Summer InstituteThe first phase of training occurred in a4-week summer institute funded by the Ed-ucation for Economic Security Act, Title II.The instructor was a science educator andexperienced primary teacher who had de-veloped and piloted the demonstration unit.Teachers were paid a stipend through theirdistricts and received four graduate creditsfor the summer program. Before the insti-tute began, we asked teachers if we couldvideotape them teaching a science lesson intheir classrooms. Teachers were also inter-viewed. The interview, which was con-

ducted by the institute instructor, assessedthe teachers' knowledge and beliefs aboutscience and science teaching generally,about the physics of light and shadows,about children's knowledge of light andshadows, and about how one would bestteach children this content. Videotapes andinterview results were examined to identifyeach participant's knowledge and skill withregard to the objectives of training (seeSmith & Neale, 1989, for details).During the first week of the institute,teachers read and discussed research onchildren's misconceptions and on teachingstrategies that facilitate conceptual change.Readings were chosen to provide evidencethat contradicted teachers' views. Thenteachers conducted interviews about lightand shadows with children who were partof a summer science camp associated withthe institute. The purpose was to sensitizeteachers to student conceptions and how toelicit them.Teachers also tested their own knowl-edge of light and shadows in activities thatrevealed their misconceptions and providedopportunities for them to construct moreadequate conceptions. The activities forboth the teachers and the campers were de-signed to model procedures of conceptualchange teaching. Lessons usually beganwith a problem or puzzle to explain; pro-ceeded to predictions, explanations, and de-bate; led to small-group inquiries wheredata were collected and represented byNOVEMBER 1990


8/24

SCIENCETEACHING 115

writing, graphing, or tracing; and endedwith discussion of results in relation to orig-inal ideas.During the second and third weeks,teachers taught light and shadows contentto small groups of summer campers, ages5-8, in the morning. Two teachers were re-sponsible for each group. One taught whilethe other coached, with the teachers re-

versing roles during the third week.In the lessons, teachers generally movedthrough several lesson segments. An open-ing meeting with the group served as a re-view of previous lessons and as a source ofinformation about children's thinking. Thisdiscussion usually moved into the teacher'spresentation of a related puzzle or problemto be solved, with children giving their pre-dictions and explanations for what theythought would happen. Next, childrenmoved into small groups of three or fourfor activities. They manipulated materials toproduce effects and recorded their results insome way (tracing, graphs, writing, pic-tures). Finally, in the concluding meeting,children reported their results and com-pared them across groups, discussed anom-alies, and considered further questions to besolved.Each teacher was videotaped at leastonce, with the videotape becoming the sub-ject of discussion in the afternoon (teacherand coach present; others welcome). After-noons were also used for further explora-tion of light and shadow activities, discus-sion of issues raised by morning sessions,and preparations for the next day's teach-ing. The purpose of the small-group teach-ing was to provide a safe opportunity, withlimited cognitive demands and plenty ofscaffolding, for teachers to begin to con-struct the knowledge needed for teaching aconceptual change unit in their own class-rooms. Teachers made regular entries inlogs to help them reflect on their experi-ences.In the final week of the institute, teach-ers interviewed children again about lightand shadows content to assess the progress

they had made at the camp. Then teachersmet in grade-level teams to discuss plansfor a 2-week unit each would teach in herclassroom during the coming year. Finally,with consultation from the institute instruc-tor, teachers developed preliminary plansfor teaching their units.Although teachers generally addressedthe common misconceptions found at theirgrade levels and designed activities to helpchildren construct scientific ideas aboutlight and shadows as in the demonstrationunit, there were differences across grades intasks that addressed similar ideas. For ex-ample, the kindergarten teachers' plans in-cluded an activity in which children tookwooden blocks outside to construct castlesand traced the outline of their castles' shad-ows on paper laid on the sidewalk next tothem. As children worked, teachersplanned to talk with them about variousopenings-windows and doors-that ap-peared in their shadows and about howthose showed up. Their intention was tofocus children's attention on the directionof the sunlight and where it traveledthrough the castle openings and where itdid not. For kindergarten children, simplynoticing the direction and travel of the lightwas an important step.A third-grade teacher planned an activ-ity in which children also constructed cas-tles, with one important requirement: theywere to figure out a way for light to gothrough three windows in the middle andtwo ends of a wall of the castle when thelight was turned on. This task was designedto provide evidence that contradicted thechildren's common idea that light travelsstraight ahead in a tunnel from the source(instead of diverging in all directions).When children turned on the light, theywould be surprised that the light went notonly through the middle window, but alsoin a thin sliver through the end windows.Their tracings of where the light did gothrough, for the three windows, would pro-vide the evidence for a discussion of thedirection and travel of the light. So, al-


9/24

116 THE ELEMENTARYCHOOLJOURNAL

though children at both grade levels wouldbuild castles and talk about the direction oflight's travel, the two tasks differed in im-portant ways depending on the children'scurrent understanding.

Academic-Year ActivitiesThe second phase of training was a fol-low-up program during the academic year.Monthly teacher meetings were held to con-tinue reading and thinking about the phys-ics of light and shadows, about the princi-ples of conceptual change teaching, andabout the translation of principles into plans

for an instructional unit. Preliminary plansthat had been developed during the sum-mer were revised based on feedback fromthe instructor and peers; the instructor alsomet with each teacher for two 3-hour plan-ning sessions prior to the unit that wastaught. Units were scheduled so that theinstructor could monitor, videotape, andgive feedback on each teacher's unit as itwas being taught. After at least four of the10-unit lessons, the institute instructorcoached the teacher, giving encouragementand suggestions about next steps. Teacherskept journals of their experiences. The pres-ent study investigated how successfully theteachers who were involved in such trainingwere able to implement a conceptualchange unit on light and shadows in theirown classrooms.

Data AnalysisSome of the data about changes inteacher knowledge are reported elsewhere(Smith, 1987, 1989; Smith & Neale, 1989).The present article focuses on how suc-cessful the teachers were in implementingthe conceptual change units in their ownclassrooms and the extent to which teachersimplemented conceptual change strategiesboth in other subjects and in the year thatfollowed. Evidence about implementationcame from an analysis of videotaped les-sons before and after training, from inter-views with students before and after theunit was taught, from teachers' own written

self-evaluations, and from interviews withthe teachers.Analysis of videotapes was undertakenby constructing an innovation configura-

tions checklist (Hall & Hord, 1987, chap. 5),which was used by a trained rater to ratelesson features in the lesson submitted priorto the summer institute and in the fourthand ninth lessons of the 10-lesson unit onlight and shadows that teachers taught intheir classrooms during phase 2 of the pro-gram. These lessons were selected becausethey occurred near the end of each week'steaching and sampled from different por-tions of the unit.The configurations checklist was de-rived from observation procedures used byHollon, Anderson, and Smith (1980) asadapted to our analysis of conceptualchange teaching in primary science (Neale,1987). Three levels of implementation werespecified for each of 31 features of concep-tual change teaching so that lessons couldbe rated on each feature as showing either(1) no implementation, (2) partial imple-mentation, or (3) high implementation.High levels of implementation are definedfor each feature in the Appendix.The 31 features, which were grouped insix categories, included eight lesson seg-ments that had been identified in the anal-ysis of expert performance in the demon-stration unit as well as indicators of thedesirable features of conceptual changeteaching given above.Items on the configurations checklistwere developed through a series of ratingsof sample lessons chosen to represent bothfull and partial implementation. Ratingswere made after inspecting a videotape re-cording and a printed transcript of each les-son. After appropriate training, ratersachieved high interobserver agreement(.94-1.00 for pairs of raters on the 31 fea-tures).Success in implementation was also ex-amined by measuring children's concep-tions of light and shadow phenomena bothbefore and after teachers taught their ver-NOVEMBER 1990


10/24

SCIENCETEACHING 117sions of the light and shadows unit. A ran-dom sample of one-half of the children ineach class, stratified by gender and ability(high, medium, and low according toteacher rating), was interviewed before andafter the units. One third-grade class andone first-grade class were selected for in-tensive study, and all children in theseclasses were interviewed. Children were in-terviewed individually in a quiet setting bytrained interviewers who were former ele-mentary teachers. Children were shownmaterials (e.g., a doll and light) in differentrelationships. Then they were asked to pre-dict what would happen if various manip-ulations were made and to explain whythey thought that would happen.The kindergarten children were given abriefer version of the interview, withoutmore difficult questions on topics such asmerged shadows and the divergence oflight. First and third graders received alonger interview covering a broader rangeof topics.

For example, children were asked,"When it's night time, and you're in bedand all the lights are turned off, do you havea shadow?" to probe for the misconceptionthat a shadow is a concrete entity that existsindependently of the light. They were alsoasked about situations like the following. Adoll is placed behind a tensor light with ashade, but facing the back of the light. Thechild was asked, "If we turned this light on,could you make a shadow with the dollhere?" (pointing to the doll behind thelight). This question probed for the child'smisconception that the shadow is a part ofthe doll and projects from the front, in thedirection the doll is facing.Children's responses were scored bytwo trained research assistants (graduatestudents in measurement and evaluation),both for accuracy of predictions and forquality of explanations, and a total scorewas obtained to represent each child's un-derstanding of unit concepts.In the examples of questions givenabove, suppose a child said that her shadow

was present at night and returned inside herbody or was present but could not be seenbecause there was no light. Her answerwould be scored as inaccurate in prediction(shadows are not present at night) and in-adequate in explanation (shadows requirelight both to be made and to be seen; theyare not like a tree, which is there whetherlight is available to see it or not). In contrast,the answer of a child who responded thather shadow was not there at night, in thedarkness, because there was no light to beblocked to make it, would be scored as ac-curate and as providing an adequate expla-nation. Her reason refers explicitly to theaction of light in making shadows and pro-vides enough information to distinguish itfrom the common misconception that thelight makes the shadow by pushing it outof the body.In the second example, a child might re-spond that the doll behind the light willmake a shadow when the light is turned on,because she is facing that way. This answeris inaccurate (no shadow will appear), andthe explanation suggests that the shadowlocation is controlled by the position of thedoll's face, a common misconception. An-other child might respond that there wouldbe no shadow when the light is turned on,because the light is not going that way(pointing to the doll). This answer wouldbe scored as accurate and as referring ad-equately to the direction of the light and therelationship between the light and the ob-ject in making the shadow.Teachers' self-evaluations of implemen-tation were summarized from written re-ports made following the completion of theunit. Teachers were asked to make a finalevaluation of their success in teaching theirown version of the light and shadows unitand of what they had learned as a result ofparticipating in the institute program.Teachers' opinions were also sought 1year later in a structured interview. Trainedinterviewers (two graduate students in ed-ucation) met individually with teachers inJune to ask about their use of the unit on


11/24

118 THE ELEMENTARYCHOOLJOURNALlight and shadows and of conceptualchange principles generally during the pre-vious year (year 2). Also, teachers wereasked to give a retrospective evaluation oftheir training program and to summarizetheir plans for using conceptual changeteaching during the upcoming year (year 3).ResultsIn this section, we present the results of theconfigurations checklist ratings of eightteachers' lessons and briefly discuss threeindividual cases. Next, we report results ofthe interviews conducted with the childrenby trained interviewers before and aftereach unit was taught. Finally, we summa-rize teachers' own evaluations of the unitand the summer institute.

Configurations Checklist RatingsRatings from the configurations check-list are summarized in Table 2 for the eightteachers who implemented the unit. Aver-age ratings over the 31 features of lessons

rated after training ranged from 1.8-2.8, in-dicating moderate to high implementation.The mean rating for all rated lessons aftertraining was 2.4 on the three-point ratingscale.For the six teachers who had submittedtapes of their science teaching prior to theinstitute, comparisons were made between

ratings of lessons taught before and aftertraining. (Two teachers joined late and didnot submit tapes of prior teaching.)Each of the six teachers made substantialprogress in implementing conceptualchange teaching, as measured by the meanratings for all 31 features combined. Despitethe small sample size, differences betweenmean ratings before and after trainingproved to be statistically significant for thegroup, as indicated by the Wilcoxon signed-ranks test (z = -2.20; p = .028).Increased implementation occurred inevery category of lesson features. Based onthe mean ratings of features within each cat-egory, these differences between ratings be-fore and after training were all significant(p < .05), as indicated by the Wilcoxonsigned-ranks test.A more detailed picture of the ratings isgiven in Table 3, which reports mean rat-ings of all six teachers on each of the 31features both before and after the summerinstitute. Although, on average, teachersshowed at least partial implementation ofall the rated features, they appeared to bemore successful with some features thanwith others. Teachers were more successfulwith lesson segments such as "review,""development," and "investigations" thanthey were with "summary." Likewise,teachers implemented content features such

TABLE. Mean Ratings for Each Teacher on All 31 Features of ConceptualChange Teaching before and after Training

After TrainingBeforeTeacher Grade Training Lesson 4 Lesson 9

Betsy 3 1.5 2.8 2.8Nan 3 1.8 2.7 2.8Carol K 1.3 1.8 2.0Denise 1 1.5 2.5 2.6Gail 1 1.8 2.3 2.4Helen K 1.7 2.4 2.5Six teachers 1.6 2.4 2.5Ann 1 N.A. 2.4 2.2Barbara 1 N.A. 1.8 2.2Eight teachers 2.3 2.4

NoTE.-1 = no implementation; 3 = high implementation; N.A. = not available.NOVEMBER1990


12/24

SCIENCETEACHING 119TABLE. Mean Ratings on Each Category of Configurations Checklist forSix Teachers Combined, before and after Training

After TrainingBeforeCategory Training Lesson 4 Lesson 9Lesson segments:1. Introduction 1.5 1.5 2.22. Review 1.3 2.7 2.73. Development (focus) 1.0 2.0 2.84. Development (elicit) 1.8 2.8 2.55. Investigations 2.2 2.5 2.86. Representation 1.2 3.0 1.87. Discussion 1.3 2.3 2.28. Summary 1.5 1.7 1.7Content:1. Conceptual emphasis 1.0 2.2 2.32. Accuracy 2.5 3.0 2.8

3. Scientific emphasis 1.3 2.7 2.74. Appropriate representations 1.2 1.3 2.35. Ties concept across lesson(s) 1.2 2.2 3.0Teacher role:1. Elicits students' conceptions 1.7 3.0 2.82. Provides discrepant events 1.5 3.0 2.23. Facilitates students' constructions of newconcepts 1.7 2.2 2.5Student role:1. Predict, explain 1.5 2.8 2.82. Test predictions 2.0 2.5 2.53. Represent results 1.2 2.8 2.54. Describe, discuss results 1.5 2.3 2.05. Apply everyday experience 1.0 1.2 1.36. Cooperate in small groups 1.8 2.3 2.5Activities/materials:1. Permit salient effects 2.0 2.5 3.02. Related to lesson concept 2.0 2.5 3.03. Include discrepant event 1.2 2.2 2.5Management:1. Workspace, materials ready 2.2 2.8 2.82. Rules, routines discussed 2.3 2.8 2.73. Rules, routines in place 1.8 2.5 2.74. Monitoring, consequences 2.0 2.7 2.75. Students take responsibility 1.8 1.8 2.26. Students on task 1.8 2.2 2.8

NoTE.-1 = no implementation; 3 = high implementation.as "accuracy" and "scientific emphasis"better than "conceptual emphasis" and"developmental appropriateness of exam-ples."With respect to teachers' roles, teacherswere better at eliciting and identifying stu-dents' misconceptions and in presentingdiscrepant events than in helping studentsto construct new knowledge. With respectto students' roles, particularly high imple-mentation was observed in getting studentsto make predictions, but low implementa-tion was noted on applications to everyday

experience. Teachers had good success, asmeasured by these ratings, in providing ap-propriate activities and materials and inmanaging their classrooms.Although group averages are useful injudging the general effects of training, theyconceal important differences among indi-vidual teachers. Therefore, three individualcases are presented briefly to illustrate thevariety and complexity of training effects.

Betsy-Dramatic SuccessBetsy's (grade 3) ratings showed a dra-matic change in lessons taught before and


13/24

120 THEELEMENTARYCHOOLJOURNALafter intervention. In both lessons 4 and 9of her light and shadows unit, she receivedthe highest possible rating on 27 of 31 fea-tures. By contrast, before training, her les-son received only one such rating.Betsy was one of those who came to theprogram quite dissatisfied with her pastpractices in science. Before training, her sci-ence lesson revealed serious managementproblems, especially in small-group activi-ties, and her subject-matter knowledge oflight and shadows was weak. She eagerlyembraced the conceptual change orienta-tion and used opportunities in the programto construct the various kinds of knowledgeneeded to implement the light and shadowsunit. In particular, by the end of the aca-demic year, she was one of the most knowl-edgeable in physics. At the conclusion ofher unit, she wrote, "I would tell anyteacher who comes to me for advice aboutthis unit that if I can do it, so can she, butshe will have to be really interested andwilling to work hard. I will tell her that shecan get ready for a unique opportunity tosee kids working and learning in smallgroups. She'll be amazed at what kids cando in the right setting and that teaching forconceptual change does work."

Nan-Substantial ImplementationNan was another third-grade teacherwho was successful in implementing thelight and shadows unit. In contrast to Betsy,Nan was relatively confident about scienceand had a good background in the physicsrelated to the content of the unit. In inter-views prior to training, Nan expressed astrong content-mastery orientation towardscience teaching and an emphasis on chil-dren's understanding of science content.Ratings of her pretraining science lessonshowed strong classroom-management fea-tures during whole-group instruction andintermediate implementation on manyother features.After the summer institute, Nan's les-sons showed high implementation of al-most all the rated features. Her confidence

in her subject-matter knowledge enabledNan to generate demonstrations and ex-amples to address unexpected events andissues in lessons. For example, she encour-aged children to discuss whether a handheld flat on a surface could have a shadow.In her report after teaching the light andshadows unit, Nan commented, "In gen-eral, I felt fantastic about the unit, the kids'work in it, what they learned, and what Ilearned.... There was so much going on,so much to be pleased with, and so muchto notice because it was all new. I felt likethe kids were really learning some scienceconcepts and without the encumbrance ofa text and the reading and vocabulary prob-lems that accompany it."Despite the success Nan experienced,conceptual change teaching proved to be adifficult transition that was far from com-plete. As reported in an intensive case study(Smith, 1989), Nan experienced difficulty ingiving up previously successful and well-practiced teaching and management rou-tines that were more compatible with con-tent-mastery approaches than with concep-tual change strategies.

Carol-Limited ProgressCarol, a kindergarten teacher, was oneof those who, despite interest, dedication,and hard work, had difficulty implementingthe light and shadows unit. Although herlessons after the institute received higherratings than her pretraining lesson, she im-plemented no more than seven of the 31features at high levels.At the beginning of the program, Carol'sscience teaching emphasized "hands-on"activities in which children made individualdiscoveries. Her science background wasweak, especially in physical science. (WhenCarol found out that physics was involvedin the unit, she nearly dropped out of theprogram.) As indicated in her presummerscience lesson, management of science ac-tivities was a weakness.

Although Carol increased her knowl-edge in each of the areas of training, sheNOVEMBER 1990


14/24

SCIENCETEACHING 121was handicapped in implementing the lightand shadows unit by student misbehavior.Carol was unsuccessful at implementing thenecessary classroom rules and routines;consequently, other features of conceptualchange were difficult or impossible to im-plement. As Carol herself put it, "I thought(to myself) that I was a pretty good teacheras far as my management went-until youopened my eyes. Now I see my flaws, andsince last summer, have tried to be moreconscious of the skills necessary to developa good unit. ... My problem with disciplinewas because I'm just too easy."

In summary, the configurations checklistratings indicated substantial success amongthe eight teachers in implementing concep-tual change teaching, although progresswas varied and idiosyncratic. Teachersstarted with varied levels of knowledge andprogressed in different ways.Interviews with ChildrenResults of interviews with children be-fore and after teachers had implemented theunit were available for six of the eight teach-ers discussed above. (One teacher borrowedher posttest results from us and misplacedthem; interviewers failed to obtain pretestresults on another teacher's class.) For eachgrade level, significant gains in mean scoresfrom pre- to posttest were observed (see Ta-ble 4). For the first- and third-grade classes,who had the same interview, a grade Xability X gender analysis of covariance withthe pretest as a covariate showed significantdifferences (p < .05) for ability group and

gender, but no significant grade effect andno significant interactions. Adjusted meanscores were greater for boys than for girlsand for higher-ability groups than for thelow-ability groups. Thus teachers at eachgrade level appeared to be successful inbringing about conceptual change in stu-dents, as well as in implementing featuresof conceptual change teaching strategies.We are currently conducting a more de-tailed study of the changes in children'sthinking.

Teachers' Evaluations of UnitsA compelling supplement to the check-list ratings and interviews with students areteachers' comments written at the conclu-sion of their units. They record the ups anddowns of teaching plus the strains of tryingsomething new and different. Teacherswere asked in making their reports to eval-uate the success of their unit on light andshadows, to comment on the factors that ledto successes and failures, to judge the effects

of the unit on students' thinking, and toevaluate the extent to which they had beenable to incorporate in their unit the featuresof conceptual change teaching as presentedduring the summer institute.As noted earlier, two of the original 10teachers failed to implement the unit. Oneof these, Fran (grade 1), dropped from theprogram after the conclusion of the summerinstitute. She had been asked by her districtto participate in a second summer workshopand reported that her inability to continuewas due to other heavy commitments in the

TABLE. Mean Total Scores and Standard Deviations for Children'sInterviews before and after Unit on Light and ShadowsPretest Posttest

Grade N M SD M SD p ValueThird 34 30.4 6.1 41.1 7.0 .0001aFirst 38 26.4 6.3 35.6 6.5 .0001aKindergarten 12 16.0 3.2 23.3 2.7 .0001b

at test.bWilcoxon Matched-Pairs Signed-Ranks Test.


15/24

122 THE ELEMENTARY CHOOLJOURNALfall. The other, Pat (grade 1), was havingdifficult management problems with herclass during the year and decided, with theconcurrence of the institute instructor, notto risk implementation.The eight teachers who implementedthe unit all judged it to be a success. Com-ments ranged from Nan's "I felt fantasticabout the unit" to Carol's "Much was ac-complished by myself and the children."Without exception, they reported specificevidence that convinced them that chil-dren's thinking had changed. Denise's(grade 1) comments were typical. "Ican def-initely see that many of the kids made prog-ress in understanding light and shadows. Ihave observed this during group lessonsand activity time. I've observed kids helpingone another, explaining and predicting. Thekids have documented their knowledge onMagic Slate disk. They have also writtenstories and drawn pictures of what theylearned in the various lessons. It was sat-isfying to see kids like Rob, Ada, Amy, andSam become dissatisfied with misconcep-tions about light and shadows after dis-cussing and working on activities."Some teachers had conducted discus-sions about light and shadows with theirstudents at the beginning and end of theunit. These discussions were frequentlymentioned as convincing evidence of stu-dent progress. However, teachers appearedto be even more convinced by children'sresponses to daily activities and by theirdaily creations. Some, like Barbara (grade1), were eloquent about children's re-sponses. "But the results-ah, what results!Kids who don't talk, talking; kids who don'tdo, doing; kids who are loners, joining. Alldone by them! I didn't say talk, do, join--they did. I didn't say learn, understand, dis-cover-they did! It felt like opening a doorin invitation and having guests rush in! Itfelt the way that the books in college alwayssaid that teaching would feel-warm, re-warding, active. It felt alive!"Others, like Helen (kindergarten) em-phasized daily assignments. "The children

were given an assignment at the end of eachlesson-draw, color, cut, paste, relate whathappened in class today. We used this asan evaluation tool and a bit of evidence. Itis really wonderful! The [children's] bookswere impressive and showed a great dealof understanding. One of the most inter-esting conclusions-shadows don't havecolor of their own-was beautifully illus-trated by a child who colored the object (anorange) and repeated the same objectshadow-in black! Looking through thebooks would be enough to show thegrowth."

Teachers were sure that they could con-vince others of students' growth in under-standing. Carol (kindergarten), for example,wrote, "My children made progress-verydefinitely. I can prove it because Ikept noteson each kid; their art work over the 2-weekperiod and their class participation are themajor ways. I could show others this sothey'd know, or show the tapes."All eight teachers believed that they hadbeen able to employ the conceptual changeteaching strategies that had been presented.They made special mention of the strategiesfor eliciting and challenging misconcep-tions, for asking students to represent theirthinking, for selecting activities and mate-rials, and especially for managing the class-room. However, every teacher reported dif-ficult days or specific problems in using thestrategies.The most common problem mentionedwas time-never enough. Teachers hadtrouble estimating the time needed for les-sons, so they were often caught short. An-other common problem was management,especially the distraction of individual chil-dren who had trouble cooperating in smallgroups.A final problem was that the new strat-egies seemed strange, awkward, or difficultto coordinate. Teachers would forget to usea strategy or inadvertently fall into oldways. Looking back, Carol recognized that"I did often forget to ask 'Why?' and havethe children give their reason. This wasNOVEMBER 1990


16/24

SCIENCETEACHING 123

lacking on my part, but I got better towardthe end-it is something for me to contin-uously strive for and include in the future."Betsy commented on the difficulty of co-ordination:

Some of the factors hat seemed to makesomelessons moresuccessful hanotherswere when a sense of timing developedandI knew how it felt to have thingsfallin place.Once that sense was intact,eachday was a constanthoning and refiningprocess. Being able to keep the lessonmoving was, for me, an integral partofsuccess.It was difficultat firstnot to bedistractedby all of the different acets ofthe subjectmatter hat the studentskeptraising in the discussions. Keepingthe"BigIdea"for the day in mindand mak-ing the activity simple but appropriatewithexcellentmanagementweresome ofthe keys to betterlessons.One strong impression that comes fromreading teachers' comments is that imple-menting new practices such as conceptualchange teaching strategies, when viewedfrom the perspective of the teacher, is a ma-jor undertaking. At the end of the first year,no teacher had mastered the new strategiescompletely despite spending considerabletime and energy. Yet the progress made,even when moderate as viewed from theoutside, was a dramatic transformation asviewed from the inside. Barbara (grade 1)is a case in point:Teaching cience n thisway was freeing!Away fromthe way I had been teachingscience-I would show and tell, ask stu-dents questionsabout what I had beenteaching(not to test for understanding,but to see who had bothered to pay at-tention),andpass out a ditto on the sub-ject for the students to complete. Thiswasn't fulfilling-but it's the way I sawothersdoing it, and since I hadn'tfounda sourceof informationon otherways, Ifell intothe rut.Ididn'teven knowwhereto begin-but now I'll never go back.Did this unit feel good?In ways I canhardlyexplain! t feltgood to see kids soalive. It felt good to push myself, to seewhat I could do, what I could handle.

Fallingon my face,figuringout whathadgone wrong,andre-doing he lessonsuc-cessfullywas terrific!This whole processof reevaluatingmy philosophyof teach-ing, my goals for myself within thisprofessionis somethingthathardlyany-one gets a chanceto do. And that is whatis happening to me-I've been exposedto ideas which contrastso harshlywithwhat I've been doing, that I've had torethink through a lot of the bull I'vetaken and given and done.Final Interviews with TeachersAlthough one of the main purposes ofthe study was to determine the extent to

which teachers had successfully imple-mented the unit on light and shadows dur-ing phase 2 of their training, we also wereinterested in whether or not they would usethe unit subsequently and whether theywould use principles of conceptual changeteaching in other aspects of their work. Al-though we were unable to observe theirlater teaching, we did conduct extensive in-terviews with the teachers 1 full year afterthe year of classroom implementation. Inthis final interview, trained interviewersasked teachers a variety of questions con-cerning the kinds of subject-matter knowl-edge and their beliefs about science and sci-ence teaching; these data will be reportedelsewhere. Also, interviewers asked teach-ers about their use of conceptual changeteaching during the academic year just past(year 2 of the project) and about their plansfor using conceptual change teaching dur-ing the upcoming year (year 3 of the proj-ect). Two of the original 10 teachers werenot available for interviews.Out of the eight teachers participating inthe final interviews, six indicated that theyhad taught the light and shadows unit dur-ing the previous year (year 2). All eightmentioned specific ways in which concep-tual change teaching principles had influ-enced their teaching outside of the light andshadows unit, and all eight had definiteplans to use the unit in the upcoming ac-ademic year (year 3). Five of the eight teach-ers mentioned clear plans for using concep-


17/24

124 THE ELEMENTARYCHOOLJOURNALtual change principles outside the unit itself.Although no teacher developed and taughtanother conceptual change unit in science(or in any other subject), each teacher men-tioned significant effects. Forexample, Den-ise reported changing her whole manage-ment system based on what she hadlearned. Carol said she had learned to waitlonger for student responses. Barbara re-placed math workbooks with activitiestwice a week. Ann said she began to inter-relate science, math, and language arts.The six teachers who had implementedthe unit a second time reported good suc-cess, although four of them thought thingshad gone better the first time. Betsy, for ex-ample, tried the unit early in the year andreported that her third-grade class had notyet mastered classroom rules and routines.She planned to schedule the unit later inthe year next time around. Nan reportedsome slippage because, without coaching,there was a loss of "pressure to try to doeverything exactly the way it should bedone." Denise was less satisfied because ashortened time schedule in the afternooncut off unit activities. Gail reported usingonly "bits and pieces" of the unit. Although"kids liked it," she did not see strong evi-dence of conceptual change.Two teachers reported better success thesecond time around. Carol, for example,said the unit was "more successful" andthat "the kids loved it." Barbara was ex-tremely satisfied with her use of the unit inan entirely new teaching assignment.The two teachers who reported not us-ing the unit gave as their reason competingpressures of other projects at their schools.Ann's school had new programs in science,math, and reading. Pat had moved to a newschool with "too much going on at gradelevel to get organized." It is interesting thatboth teachers planned to implement theunit in the following year.

Beyond the specific changes reported byteachers, a general theme emerged. Alleight teachers said, in one way or another,that the greatest changes occurred in the

new attention that they gave to students'thinking. Denise put it this way: "I learnedhow to listen more carefully to children,think about what they're saying, let themgo on their predictions, to be patient, to planmy lessons according to what the classknows and not what I think should betaught, and not take for granted that theyalready know something." Betsy noticed inher math teaching that, "I used to just getup and do it; it never occurred to me to findout what they already knew. I now realizethe amount of misconceptions that childrencome to school with, and it was somethingthat I never thought about before.... It'samazing to me-the depth of children'sthinking; it is not always "correct,"but theyjust feel so strongly or think so stronglyabout their ideas, and maybe I just nevergave it credibility."Even Pat, who had yet to implement herunit, mentioned trying "to allow more timeto get ideas from children" and "to askmore appropriate questions to find out whatthey are really thinking."

Asked about their plans for the comingyear, all eight teachers said that they in-tended to teach the light and shadows unit.However, only Betsy, Nan, and Denisementioned plans to extend such teaching toother science units. No one mentioned spe-cific plans to develop conceptual changeunits in other subjects. The stage of teach-ers' developing expertise was best charac-terized by their responses to a questionabout whether they would be interested inteaching a new unit on heat and tempera-ture, if it were available. Nan spoke for themost confident in saying, "Definitely.Where is it?" Ann spoke for the majority insaying, "I would try." Although they werecomfortable with the light and shadowscontent, most teachers believed that theywould need help with subject-matterknowledge and familiarity with specific ac-tivities and materials before taking on a newconceptual change unit.Teachers' Evaluations of the ProgramDuring the final interview, the eightteachers were asked to rate how comfort-NOVEMBER 1990


18/24

SCIENCETEACHING 125able they had become with various activi-ties related to conceptual change teaching.All eight reported feeling much more com-fortable with understanding children'sthinking and with planning and teachingconceptual change lessons. With one or twoexceptions, they all felt more comfortable inunderstanding the physics of light andshadows, choosing curriculum and mate-rials, organizing and managing smallgroups, and conducting discussions.When asked to rate components of theprogram in terms of helpfulness, the unan-imous top rating was given to the planningsessions with the expert coach. In order ofhelpfulness, teachers rated other compo-nents as follows: readings about children'sthinking, firstweek together in the summer,teaching small groups of children in sum-mer camp, having coaching while teaching,watching videotapes of one's own teaching,readings about conceptual change teaching,writing in log or journal, writing curriculumin grade-level team, coaching others, andmonthly seminars. All of the activities wereconsidered to be at least "somewhat help-ful," and all except the last three were givenaverage ratings of "most helpful."Finally, teachers were asked to rate ac-tivities that might be helpful to them interms of their continuing growth in teachingscience to young children. Teachers gavetop priority to workshops with materialsand activities on science. Next, in order ofhelpfulness, were readings about children'sthinking, revisions in state or district cur-riculum, new units on other topics, ongoingdiscussion with other teachers, the princi-pal's support and feedback, regular coach-ing, and developing curriculum oneself. Allof the items were considered to be eitherhelpful or very helpful.Summary and DiscussionThe major result of the study was to doc-ument that eight of 10 primary teacherswho attended the summer institute wereable to implement a conceptual change uniton light and shadows in their classrooms.

Although implementation varied amongparticipants and among features of theteaching strategies, the results indicate thatteachers in the primary grades can use con-ceptual change strategies successfully.Granted, the teachers in this study had un-usually strong support, including a dem-onstration unit and coaching; nevertheless,the results are highly encouraging com-pared to the dismal record of past efforts toimprove elementary science teaching (e.g.,Stake & Easley, 1978). Furthermore, basedon interviews with the teachers 1 year later,apparently all eight of those interviewed,even without coaching, continued to use thestrategies they had learned. As always, suchreports must be interpreted with caution.There did seem to be some slippage forteachers during the second year. Yet be-cause their practice is supported by changesin their underlying knowledge, we are op-timistic about their chances for continuingto practice and learn from this new way ofteaching.

In some cases the training seemed to beaccompanied by less observable but no lessdramatic shifts in teachers' orientationstoward teaching science and other subjects.The most general shift was a new emphasisgiven to the role of students' conceptions.Almost all teachers mentioned new sensi-tivity and new willingness to listen to chil-dren's ideas as a central outcome of theirexperience. This shift is similar to the resultdescribed by Carpenter et al. (1988) withfirst-grade teachers who were taught aboutchildren's thinking in mathematics, eventhough the intervention strategies were dif-ferent.

Despite the positive effects of the train-ing, results also indicate that teachers hadto make strenuous efforts to construct theknowledge required for this type of teach-ing. An interesting aspect of the researchprogram has been the attempt to documentteachers' views of the changes. Subject-matter knowledge, including knowledgeabout students' misconceptions and how tochallenge them, did not come easily for


19/24

126 THE ELEMENTARYCHOOLJOURNALmost. Some teachers were still workinghard on their understanding of the physicscontent at the end of the first year. Knowl-edge about how to involve children in in-quiry and about how to get young childrento represent their new knowledge was like-wise difficult to develop. If teachers in theproject are typical, a feeling of inadequacyin these aspects of science may be one rea-son why teachers rely on fact-oriented, text-book instruction or offer loosely structured,"hands on" science activities that often con-tribute little to children's conceptual devel-opment.

As illustrated in the present project,teachers need not only a vision of some-thing better but also various kinds ofknowledge that they can use to bring thevision to life. As suggested by the cognitive-apprenticeship model, just like the knowl-edge of any novice craftsman or profes-sional, such knowledge appears to be ac-quired by degrees and graduallycoordinated for flexible use. Propositionalknowledge about students' ideas and aboutteaching strategies will be insufficient forchange in teachers' practice unless it is aug-mented by procedural knowledge of how toplan for and implement teaching strategieswithin an effective lesson structure. Alsoneeded is the conditional knowledge aboutwhen to use such strategies appropriately(Berliner, 1989; Brophy, 1988).To acquire the propositional, proce-dural, and conditional knowledge, teachersneed scaffolding for support. Among thesupports teachers in this project valuedwere those that gave them specific infor-mation about children's thinking, as well asmaterials and activities they could use inteaching. Also valued were discussions withother teachers and especially interactionwith an expert coach while developingteaching plans. In addition, teachers appre-ciated opportunities to teach small groupsand to receive feedback on their videotapedteaching. These evaluations are consistentwith the assumptions of cognitive-appren-ticeship training, in which novices are given

help and guidance by an expert as they un-dertake a series of tasks that are graded indifficulty. In this case, implementing a uniton light and shadows after training with ademonstration unit appeared to be a sig-nificant, though difficult, step toward ex-pertise in conceptual change teaching forthese primary teachers.Focusing on a single topic, light andshadows, was a significant strength and alsoa limitation of the program. By limiting thescope of subject-matter knowledge, primaryteachers could construct the subject-matterknowledge needed for successful imple-mentation, even while they were acquiringknowledge of teaching and managementstrategies. With the possible exception ofNan, who had good subject-matter knowl-edge to begin with, the strategy appearedto be an essential part of the scaffoldingprovided. Other teachers were perilouslylow on subject-matter knowledge, and theyknew it. Giving them a realistic, positiveopportunity to acquire such knowledge inconnection with a teaching unit appearedto be a highly successful feature of the train-ing.The limitation of the strategy is, ofcourse, that the changes in science teachingoccur primarily, perhaps only, within thesingle unit. Based on our experience,changes were primarily within the unit onlight and shadows, although teachers didreport influences on their beliefs and teach-ing practices generally. We believe that suchlimitations are not a flaw in the trainingstrategy as much as a recognition of howdifficult it is for primary teachers to revo-lutionize their science teaching and to im-prove their subject-matter understandingsignificantly.Even the present project, with a limitedfocus on a single unit, with a demonstrationmodel and coaching, with careful scaffold-ing tailored to individual needs-even suchtraining did not produce teachers who wereexperts. Although many of the teacherswith whom we worked were recommendedas highly competent in traditional areas ofNOVEMBER 1990


20/24

SCIENCETEACHING 127their teaching, most were relative novicesin teaching for conceptual change in sciencelessons. In terms of Berliner's (1989) dis-cussion of stages of expertise, these teachersappear to have acquired new patterns ofthinking and teaching, but these patternshave not yet become fluid and coordinated.Instead, teachers appear to have taken adramatic, yet strictly limited, step towardexpertise in conceptual change teaching.Before conceptual change strategies arefully implemented throughout their scienceprogram, these primary teachers wouldneed to have much additional support of asimilar kind.Is it realistic to think that the kinds ofsupport given to teachers in this programcan be made available to teachers generally?Perhaps not in the intensive format de-scribed here. We are developing variationson the program and working with smallgroups of teachers in our area. Conceptualchange units in earth science and weightand balance are the basis for both summercourses and in-service programs during theyear. In the latter case, teachers bring theirstudents to the curriculum laboratory class-room where both the demonstration teacherand the regular class teacher share the plan-ning and teaching. In this setting, the scaf-folding is somewhat different from the sum-mer institute program, but the modelingand coaching process is basically the same.We believe that existing resources now de-voted to in-service teacher education couldbe redirected along these lines to have asignificant effect on many more teachers. Assome teachers become more knowledgeableand expert in their science teaching, for ex-ample, they might become demonstrationteachers and coaches for others.As far as the provision of continuingsupports for such teaching in schools goes,we are less sanguine. If science is taught 2days per week between 2:00 and 2:30 P.M.,if principals and parents care mainly aboutreading and math achievement, and if ap-propriate equipment and supplies are notavailable, even teachers who are enthu-

siastic about this type of teaching will ex-perience "slippage." Furthermore, if teach-ers are not given the opportunity, time, andsupport to construct the substantive and pe-dagogical-content knowledge underlyingthe curriculum they are asked to teach, thenno doubt we will continue to see short-cutteaching strategies that emphasize superfi-cial "coverage" of science textbook topicsor loosely structured, hands-on activitiesthat often add little to students' understand-ing of science. We must recognize that suchproblems do not represent defects in ourtraining strategies; rather, they are failuresto use and support the strategies we have.What implications do we see for pre-service teacher education programs? This isa more difficult question. Surely, the explo-ration of such a question would requiremore information than we have on theknowledge and beliefs that preserviceteachers bring to their teacher training. LikeShulman (1986), we see the need to em-phasize the development of content knowl-edge, especially pedagogical contentknowledge. With elementary teachers, suchan emphasis presents difficult problems be-cause of the wide range of content theycould be asked to teach in science as wellas in other subjects. Furthermore, collegescience courses commonly include teachingstrategies and forms of representations thatare far removed from those likely to be ef-fective in primary classrooms. Specialcourses that teach substantive knowledge inconjunction with both pedagogical-contentknowledge and pedagogical knowledge ap-pear to be desirable. Based on our experi-ence, such courses need to have practicalcomponents so that students may constructthe subject-matter knowledge in relation tolesson structure and so that the knowledgehas some flexibility in use. However, wewould expect, along with Berliner (1989),that preservice teachers would have evenmore difficulty than did our experiencedteachers in coordinating the various formsof knowledge. As novices, they lack thefund of episodic memories, practical knowl-


21/24

128 THE ELEMENTARYCHOOLJOURNALedge, and frameworks for structuring andinterpreting classroom events that charac-terize the thinking of expert teachers.Therefore, objectives for preservice teachereducation would need to be modest, andthe scaffolding would need to be plentiful.Training would emphasize learning to learnabout science teaching and would rely onthe intensive study of a small number oftopics rather than on broad coverage ofmany topics. For this purpose, case knowl-edge, such as that contained in the dem-onstration unit used in this study, shouldprove extremely useful.Now that we have documented changesin teachers' knowledge and teaching prac-tice, we plan to examine our data in detailedcase studies of the teachers involved to seemore specifically how knowledge changeswere related to the classroom science teach-ing we observed (e.g., Smith, 1989). As To-bin and Fraser (1987, p. 213) put it, "The

challenge that faces researchers in the wakeof the Exemplary Practice in Science andMathematics Education Study is to identifyhow teachers can construct knowledgeabout content and teaching so that theirteaching performance improves .... Wemust learn more about the role of differentknowledge forms in teaching."How does teachers' knowledge guidetheir planning and affect what they noticein classroom lessons? What role does it playin their decision making while teaching? Towhat extent does one kind of knowledge(e.g., pedagogical knowledge of lessonstructures) influence teachers' ability to useother kinds of knowledge (e.g., knowledgeof teaching strategies)? These and other re-lated questions remain to be investigated inmore detail as we continue our attempts tounderstand the role of teachers' knowledgein primary-grade science teaching.

AppendixFeatures of Conceptual Change Teaching Rated on Configurations Checklist

Lesson Segments1. Introduction Teachercomments on lesson's topic and activi-ties with reference o conceptualcontent.2. Review (if not firstlesson) Teacherasks students fordescriptionof pre-vious lessons' conceptualcontentandprobestheirunderstanding.3. Lessonfocus Teacherclearlypresentsa single focus to aconceptor problem.4. Lessondevelopment Teacherelicitsstudents' deas aboutconcept,gets predictionsand explanations,andprobesunderstanding.5. Investigations/activities Studentsmanipulatematerials o test newideas. Teacherprobes,questions,checksstudents'understanding.6. Representation Studentsrepresentresults of activities n a waythat allows teacher'sprobeof understand-ing (e.g., writing,graphs,drawings) n thewhole-classdiscussion.7. Discussion Studentsshareresultsand discuss evidencewhile teacherencouragesdiscussionof evi-dence. Focusis on studentthinking.8. Summary Teacherencouragesstudents to summarizefindingsand connectthem to the concep-tual contentof previousand future essons.

NOVEMBER1990


22/24

SCIENCETEACHING 129Content Features

1. Teacher emphasizes students' conceptual un-derstanding and conceptual change in pre-vious ideas.2. Teacher's own presentation is conceptuallyaccurate.3. Lesson focuses on an important scientificmodel or explanation.4. Teacher uses examples, analogies, and met-aphors that are conceptually and develop-mentally appropriate.5. Teacher ties conceptual content across lessonsegments and/or across lessons.Teacher Role

1. Teacher elicits and diagnoses students' con-ceptions of content under study and asksthem to predict and explain events.2. Teacher contradicts students' thinking byproviding discrepant events, encouraging de-bate, and/or challenging students' ideas.3. Teacher facilitates students' construction ofscience conceptions by contrasting ideas, en-couraging discussion, asking for applications,and/or modeling cognitive processes.Student Role1. Students verbalize ideas by predicting, ex-plaining, and discussing ideas relating to con-

cept under study.2. Students work with materials to investigateideas and test predictions.3. Students represent (e.g., writing, drawing,graphing) results of investigations for latersharing and discussion with teacher and otherstudents.4. Students describe, demonstrate, and discussresults with teacher and other students in re-lation to own ideas.5. Students apply concepts to everyday eventsand explain how concepts are related to ev-eryday events.6. Students work and discuss ideas in small co-operative groups or pairs, as well as in whole-group meetings.

Activities and Materials1. Materials and activities are familiar and allowstudents to produce immediate, observable,salient, and varied effects.2. Activities are clearly tied to concept in lesson.3. Activities include discrepant events thatclearly contradict students' thinking.Management Features1. Teacher has arranged classroom so that workspace is available, materials are ready and ac-cessible, traffic flows smoothly.

2. Teacher has established positive expectationsfor cooperative work, involves students indiscussions of needed rules and routines, andof reasons why these are important for theirwork (or this appears to have been done pre-viously).3. Teacher teaches, explains, or provides feed-back on needed rules and routines (or theyseem to be in place).4. Teacher consistently monitors students' be-havior, acknowledges appropriate behavior,and applies agreed-on consequences.5. Students take responsibility for maintainingcooperative work environment.6. Students generally follow rules and routines,monitor their own behavior, remind friendsof rules and routines, are generally on task.

NoteThis article is adapted from a paper pre-sented at the annual meeting of the AmericanEducational Research Association, San Fran-cisco, March 1989. The research was supportedby an Education for Economic Security Act TitleII grant; the opinions expressed do not neces-sarily reflect those of the funding agency. We

wish to thank the teachers involved in the proj-ect.

ReferencesAnderson, C., & Barufaldi,J. (1980). Research onelementarychool cience eaching:A studyus-ing short-term utcomemeasuresOccasionalPaper No. 37). East Lansing: Michigan StateUniversity, Institute for Research on Teach-ing.Anderson, C., & Smith, E. (1983a). Children'sconceptions f lightand color:Developing heconcept of unseen rays. Paper presented at theannual meeting of the American EducationalResearch Association, Montreal.Anderson, C., & Smith, E. (1983b). Teacher be-haviorassociatedwithconceptualearning.Pa-per presented at the annual meeting of theAmerican Educational Research Association,Montreal.Anderson, C., & Smith, E. (1985). Teaching sci-

ence. In V. Koehler (Ed.), The educator'shandbook: research erspectivepp.80-111).New York: Longman.Berliner, D. (1989). Implications of studies of ex-pertise in pedagogy for teacher education


23/24

130 THEELEMENTARY CHOOLJOURNALand evaluation. In New directionsfor teacherassessment, Proceedings of the 1988 Educa-tional Testing Service Invitational Confer-ence (pp. 39-68). Princeton, NJ: EducationalTesting Service.

Brophy, J.(1988). Educating teachers about man-aging classrooms and students. TeachingandTeacherEducation, (1), 1-18.Carey, S. (1986). Cognitive science and scienceeducation. American Psychologist, 41(10),1123-1130.Carpenter, T. P., Fennema, E., Peterson, P. L.,Chiang, C., & Loef, M. (1988, April). Usingknowledge f children'smathematicshinkingin classroomeaching:An experimentaltudy.Paper presented at the annual meeting of theAmerican Educational Research Association,

New Orleans.Champagne, A., Klopfer, L., & Anderson, J.(1980). Factors influencing the learning ofclassical mechanics. American Journal ofPhysics, 48(12), 1074-1079.Clark, C., & Peterson, P. (1986). Teachers'thought processes. In M. Wittrock (Ed.),Handbookf research n teaching 3d ed., pp.255-296). Chicago: Rand McNally.Clement, J. (1983). Students' alternative concep-tions in mechanics: A coherent system of pre-conceptions?Proceedings f theInternationalSeminar,MisconceptionsnScienceandMath-ematics (pp. 310-315). Ithaca, NY: CornellUniversity, Department of Education.Collins, A., Brown, J.,&Newman, J.(1989). Cog-nitive apprenticeship: Teaching the crafts ofreading, writing and mathematics. In L. B.Resnick Ed.),Knowing,earning, ndinstruc-tion (pp. 453-494). Hillsdale, NJ: Erlbaum.DeVries, R. (n.d.). Children's conceptions ofshadow phenomena. Unpublished manu-script. Houston: University of Houston.Driver, R., & Erickson, G. (1983). The study ofstudents' onceptualrameworksnscience.Pa-per presented at the annual meeting of theAmerican Educational Research Association,Montreal.Eaton, J., Anderson, C., & Smith, E. (1984). Stu-dents' misconceptions interfere with sciencelearning: Case studies of fifth-grade students.ElementarychoolJournal, 4(4),365-379.Eylon, B., & Linn, M. C. (1988). Learning andinstruction: An examination of four researchperspectives in science education. Review ofEducationalResearch,58(3), 251-301.

Fullan, M., Bennett, B., & Rolheiser-Bennett, C.(1989).Linking lassroom ndschool mprove-ment. Paper presented at the annual meetingof the American Educational Research As-sociation, San Francisco.

Fullan, M., & Pomfret, A. (1977). Research oncurriculum and instruction implementation.Reviewof EducationalResearch, 7(1), 335-397.Gott, S. P. (1988). Apprenticeship instruction forreal-world tasks: The coordination of pro-cedures, mental models, and strategies. In E.Z. Rothkopf (Ed.), Review of research in ed-ucation (Vol. 15, pp. 97-169). Washington,DC: American Educational Research Asso-ciation.Griffin, G. (1983). Implications of research forstaff development programs. ElementarySchool Journal,83(4), 414-426.Guesne, E. (1985). Light. In R. Driver, E. Guesne,& A. Tiberghien (Eds.), Children'sideas in sci-ence (pp. 10-31). Philadelphia: Open Uni-versity Press.Hall, G., & Hord, S. (1987). Change in schools:Facilitating the process. Albany: State Uni-versity of New York.Hollon, R., Anderson, C., & Smith, E. (1980). Asystem orobserving ndanalyzing lementaryschoolscienceteaching:A user'smanual.EastLansing: Michigan State University, Institutefor Research on Teaching.Hord, S., & Huling-Austin, L. (1986). Effectivecurriculum implementation: Some promisingnew insights. ElementarySchool Journal,87(1), 97-115.Inhelder, B., & Piaget, J. (1958). The growth oflogical hinking romchildhoodoadolescence.New York: Basic Books.Jones, L. V. (1988). School achievement trendsin mathematics and science, and what canbe done to improve them. In E. Z. Rothkopf(Ed.),Reviewofresearchn educationVol.15,pp. 307-341). Washington, DC: AmericanEducational Research Association.Leinhardt, G., & Greeno, J. (1986). The cognitiveskill of teaching. Journal of EducationalPsy-chology, 78(2), 75-95.Leinhardt, G., & Smith, D. A. (1985). Expertisein mathematics instruction: Subject matterknowledge.Journal fEducationalPsychology,77, 241-271.Lieberman, A. (Ed.). (1986). Rethinkingschool im-provement. New York: Teachers CollegePress.Minstrell, J. (1984). Teaching for the develop-ment of understanding of ideas: Forces onmoving objects. In C. Anderson (Ed.), Ob-servingscienceclassrooms:Observingcienceperspectivesrom researchand practice(pp.55-74). Columbus, OH: ERIC/SMEAC.Morrison, P. (1968). Light and shadows. NewYork: McGraw-Hill, Webster Division.Neale, D. C. (1985, October). Cognitive demandsof teaching or conceptual hange:Theexpert

NOVEMBER1990


24/24

SCIENCETEACHING 131teacher.Paper presented at the annual meet-ing of the Northeastern Educational Re-search Association, Kerhonkson, NY.Neale, D. C. (1987, April). Primaryteachers'cur-rentpracticesand neededexpertise n sciencelessons. Paper presented at the annual meet-ing of the National Association for Researchin Science Teaching, Washington, DC.Neale, D. C., Bailey, W. J., & Ross, B. E. (1981).Strategiesorschool mprovement.oston:Al-lyn & Bacon.Neale, D. C., Smith, D., & Wier, E. (1987, April).Teacherhinkingnelementarycience nstruc-tion. Paper presented at the annual meetingof the American Educational Research As-sociation, Washington, DC.Piaget,J.(1930).Thechild'sconception fphysicalcausality. Totowa, NJ: Littlefield, Adams.Piaget, J., & Garcia, R. (1974). Understandingcausality. New York: Norton.Porter, A. C., & Brophy, J. (1988). Synthesis ofresearch on good teaching: Insights from thework of the Institute for Research on Teach-ing. EducationalLeadership,45(8), 74-85.Posner, G., Strike, K., Hewson, P., & Gertzog,W. (1982). Accommodation of a scientificconception: Toward a theory of conceptualchange. Science Education,66(2), 211-227.Shulman, L. (1986). Those who understand:Knowledge growth in teaching. EducationalResearcher,15(2), 4-14.Siegler, R. (1981). Developmental sequenceswithin and between concepts. Monographs oftheSociety orResearchn ChildDevelopment,46 (2, Serial No. 189).Smith, D. (1987). Primary teachers' misconcep-tions about light and shadows. In J. Novak(Ed.),Proceedings f the Second nternationalSeminar:Misconceptions and EducationalStrategies n Science and Mathematics pp.461-476). Ithaca, NY: Cornell University,Department of Education.

Smith, D. (1989). The roleof teacherknowledgein teachingconceptual hange cience essons.Unpublished doctoral dissertation, Univer-sity of Delaware, Newark.Smith, D., & Johns, D. (1985, October). Teachingfor conceptual hange:Rationaleand descrip-tion of activitiesin a first-gradecienceunit.Paper presented at the annual meeting of theNortheastern Educational Research Associ-ation, Kerhonkson, NY.Smith, D., & Neale, D. C. (1989). The construc-tion of subject matter knowledge in primaryscience teaching. Teaching and TeacherEdu-cation, 5(1), 1-20.Smith, E., & Anderson, C. (1984). The planningand teaching ntermediateciencestudy:Finalreport.East Lansing: Michigan State Univer-sity, Institute for Research on Teaching.Smith, E., & Sendelbach, N. (1982). The pro-gramme, the plans and the activities of theclassroom: The demands of activity-basedscience. In J. K. Olson (Ed.), Innovation in thescience curriculum: lassroom nowledge ndcurriculum change (pp. 72-106). London:Croom-Helm.Stake, R., & Easley, J. (1978). Case studies in sci-ence education.Urbana: University of Illinois,Center for Instructional Research and Cur-riculum Evaluation.

Stallings, J.,& Krasavage, E. (1986). Program im-plementation and student achievement in afour-year Madeline Hunter Follow-ThroughProject. ElementarySchool Journal,87(2),117-138.Tobin, K., & Fraser,B. (1987). Exemplarypracticein science and mathematics ducation.Perth,Western Australia: Curtin University ofTechnology, Science and Mathematics Edu-cation Centre.Vosniadou, S., & Brewer, W. F. (1987). Theoriesof knowledge restructuring in development.ReviewofEducational esearch, 7(1),51-67.

Documents

Assessing Misconceptions