7/30/2019 Studying and Using Multiple Documents in History Effects of Discipline Expertise http://slidepdf.com/reader/full/studying-and-using-multiple-documents-in-history-effects-of-discipline-expertise 1/23 Studying and Using Multiple Documents in History: Effects of Discipline Expertise Author(s): Jean-François Rouet, Monik Favart, M. Anne Britt and Charles A. Perfetti Source: Cognition and Instruction, Vol. 15, No. 1 (1997), pp. 85-106 Published by: Taylor & Francis, Ltd. Stable URL: http://www.jstor.org/stable/3233756 . Accessed: 10/04/2013 23:53 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. . Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to Cognition and Instruction. http://www.jstor.org This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PM All use subject to JSTOR Terms and Conditions
23
Embed
Studying and Using Multiple Documents in History Effects of Discipline Expertise
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
7/30/2019 Studying and Using Multiple Documents in History Effects of Discipline Expertise
Studying and Using Multiple Documents in History: Effects of Discipline ExpertiseAuthor(s): Jean-François Rouet, Monik Favart, M. Anne Britt and Charles A. PerfettiSource: Cognition and Instruction, Vol. 15, No. 1 (1997), pp. 85-106Published by: Taylor & Francis, Ltd.
Stable URL: http://www.jstor.org/stable/3233756 .
Accessed: 10/04/2013 23:53
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp
.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
.
Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to Cognition and
Instruction.
http://www.jstor.org
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions
tion. Furthermore, istoryinvolves permanent econstruction ndreinterpretationof pastevents, and our knowledgeof most history topics entails some degreeof
with incompleteness,inconsistencies,or bias. Thus, skilled learningof historyincludes the ability to integrate,to complete, and to challenge the knowledge
conveyedthroughmultiplehistoricaldocuments.What s the nature f thecognitiveskills involvedin usinghistoricaldocuments?Do they developas partof students'
general iteracy,ordo they belong exclusivelyto thedisciplineof history?We examine this questionwith a studyof history disciplinespecialists,asking
how suchdisciplinespecialistsdifferfromacademically omparable onspecialistsin their
approacho a
particularistorical
problem.A
keyidea in ourexamination
is the distinctionbetweendomainanddisciplineexpertise.Academic rainingeads
to expertise in specific domains.For a historian,this domainexpertisemay be
19th-centuryAmerica,post-WorldWarI Europe,and so on. Ourassumption s
thatdiscipline specializationprovidesan overlapping butnot identical)expertisein the tools of the discipline. These include its characteristicproblem types,information ources,methodsof analysis,and standards f evidence. To capturethe character f disciplineknowledge, tis necessary ogainsome controlof domain
or topicknowledge.
Domainknowledge,not discipline knowledge,has been the objectof studyinmost researchon text understanding.Thus,for example,Voss and his colleagues
(Spilich,Vesonder,Chiesi,& Voss, 1979;Voss, Vesonder,& Spilich, 1980) have
shown that individualswith more initialknowledgeof the game of baseballhave
a bettermemoryfor verbalaccountsof a particular ame.Recht and Leslie (1988)found similarresultswith 12-year-oldreaders.They showed that the influenceof
domainknowledge s independentromreadingabilityandtypeof posttest(verbalvs. pictorial).Online studies haveshown thatdomainexpertsapplymore efficient
text comprehension trategies(Afflerbach,1989) becausethey can activate more
elaboraterepresentations f the situationorproblem Means& Voss, 1985; Voss,Greene,Post,&Penner,1983).Domainexpertsalsoapplymoreflexible ruleswhen
evaluating he importanceof statements n a scientific text (Dee Lucas& Larkin,
1986, 1988). Thus, in additionto content areaknowledge, domainexpertsmay
possessmoreelaboratemodelsof the discoursestructuresypicalof theirdiscipline
(Dillon, 1991).
History expertisehas been studiedas ablendof domainanddisciplineknOwl-edge. Wineburg(1991) observed whatwe call disciplineexpertisewhen he sug-
gested thathistoryexpertsuse specific strategieswhen studyingmultipledocu-
ments. In Wineburg's study, high school seniors and historyexperts (graduatestudents and faculty) were asked to study a set of paintings and documents
representing he battleof Lexington.Expertsshowed three distinctstrategiesnot
sharedby novices. First,they focusedon the sourceinformationattached o each
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions
among different types of evidence and can interpret nformationfrom amongsources.
Other tudies,however,provideevidencethat nexpert tudentshave somesense
of documentsas evidence.Perfetti,Britt,andGeorgi(1995) observed hatcollege-level studentsareable to detect bias inhistoricalaccounts,although heypay little
attention o primary vidence. Rouet, Britt,Mason,and Perfetti 1996) found that
college-level studentswithlittleexperience n historycandistinguishamongtypesof historical sources. When asked to evaluate several documentsrelated to a
correspondence)more useful andtrustworthyhanopinionatedhistoricalessays.Furthermore,he studentssometimesreferred o thetypeof documentperse (e.g.,"Itis an official document")n order ojustify theirpreference.The studentsalso
the separationbetweendiscipline specializationand domainknowledge,however.
As we noted earlier,the two overlapbut are not identical.Whetherdisciplinespecializationbringscharacteristictrategiesbeyondassociateddomainknowledgeis an interestingand difficult issue. It is interestingbecauseevidenceof discipline
knowledge would instantiatea kindof generalized earning,one thatcomes from
trainingin the methodsof a discipline. It is a difficult issue becausediscipline
knowledgecannot be separated asily fromdomainknowledge.Inthis study,our
strategy is to comparestudents of comparableacademic level-two groups of
graduatestudents-who differprimarily n the type of training hey have chosen
to take.Althoughgraduate tudentscannotbeconsidered xperts nthe usualsense,
they have been extensively exposed to problemsanddiscoursestructuresypicalof thedisciplineof history(e.g., studyingcontroversial ventsbased onconflicting
interpretations).Werefer o theseexpertstudentsasdisciplinespecialists.Students
fromsome otherdiscipline-in ourcase,psychology-are referred o asdiscipline
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions
of seven documents(or study set) used in a previousstudy (Rouet, Britt,et al.,
1996).The
compositionof each
studyset was as follows:
1. Two historianessays: Historianessays (HEs)explicitly supportedone side
of thecontroversy e.g., they argued hat he interventionwas or was notjustified).The two essays gave oppositeviewpoints. They were both secondhandaccounts
and werewrittena long timeafter he events.
2. Twoparticipantaccounts:Participantccounts PAs)were writtenby peoplewho directlytook partin the events. These accounts were written at the time or
some time afterthe events.Like the HEs, the PAs were opinionated, hatis, they
explicitly supportedone side of thecontroversy.3. Twoofficialdocuments:These documentswerewrittenbefore or at the time
of the events andactuallyplayeda role in the controversy e.g., the 1846 United
States-ColombiaTreaty tating heirrightsandduties nPanama).The two official
documents ODs) were mentioned n the HEsand PAs in supportof their nterpre-tations.
4. One textbook-likeexcerpt:This passage, writtenby the researchers,was
basedon severalaccountsandgavea basicdescriptionof the events. The textbook-
like excerpt(TB) wasmostlyfactualanddidnot referexplicitlyto thecontroversy.
The study set was presentedon a computerscreen using a simple hypertext
system(fora more detailedpresentation f thesystem,see Britt,Rouet,& Perfetti,
1996). Based on studies of hypertextusability (Rouet, 1992), the system was
designedin order o be easily masteredby novice users.Thesystemfirstpresentedachronologyof the mainfactsanda list of the availabledocuments.Thedocuments
were listed in a fixed order:HEs, PAs, TB, and ODs. Each document was
pantscould select and read the documentsone by one, in any order,and as manytimes as they needed. Each documentselectionresulted n the presentation f the
extended source (e.g., "Excerptfrom U.S. PresidentT. Roosevelt's speech to
Congress[January , 1904].The authorarguesthatthe U.S. military ntervention
in the Panamanian evolutionwasjustified.").The participants ould thendecide
either to readthe documentor to go back to thechronologyand document ist.
Procedure
The experimentwas runindividuallyandinvolved two sessions. In Session 1, we
was madeof 14short-answer uestions e.g., "Whocontrols hePanamaCanal?").In addition, he
participantswere asked to locate nine countrieson a blankworld
map.2Theparticipantshen reada backgroundext about hehistoryof Panamaand
answereda seriesof comprehensionandopinion questions.Session 2 was run from a few minutesto a few hours afterSession 1.First,the
participantswere trainedon the hypertext ystem using an unrelated et of docu-
ments.Theexperimentermadesurethat heparticipantswerecomfortablewith the
hypertext nterface.Participantswere thenasked to studythe two controversies.
The procedurewas identicalfor each controversyandincluded three main tasks:
studyperiod,essay writing,and documentevaluation.
Study period. Participantswere given 15 min to studythe list of facts and
the set of documentsusingthe hypertext ystem. They were asked to select each
documentat least once in order o read he extendedsource(however,theydid not
haveto read he actualdocument).Theparticipantswere told that heycould select
thedocuments n anyorderand as manytimes as theyliked.Theywere advised to
do so wisely becauseof the time constraint.The participantswere free to stop as
soon astheyfelt theyhadaninformedopinionabout hecontroversy,provided hat
theyhad selectedeachextendedsource at leastonce. Duringthe studyperiod,the
controversy tatementwas accessibleusinga "read hequestion"button ocatedoneverypage of thehypertext.
Essay writing. Afterthe studyphase,theparticipantswere given 10 minto
write aone-pagedraftessayexpressing heiropinionon thecontroversy.Thestudyset was no longer availableto the participants ecausewe wanted them to draw
argumentsrommemory.However,theparticipantsweregiven a reminder tatingthe controversyanda list of the extendedsources as well as the importantnames
mentionedin the study set. The participantswere also told that they were not
expectedto quotefromthedocuments.
Document evaluation. Theparticipantsweregiven a sheetlistingthe seven
extended sources. They were asked to rank the documentsaccordingto their
usefulness(i.e., "totheextenttheyhelpedyoubuildupan informedopinionduringthe study").The number7 was to be given to the most usefuldocument,6 to the
next most useful, and so on to 1 for the least useful document.In addition,the
participantswere askedtojustify brieflythe rankgiven to each document.Then,
theparticipantswereagain giventhe sevenextendedsourcesand askedto rank he
2The materialsare availableon request o the firstauthor.
3The timelimitsusedin Session2 were setaftera pilotstudy.Theywere used as baselines,andtheywereappliedwith some flexibilityso thatall the participantsouldcompletethetasks.
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions
selection, students'evaluationsof the documentswith respectto usefulness and
trustworthiness, nd students'expressionsof and support or an opinion in their
essays.
Did Noviceand ExpertStudentsDiffernTheirInitial
Knowledge?
The historyspecialists outperformedhe novices on the general historytest (65%vs. 35%, respectively),t(17) = 4.19, p < .001. However, the specialistsand the
novices bothperformedpoorlyon the Panamahistorytest (33%and23%correct,
respectively) and did not significantlydiffer from each other, t(17) = 1.41, ns.
Finally,thespecialistsandthe novices differedonly marginally rom eachotheron
the maptest (63%vs. 82%,respectively), (17) = 1.95,p < .10. Thus,as expected,the specialistshad more generalknowledge in historythan the novices, but the
novices and the specialistshadequallylittlepreviousknowledgeof the historyofPanama.
How DidStudentsManagethe StudyPeriod?
Therewas a lot of variation n students'managementof the studyperiod,but no
consistentdifferenceacrossgroups. Studytimeranged rom4 to 20 min,'with an
averageof 11.3 min, and did not significantlydiffer acrossgroupsor problems.
Students' order of selection of documentsalso varied considerably. Only twostudents(one in each group)selectedthe documents n theirorderof presentation
40n five occasions,studentsrequestedandweregranted xtrastudytimepastthe 15-min ime limit.
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions
on the computerscreen,andtheydid so for bothproblems.All the other students
selectedthe documents n differentorders.Therewere a few consistentpatterns
n
students' selection orders.Fifty-threepercentof the students(6 novices and 4
specialists)selected the TB in thefirst,second,or thirdpositionforbothproblems.An additional25%(4 novices and 1 specialist)did so either for Problem1 or for
Problem2. In Problem2, 68% of the studentsselected both ODs amongthe first
Whatmakes a document useful? Table2 presentshefrequencyf each
categoryof justificationfor each type of document.A frequency analysison the
pool of justifications,collapsedacrossdocumentsandproblems,showedthatthe
distributionof justificationsdifferedsignificantlyacrossgroups,X2(3,N= 352) =21.78, p < .01. The novice studentsmostly justified their rankingsusing the
documentcontentas a criterion 60%on average).Incontrast, he specialistsused
a roughly equal proportion f eachcategoryof justifications(i.e., content,source,andtask).
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions
alsofoundthese two documentsvery untrustworthyomparedwiththespecialists.Thatthe documentauthorswere involved in the events may have led the novice
students o distrust he documentcontent.This may also explainwhy the novices
did not findPAs useful (see Table 1).
Overall,the students n bothgroupswere sensitive to the opinionexpressed n
thedocuments.Neutraldocuments, hat s, documents hatdid not takeany explicitstanceon the controversy TB, ODs), were foundmoretrustworthyhanopinion-
ateddocuments i.e., PAs, HEs).
Whatmakes a documenttrustworthy? Table4 presentshefrequencyfeach categoryof justificationfor eachtype of document.The patternof justifica-tionswas verydifferent romthe one foundwithusefulness.First, hereweremanymore justificationsbased on source information 44.5% vs. 23%). In addition,novices andspecialistsdid not differsignificantly n theirtypes of justifications,
X2(3,N= 311) = 7.07, ns.
As shownin Table4, source ustificationswere mostfrequent or PAs and ODs.Source was also the second most important ategoryfor the TB and HEs.
The datapresented n Table4 suggestthat source characteristicse.g., who the
authoris) were salient when evaluating trustworthiness,whereas task charac-
teristics-which the specialistsdid use to evaluate usefulness-were less salient.
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions
ple, whether he authorhadaninterestat stakeor whether he document s official.
Therewere somediscrepanciesn the evaluationof documentswithinacategory.Studentsjustified their preferencefor OD2 (1846 Treaty)over OD1 (military
correspondence)n Problem1 by stating hatOD1 could be biased(as one student
put it, "I do not trustthe so-called official sources of the military").Similarly,studentsjustified their preferencefor Senator Carmack(PA1) over President
Roosevelt (PA2) by statingthe position of the latter(e.g., "One knows that a
President s notgoing
todisparage
a decision takenby
himself and hisleaders").
Conversely,a few studentsquestionedSenatorCarmack' background e.g., "takes
a con stancebut is he a political opponent o Roosevelt?"). n thiscase, the lack of
detailed informationaboutthe authorcreated some uncertaintyabout the docu-
ment' trustworthiness.
Summaryof evaluationdata. ODswere oundmostuseful nd rustworthyby both novices and specialists.The specialists,however, found the PAs more
useful than the novices did, which they justified using source characteristics.Novices evaluateddocumentusefulnessusingmostlycontent nformation,whereas
the specialistsused a varietyof criteria.Both groupsfocused on sources when
evaluating trustworthiness.This finding suggests that students may focus on
differentdocument eaturesas a functionof taskrequirements.
How DidNovices and SpecialistsComposeTheirEssays?
Students'essayswerefirstanalyzedusingaframework esigned n an earlier tudy(Rouet,Marron,Mason,& Perfetti,1993).Thisframework imsatidentifying he
claims andtypesof arguments sedby students n theiressays.Eachessaywasfirst
decomposed ntoa set of argumentationnits,whichgenerally ncludeda smallset
of semanticpropositions.Then,each unit was assignedto an argument ategory:
How did students express an opinion? Opinionsin each essay were as-sessed througha subcategoryof evaluationstatements alled claims. Withrespectto thepresenceandtypeof claimstatements, heessays collected in this studyfell
into threebroadcategories:
1. Full claim: Some of the students ook anexplicitstance on thecontroversy,forexample,"TheU.S. interventionwas notjustifiedbecausetherewere no signsof a revolution n Panamabeforethemarines anded" N5, P1).5
2. Restrictedclaim: Some of the students ook a moderatepositionor did not
personallyassumethe expressed position, for example, "If one considers that arevolutionwould havethreatenedransitn theIsthmus, henonemayhypocriticallyconcludethatthe interventionwasjustified"(E4, Pl).
3. No claim: Some of the studentsdid nottake a stance on thecontroversybut
ratherpresenteda generalevaluationof the controversy, or example, "TheU.S.
interventionn Panamagave birth o two opposedinterpretations"E5, P1).
The frequencyof full claims,restricted laims,andno claimessays in the two
groupsarepresented n Table 5.
A frequencyanalysison the pool of 38 essays showed that the distributionof
claims was significantlydifferent n thetwo groups,X2(2,N = 38) = 6.44, p < .02.
whereas 69% of the specialists' essays began with eithera contextual statement
(e.g., "OnNovember2, 1903,arevolutionbrokeoutinPanama") r acommentaryon theproblem tatement e.g.,'"The ontroversybasicallyrests on anambiguityof
the 1846 Treaty").Thus,thespecialists ended o introduceordiscuss theproblem
spacebeforegoing into furtherdetails.The distribution f initial statements claimvs. otherstatement)was significantlydifferentacrossgroups,X2(1,N= 38) = 3.89,
The specialistsincludedmore contextualstatements hanthe novice students:
1.75 versus 0.68 per essay, respectively,F(1, 17) = 5.94, p < .05. Contextualstatementswere also morefrequent n Problem2 than in Problem 1: 1.75 versus
0.68 per essay on average,respectively,F(1, 17) = 4.62, p < .05. There was no
herevaluationof U.S. involvement nthe revolution:"Thearrival f aU.S.warship,even with a small troop on the eve of the revolution and the recognition of
[Panama's] ndependence hreedays later nvolves the U.S. in thisevent"(E4).In summary,contextualstatementswerenotonly morefrequent n the special-
ists' essays butalso more focusedandmore elaborate.
Sourcing heuristic. Thesourcingheuristic onsists of payingclose attention
to the sourceof information e.g., theauthorof adocument).This source informa-
tion is vitalin anessay taskif students ntend o cite thesourceof their nformation(i.e., thespecificdocument hatcontained heargument rprovidedsupport orit).
Consequently,we considered hepresenceof references o documentsas evidence
of asourcingproductionheuristic.Onereferencewas scoredeach timeaparticipantmentioneda documentor the nameof an author.
We did notfindanydifferenceacrossgroups n thenumber rtypeof references.
Ninety-two percentof theessays containedat least one reference,and theaveragenumberof referencesperessaywas2.05. Bothgroupsreferredmainly o ODs(62%of thereferences).Thus,based on thestudyof references,we didnotfindevidence
for sourcingdifferencesbetween novices andspecialists.
Corroborationeuristic. Thecorroborationeuristic onsistsof checkingseveralsourcesof information gainsteachother.Theessaysincludedseveral ypesof corroborations s shownin Table6.
Some studentsmerely expressedthe existence of two opposingviews or inter-
pretations,whereasothersexplicitly noted the type of connection betweentwo or
more documents. Some mentioned the positive connectionsbetween consistent
sources,thatis, they tended to groupthe documentsexpressingsimilar views onthe controversy.One studentmentioneda negativeconnection; hatis, he pointedout thediscrepancyor conflictbetweentwo sources(thesetwo cases werefrom the
samestudent,one in eachproblem).Finally,somestudentsmadegeneral referencesto a groupof sourcesbackinganassertion.
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions
frequentin both groups. Some students used a different means to corroborate
information.They pointed out that some evidence was missing or that moreevidence of a certain ypewouldhavehelped.InProblem1,forinstance,aspecialistnoted: "Whether r not the U.S. really respected he Bidlack-MallarinoTreaty s
a difficultquestion,since we do not havethis treaty n its entirety" E7).Sometimestheneedforadditional nformationwas stressedeven morespecifi-
cally, as in this specialist's essay:
The problem s to know to whatextent the Panamanianswere readyat the
time of the revolution to pay thatprice for theirindependence.We would
need to know to whatextentthe Panamanianswere againstthe Colombiangovernment,whether heywereready o sacrificeeverything nexchange or
a free country; hus,to know thePanamanians' eactions o the Treaty,and
whether n November1903they thought heyhad made a "gooddeal."(E5,
P2)
Inthis case, the specialistidentifieda potentialsourcethat was not representedn
the studyset. Although he two examples ust presented ame fromspecialists,we
also foundsimilarstatementsn two novices' essays.Thus,ourdatado notsupport
the conclusionthatcorroboratingources or identifyingmissing information s askill specific to experthistorystudents.
Complexreasoningheuristics. Ina fewessays,participantsombinedev-
eral of the heuristicsdescribedearlier.For instance,a specialist discussed two
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions
context of the events, whereas the novices used various types of contextual
information.Some specialistsalso used more sophisticatedreasoningstrategies,
combiningthesourcing,corroboration, ndcontextualization euristics n a single
threadof argumentation.
DISCUSSION
We examinedthe effects of students'expertisein historyon theirevaluationand
use of documentinformation.The studentswere all graduatestudents,and theydifferedonly in theirdisciplineof specialization i.e., psychologyvs. history).We
asked the students to study a set of documents, to evaluate each document's
usefulness and trustworthiness, nd to write a draftopinion essay based on the
documents.Threequestions were addressed: a) Do discipline specialistsdiffer
from novices in theirstudy strategies?(b) Do discipline specialistsuse specificknowledgewhenevaluatingdocument nformation? nd(c) Do discipline special-ists use documents n a specificway whenwritinganessay?
Theanalysisof onlinestudyparameters id notshowany significantdifferences.
Most students constructedidiosyncratic study sequences, and they selectively
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions
reviewed some of the documents.Novices andspecialists ended o examineeither
the TB or ODs first. These documentswere generally found most useful and
trustworthy. n addition,most studentscheckedthe problemstatementat specifictimes, eitherat the beginningor at the end of the study sequence,a monitoring
strategyobserved n othertextcomprehension tudies(Wright,1993).Theresults
suggest that,when asked to examinea set of documents,most graduate tudents
will use elaboratestudy strategies,whatever heir areaof specialization.General
literacyor academictrainingmay account for graduate tudents' elaboratestudy
strategies Anderson& Armbruster, 984).However,the analysisof online study
goingback andforthbetweentwodocuments.Moreover,different tudystrategiesmightstill be observed underdifferentstudy conditions(e.g., asking students to learn
directlyfrom a set of documentswithoutany backgroundnformation).There werebothconsistentsimilaritiesanddifferences n theway novices and
specialistsevaluated he documents.Ingeneral,ODs were found most useful and
trustworthy y bothgroups.The novicestudents onsistentlyranked he HEs more
useful thanthe PAs, whereasthe reversewas truefor the specialists.Finally,the
TB was found useful in Problem1 but not in Problem2, due in partto students'
findingthatthe TB was redundantwith thechronology.To the novice students,whethera documentwas useful was mostlya matterof
content. In contrast,the specialists applied multiple criteria when evaluatingdocumentusefulness.Furthermore,hespecialists'criteriavariedacrossdocument
types. Theycitedmostlycontentfor secondhandaccounts, hat s, the TB and HEs
(with manyreferencesto task),mostly source for PAs, andmostly task for ODs
(withmanyreferences o content).These resultssuggest thatnovices andspecialistsstudied the documents with
differentpurposes.The novices' main concern was to build up a mentalrepre-
sentationof whathappened n Panama.To this end, the TB andHEs were usefulsources,and content was the most salient factor of usefulness. Incontrast,history
graduatestudents focused on interpretations nd evidence. They grantedmore
importance o primary ources(e.g., PAs, ODs). They disregarded he HEs (one
specialiststated hatshe wishedshe knew the historians'credentials).Thespecial-istsalsoused a broader angeof criteriawhenevaluatingusefulness.Itis likelythat
extensive academic training allows students to build up elaboratemodels of
differenttypes of documentsand theirpotentialuse for learning,reasoning,and
argumentation. his interpretations consistentwith Dillon's (1991) proposal hat
generalknowledgeof information ndcommunication ules.It does not takemuch
expertisen
historyo
recognizethata
personnvolvedin acontroversial ssue
maynotprovidea veryreliableaccountof the events. To a largeextent,trustworthiness
is also relatedto the intrinsiccharacteristics f a document(e.g., content,source)rather han o the role the documentplaysin aparticularaskorproblem.Whatmayevolve withdisciplineexpertise,however, s how theevaluationof trustworthiness
influencesthe actualuse of a document.Novice students oundPAs neitheruseful
nor trustworthy.They rejectedthese documentspossibly becausethey could not
handle the bias in them.Expertstudentsalso distrustedPAs but still found them
useful.Theirexperience n usingvarious orms of evidencehelpedthemdealwith
the bias in PAs. They could formulatehypotheseson why and how a president,asenator,or a lobbyist would provide slantedaccounts,and they interpreted he
documentaccordingly.When writingan opinion essay on the controversies,most specialistsresisted
takinganexplicit stance.In a way, the task of expressinganopinionseemedmore
difficult to them than it was to the novices. The data suggest that novices and
specialists nterpreted ifferently hetaskof expressinganopinionon acontroversy.Becausetheyhad been confrontedwith similarproblemsmanytimes,thegraduate
historystudentsknewthey were asked to studyactualcontroversies hatcould not
be answered n a simple,straightforward anner.Tothem,"expressyouropinion"meant describe the problemspace, the claims, andarguments hatmay be stated.
In contrast, he novices probablyassumedthatsupportingone particular pinionwas relevant because this was the assigned purposeof their writing. To them,
"expressyour opinion"meantdecidewhichside is correctandexplain why. In that
sense the novices represented he historicalcontroversiesas they wouldrepresent
any opinion problem (except perhapscontroversiesin their own discipline of
Thesestrategieswere similar o the heuristics dentifiedbyWineburg 1991) basedon think-aloudprotocols.The occurrenceof similarreasoningprocessesat studytime and during essay composition is an importantfinding. It indicates that
informationsources (sourcingheuristic),relations between sources (corrobora-
edge led them to mentionany information hey thoughtwas helpful,from back-
groundproblem nformation ogeneralworldknowledge.Incontrast,hespecialistswere able to draw from a richerknowledgebase and, hence, to select contextual
elementsat a moreappropriateevel of generality.Thesourcingheuristicwas equallyfrequent n novices' andspecialists'essays.
Moreover, hepattern f referenceswas similar o the one observedbyRouet,Britt,et al. (1996) with undergraduatetudentsandhistorygraduates.As in ourstudy,
Rouet, Britt,et al. foundthatstudentsreferredmostly to official documentsand
firsthandaccounts.Thus,neitheracademic evel nordisciplineexpertiseseemed to
influencethe use of explicitreferences n a draftessaywriting ask.However,there
may be more subtle differencesin the use of sources that our analysis failed to
capture.Whatdo these results ell us about earning rommultipledocuments?Knowing
how to handle document information s an advanced learningskill that manystudents ack (Guthrie,1988).To a largeextent,highereducationpromotesdocu-
ment literacy, and university graduatesmay be considereddocument experts
regardlessof theirdisciplineof specialization.However,disciplinessuch as psy-
chology andhistorydiffer n theiruse of document nformation elative o thetypesof problemswith which they deal. In thatsense, ourresults,along with those of
acase, to borrowanexpression romVoss et al. (1983), theyarenonexpert xperts.Ourstudywas anattempto demonstratehatexpertise n thedisciplineof history
involves boththe knowledgeof specific formsof discourseand theknowledgeof
how discoursemay be used as evidencewhen solving problems.We believe thatdocumentary xpertiseplays an important ole in manyareasof problemsolving,even thoughsucha roleis mostobviousin the social sciences.We alsosuggestthat
documentaryexpertiseis useful even at intermediateevels of instruction.Text-
booksoftenincludeselecteddocuments,and he studentmustknow how toevaluate
This content downloaded from 152.74.16.35 on Wed, 10 Apr 2013 23:53:31 PMAll use subject to JSTOR Terms and Conditions