Academic Dishonesty: Honor Codes and Other Contextual Influences Author(s): Donald L. McCabe and Linda Klebe Trevino Reviewed work(s): Source: The Journal of Higher Education, Vol. 64, No. 5 (Sep. - Oct., 1993), pp. 522-538 Published by: Ohio State University Press Stable URL: http://www.jstor.org/stable/2959991 . Accessed: 14/03/2012 16:18 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 support@jstor.org. Ohio State University Press is collaborating with JSTOR to digitize, preserve and extend access to The Journal of Higher Education. http://www.jstor.org DonaldL. McCabe LindaKlebeTrevino AcademicDishonesty HonorCodesandOtherContextual Influences Researchand media reportshave establishedthe continuedpervasivenessof academic dishonestyamong studentson America'scollegecampuses[12, 13, 22, 25, 26, 33, 46]. While some classes and increased collegeshave respondedwithacademicintegrity to convincereluctant to reportstudentcheaters efforts facultymembers in theconceptof"community" as an ef[13],thereis a renewedinterest fectivefoundationforcampusgovernance.For example,theCarnegie ofTeaching'sspecialreport,Campus FoundationfortheAdvancement Life:In Searchof Community, concludes,"Whatis needed,we believe, in highereducation. is a larger,moreintegrative visionof community a place whereindividuals to thegroupand accepttheirobligations wherewell-defined governanceprocedures guidebehaviorforthecommongood" [10, p. 7]. Derek Bok, in Universities and theFutureof America,echoesthis theme: needto consider thelarger [U]niversities campusenvironment beyondthe Anobviousstepinthisdirection classroom. istohaverulesthatprohibit lying,cheating, stealing, violent behavior, interference withfreeexpression, or otheractsthatbreakfundamental norms.Suchrulesnotonlyprotect the thesupport oftheRutgers Theauthors wouldliketo acknowledge Graduate School ResearchResourcesCommittee, of Management ExxonResearchand Engineering andFirstFidelity Company, Bancorporation. DonaldL. McCabeis associate at theGraduate SchoolofManagement, professor TheStateUniversity andLindaKlebeTrevino isassoRutgers, ofNewJersey-Newark, ciateprofessor at theSmealCollegeof Business ThePennsylvania Administration, Park. StateUniversity-University Journal Vol.64,No. 5 (September/October ofHigher Education, 1993) 0 1993bytheOhioStateUniversity Press Copyright AcademicDishonesty 523 inthecommunity; ofeveryone rights theyalsosignaltheimportance ofbasic andstrengthen behavior moralobligations habitsofethical [5,pp.84-85]. Bok offersthe honorcode as perhapsthe mosteffective approachin but acknowledgesthat,"the pervasive mattersof academicintegrity, forgrades;thesize,diversity, competition natureofmany andimpersonal largeuniversities; theirlack of anyhonorcode tradition;and thewidespreaddistasteforaccusingone'sclassmates"combineto workagainst such an approach[5, p. 87]. Althoughthehonorcode tradition dates back overa century, theviabilityof suchcodes on today'scampusesis open to some question[12]. Small, relatively homogeneouscampuses have generallygivenwayto large,culturally diverseinstitutions which lack anyapparentsenseofcommunity or commonpurposeamongstua credential dentsotherthangetting and a job. natureof thisquestion,thereis a surprising Despitethefundamental ofhonor paucityofempiricalresearchwhichaddressestheeffectiveness codes.The studydiscussedhereattempts to helpfillthisgap bycomparin collegesthathave honorcodes and those ing academicdishonesty thatdo not. The fewstudiesthathave addressedthe effectiveness of honorcodes[7, 9] havegenerally considered code effectiveness independentof context.We believethatit is importantto acknowledgeand understandthe complexityof the social systemswithinwhichhonor codesareembeddedand thefactthatothercontextual factorsmaybe as importantor moreimportantthantheexistenceof an honorcode by itself.Thus thisstudyextendsbeyondpreviousworkby studying the ofhonorcodeswithina morecomplexsocialcontext. effectiveness Honor Codes in Context AcademicDishonesty ofacademicdishonesty, thedatacollecDependingon one'sdefinition tionmethodsemployed,and othervariables,priorstudiesreportthat from13to 95 percentofcollegestudents anywhere engagein someform ofacademicdishonesty [12, 17,20,21,26,30,31,42]. A majordichotomy thatseparatesthesepriorstudiesis thelevelof analysis.One streamof to be predictive researchhas focusedon individual differences of thought cheatingbehavior,suchas gender[45],gradepointaverage[1,22],work ethic[15], TypeA behavior,competitive achievement-striving [35], and self-esteem on the [44]. In contrast,otherstudieshave concentrated institutional levelof analysisand examinedsuch contextualfactorsas honorcodes[7, 8, 9], faculty responsesto cheating[26],sanctionthreats [33,42], and sociallearning[33]. 524 Journalof HigherEducation approachhelpsto understand Althoughthe"individualdifferences" to cheat,the findingsare not particularly individuals'predispositions institutional foreffective administrator searching usefulto theuniversity contextual By investigating responsesto issuesof academicdishonesty. and thatare factorsthatmaybe associatedwithacademicdishonesty thestudydiscussedheremakesan iminfluence, opento administrative ofthiscomplexissue.In adto ourunderstanding portantcontribution questionofhonorcode effectiveness, ditionto addressingtheimportant including factors, itexaminestheroleofa numberofrelatedcontextual policiesare understoodand acthedegreeto whichacademicintegrity ofthesepolicies,and theperceivedbehaviorof cepted,theenforcement factors, contextual peers.Amongpreviousstudiesthathaveinvestigated yearsago, thatreaches we are aware of onlyone [6], conductedthirty and considerstheeffectiveness beyonda singleclassroomor university variables.Thus thisstudy, of honorcodes alongwithothercontextual factors whichinvestigates therelativeinfluenceof multiplecontextual institutions of higherlearning,makesan acrossstudentsat thirty-one to theextantliterature. contribution important TheInfluenceof Honor Codes on AcademicDishonesty honorcodes can be foundin minority, Althoughstillin a significant ofhigher education[18,41]. Sheldon numberofinstitutions an increasing Steinbach,generalcounseloftheAmericanCouncilon Education,was recently quotedas saying,"Honorcodes look liketheyare in a stateof in revivalon America'scollegecampuses"[41]. This renewedinterest honorcodes impliesthe beliefthatcodes can influenceacademicdisand thelimitedempiricalresearchavailablesupportsthisbelief. honesty, For example,Campbell[8] comparedcheatingamongstudentsunder and found an honorsystemand a proctorsystemat thesameuniversity thestudentsunderthehonorsystemwereless likelyto cheat.Canning infivesociologyclassesbeforeand afteran [9] conductedan experiment honorsystemwas established.Studentswereprovidedtheopportunity to cheatby gradingtheirown papers,withoutknowingthatduplicates The incidence made and gradedbytheinstructor. had been previously and of cheatingwas reducedafteran honorsystemwas implemented afterhavingthehonorsystemin place was reducedbynearlytwo-thirds forfiveyears.A surveyof medicalstudents[7] governedby an honor ofAlabamaSchool ofMedicinemeasuredstudent code attheUniversity ofand adherenceto thecode. Ninety-two percentofthereperceptions sugthattheyhad notobservedanycode violations, spondentsreported However,thelackofa non-codecomparison gestingcode effectiveness. thelowlevelofacademicdishongroupmakesitimpossibleto attribute AcademicDishonesty 525 estyto the code in thisstudy.One study,conductedin 1964,directly comparedacademicdishonesty at code and non-codeschools[6]. This - thosewhere studyfoundthatschoolswithtraditional honorsystems studentspledgeto abide by an honorcode and takeresponsibility for ofacademicdishonesty whenitoccurs- had and sanctioning detection thelowestrateof academicdishonesty. In general,previousresearchhas donelittleto developthetheoretical oftheproposedinverserelationship betweenhonorcodes underpinnings and to explainwhystudentsare expectedto and academicdishonesty cheat less when governedby an honorcode system.At least several plausibleexplanationsare possible.First,in mosthonorsystems,studentspledgeto abidebya code thatclarifies expectations regarding apis moreclearly propriate and inappropriate behavior.Thus,wrongdoing ofwrongdoing definedunderhonorcode systems.Whenthedefinition to rationalis madeclear,itbecomesmoredifficult forpotentialcheaters ize andjustifycheatingbehavior,and theincidenceof cheatingmaybe shifttheresponsibility for loweras a result.Second,honorcode systems controlof academicdishonesty fromfacultyand administrators to stuforthedestudentsare givenresponsibility dents[6]. In mostsystems, tectionofviolatorsand forthejudicialaspectsofthesystem, suchas determining guiltand assigningpenalties[6, 32]. Researchby Schwartz to the selfif [38] suggeststhatindividualsmustascriberesponsibility moralnormsare to be activatedand to influencebehavior.Therefore, cheatingmaybe lowerunderhonorsystemsbecause studentstakereforacademicdishonesty. sponsibility Finally,studentsgovernedby an suchas unproctored exams. honorsystemarefrequently givenprivileges It is likelythattheywouldbe willingto complywithan honorsystemto preservesuch valued privileges.Thus, cheatingmay be lowerunder the honorcode systemsbecausestudentswishto protecttheprivileges systemprovides. 1: Honorcodesare associatedwithdecreased Hypothesis academicdishonesty. Becausecodes varyextensively in theircontentand implementation, othercontextual factorsmayinfluence academicdishonesty beyondthe mereexistenceofa code. For example,thehonorcode maybe adminisor by teredbyfaculty, byadministrators, by studenthonorcommittees, ofthesegroups.At someuniversities, either somecombination students signa generalpledgenotto cheator arerequiredto signa specificpledge ofhonoron individualtestsand majorwritten At otherinassignments. such as Princeton,Smith,BrynMawr,and Rice,ifstudents stitutions, forreporting observean honorcode violation,theyareheldresponsible 526 Journalof HigherEducation of Virginia, it [4]. Some honorcodes,suchas theone at theUniversity are also tied to severesanctionsystemsrequiringexpulsionfor any seriousviolation[41]. At theotherextreme, an institution mayhavean honorcode thatis detailedin thestudenthandbookbutis relatively unknownto studentsand facultybecauseit is not a livingcomponentof ofhigheredtheinstitutional culture.Also,itis possibleforinstitutions ofacademicintegrity ucationto createa strongclimatesupportive withoutactuallyhavingan explicitcode. Given the complexityof social systemsand the variationsamong honorcodes,itis relevantto ask thequestion,howimportant arehonor Thisstudyprovidesan opporcodesrelativeto othercontextual factors? contributions tunityto examinetheindependent of a numberof such in a multivariate contextualvariablesand theirrelativecontributions context. on AcademicDishonesty OtherContextualInfluences and acceptanceof academicintegrity Understanding policies. Acaintheircontentand inthe demicintegrity policiescan differ significantly mannerin whichtheyare communicated. As a result,important differences are likelyto be foundin the understanding and acceptanceof thesepolicieson different campuses.For example,althoughit maybe thatstudentsand facultywouldnotknowoftheexistenceof a unlikely formalcode, its specificprovisionsmaybe poorlycommunicated and understood.Thus, studentsand facultywillbe less likelyto adhereto policiesthat theyeitherdo not know about or do not understand. thatfacultymembersmaybe reluctant Further,researchdemonstrates to followinstitutional policieswhentheyobservea studentcheating, to settlethematter"one-on-one" preferring withthestudent[4, 26, 34]. Finally,studentsmayviewacademicintegrity policiesas "justanother hurdleto be overcomeon theroad to a career"[18, p. 33]. Thus,combinedfacultyand student and supportoftheinstitution's understanding academicintegrity policiesmaybe moreimportant thanthesimpleexistenceor nonexistence ofan honorcode. 2: Academic willbeinversely tounderstanding Hypothesis dishonesty related andacceptance ofacademic integrity policies. Enforcement of ethicalguidelines.Whetheror not an explicithonor code exists,academicdishonesty maydependuponthesanctionsystem, how effectively academicintegrity particularly rulesand guidelinesare Deterrence enforced. thatformisconduct to be inhibited, theorysuggests mustperceive,first,thattheywillbe caughtand, second, wrongdoers thatseverepenaltieswillbe imposedforthemisconduct [19].The under- AcademicDishonesty 527 lyingcognitivemechanismis the expectationof punishment and the notionthatindividualswillbehavein waysthatmaximizerewardsand minimizecosts. Tittleand Rowe [42] foundthatthe threatof being caughtand punishedsignificantly deterred testcheatingamongcollege Michaelsand Miethe[33] foundthattheperstudents.More recently, of punishment ceivedprobability and theperceivedseverity of punishmentwereinversely correlated withcheatingin an honorcode setting. Becauseacademicdishonesty can oftenbe concealedfromfaculty memof being caughtmay bers,students'perceptionsabout the certainty depend upon the likelihoodthat anotherstudentwould reportthe of academicdishonesty misconduct.Such peerreporting is partof the explicithonorcode atmanyeducational institutions where"nontoleration clauses"oftenobligeand occasionallyrequirestudentsto reportviolations[18]. willbeinversely totheperceived Hypothesis 3: Academic dishonesty related ofbeingreported certainty bya peer. Deterrence theoryalso suggeststhat,all else beingequal, an increase intheseverity ofconsequences fora deviantactshouldreducethenumber ofindividualswillingto riskit [47]. If thepenaltyis severeenough,the potentialconsequencesmaysimplyoutweigh thepotentialrewardofthe misconduct. Not surprisingly, priorresearchindicatesthatthislogicappliesto studentperceptions concerning theseverity of penaltiesforacts of academicdishonesty [33]. willbeinversely 4: Academic related totheperceived Hypothesis dishonesty ofpenalties. severity Behaviorofpeers. Academicdishonesty mayalso be influenced by oftheirpeers'behavior.Thisproposedrelationship students' perceptions is supportedby sociallearningtheory[2] and thenotionthatunethical the theinfluence ofexample[37] orthrough behavioris learnedthrough ofcheatingin peergroups.For example, acceptanceand reinforcement studiesofacademiccheatinghavefoundfraternity/ sorority membership increasescheatingbehavior[1, 21, 39]. Althoughfrequently explainedin structures termsof opportunity [21, 39], thesefindings mayalso be exassociation plainedin termsof social learningtheoryand differential theory. Social learningtheory [2] emphasizesthatmuchofhumanbehavioris learnedthroughtheinfluenceof example.Individualslearnby observforthem.Thus,accordingotherpeople'sbehaviorand itsconsequences otherscheatsuccessfully ing to social learningtheory,seeingreferent 528 Journalof HigherEducation shouldincreasethetendencyof theobserverto behavein similarways [2]. Hypothesis 5: Academic willbe positively to perceptions dishonesty related ofpeers'academic dishonesty. administrators Beyond these hypothesizedbivariaterelationships, therelative dealingwithacademicdishonesty issuesneedto understand influencesof the independent variables(existenceof an explicithonor code,understanding and acceptanceof academicintegrity policies,cerof penalties,and peers'behavior)on taintyof beingreported, severity academicdishonesty. of thisstudyis contribution Thus, an important it providesto explorethe relativeinfluencesof these the opportunity contextual variables. ResearchProcedures Sample The hypothesesweretestedusinga sampleof 6,096 studentsfrom U.S. collegesand universities. Initialsampleselection thirty-one focused on thirty-three institutions that had participatedin a conferenceon honorsystemsheldat PrincetonUniversity in 1988,and fourteen institutions withhonorcodesultimately indicated a willingness to participate. Theseinstitutions weregenerally smalland highlyselectivein theiradmissionspolicies,and we identified non-honorcode institutions with similarprofiles. ofthethirty-eight Seventeen non-codeschoolscontacted also agreedto participate. The institutional characteristics ofthesecode and non-codeschoolssuggestthatthetwogroupsare comparablewith = 3441 fornon-codeand regardto bothsize (meannumberof students 3407forcode institutions) and academicselectivity (mean SAT score= 1236fornon-codeand 1249forcode institutions). The onlysubstantive is the inclusionof five all-femaleinstitutions difference in the code schoolsampleversusnoneamongthenon-codeschools.However,all of theanalysesdiscussedbelowwerereplicatedon thereducedsampleof coeducational twenty-six and no substantive differences were institutions, foundbetweenthetwosetsof analyses. Data Collection At twenty-eight ofthethirty-one institutions, surveysweremailedto a randomsampleof500 students in thefallof 1990.Threeschoolsasked to includetheirentireseniorclass in thesurvey.Thus, 15,904surveys weremailedin total.Respondentswereasked to completethe survey and return it by mailor to specialcollectionboxessetup anonymously AcademicDishonesty 529 for at a centralcampuslocation.A totalof6,096surveyswerereturned code The return rate at honor 38.3 percent. rate of an overallresponse higherthanat non-codeschools schools(41.4 percent)was significantly wereseniors,9 perpercentofrespondents (35.7 percent).Eighty-eight centwerejuniors,and theremaining3 percentcould not be classified. and mixofmajorswere on gender,ethnicity, The profiles ofrespondents profiles(based upon pubcomparedwiththeirrespectiveinstitutional contacts). data providedbyourinstitutional lishedfigures and additional of thattheprofiles suggested thesecomparisons Witha singleexception, profiles, lesinstitutional closelymatchedtheirrespective respondents bias in the sample. seningconcernsabout potentialnon-respondent However,thereis evidenceof some genderbias. Based on available foreach school,we wouldhaveexpecteda female/ profiles demographic surHowever,femalesreturned maleratioof55/45amongrespondents. ratioamong higherrate,fortheactualfemale/male veysat a significantly was 62/38.The primaryeffectof thisbias towardhonor respondents theactuallevelofcheating and womenis to understate code institutions in the overallsample,because the resultsindicatethatcheatingis less prevalentin each of thesegroups.However,separateanalysesforthe code versusnon-codeand thefemaleversusmalesegmentsofthesamthenatureofthereaffect ple suggestthatthesebiasesdo notmaterially discussedhere. lationships Measures The dependentvariablewas a compositemeaAcademicdishonesty. usacademicdishonesty: oftwelvetypesofself-reported sureconsisting duringa test;using ingcribnoteson a test;copyingfromanotherstudent unfairmethodsto learnwhatwas on a testbeforeitwas given;copying fromanotherstudentduringa testwithouttheirknowledge;helping someoneelseto cheaton a test;cheatingon a testin anyotherway;copyorfalsifying itin as yourownwork;fabricating ingmaterialand turning substaninworkdonebysomeoneelse;receiving a bibliography; turning on an assignment collaborating helpon an assignment; tial,unpermitted askedforindividualwork;copyinga fewsentences whentheinstructor it. Respondents ofmaterialfroma publishedsourcewithoutfootnoting had which the with to engagedin they wereasked identify frequency Likertscale (fromnever= 1 to each ofthesebehaviorson a four-point by calculating manytimes= 4). A compositemeasurewas constructed and thismeasurehad a scale violations, thetotalnumberofself-reported werealso evaluatedusing Cronbach'salpha of0.794. (All relationships thetestand non-test cheatingcomponentsof thismeasureas separate 530 Journalof HigherEducation werefoundbetween dependentvariables.No substantivedifferences theseanalysesand themoregeneralresultsdiscussedhere.) The compositeacademicdishonesty measurecould rangefrom12 to 48 forany individual,with12 representing no self-reported academic ofanykindand48 representing academicdishonesty dishonesty frequent acrossall ofthecategories described.The scoresforthisvariableranged from12to 45 withan overallmeanof 15.29and a standarddeviationof 4.21. However,theacademicdishonesty variablewas highlyskewed,and inspectionof theresidualsin subsequentregression analysessuggested thatthe standardassumptionsof normality wereviolated.Thus a log transformation [11] oftheacademicdishonesty variablewas used in all statistical analyses.(The residualsin theregression analysesemploying this log transformation are normallydistributed.) This transformed measureof academicdishonesty had a mean of 2.70 and a standard deviationof0.23. Explicithonorcode.A dummyvariablewas utilizedto signify whether an individualinstitution employedan explicithonorcode. The workof Melendez[32] was usedto definetheminimum codecriteria: unproctored an honorpledge,studentreportage examinations, and (non-toleration), of theexistence a courtorpeerjudiciary.Administrators atparticipating institutions providedthedata neededto makethedetermination and,in everycase,theyagreedwiththeauthors'classification oftheirinstitution. Understanding/acceptance ofacademicintegrity policy.Understandof the academic acceptance integrity policywas measuredwithfour ing/ itemson the studentsurvey.These itemsaskedtherespondent to rate: "the averagestudent'sunderstanding of [the institution's] policieson academichonesty," "thefaculty'sunderstanding of thesepolicies,""the faculty's supportofthesepolicies,"and "theeffectiveness oftheinstitution'spolicieson academichonesty"on a four-point Likertscale from verylow to veryhigh.Cronbach'salphaforthisscalewas 0.824. of thecertainty Certainty of beingreported.Studentperceptions of beingreportedweremeasuredby a singlefour-point Likert-scale item thatasked"how likely(veryunlikelyto verylikely)is it thatthetypical studentat [theinstitution] wouldreportsuchviolations?" Severityof penalties.Studentperceptionsof the severityof their school'spenaltiesforacademicdishonesty weremeasuredusinga single Likert scaleitemwhichaskedrespondents four-point to ratethe"severity ofpenaltiesforcheatingat [theirschool]"fromverylow to veryhigh. Peers'behavior.The measureof peerbehavioremployedhereconsistedof threeitems:studentperceptions of how frequently (neverto often on a either very five-point scale) plagiarismor testcheatingoc- AcademicDishonesty 531 curredat theirschooland an objectivemeasureoftheactualnumberof times(neverto manytimeson a four-point scale) the respondent had observedanotherstudentcheatingon a testor exam. Due to thediffermetrics ofthesevariables, each ofthecomponents encesintheunderlying was standardizedbeforecombiningthemintothepeers'behaviormeaz scoreswerecalculated,thethreez scores sure.For each component, werecombined,and a constantvalueof 10was added to thesumofthe z scoresto createthepeers'behaviorvariable.This measurehad a mean of9.99 and a standarddeviationof 2.43. Cronbach'salphaforthescale was computedto be 0.715. Findings A t-testwas used to comparethe meansof code and non-codereon theacademicdishonesty variable.As predicted spondents (hypothesis 1),self-reported cheatingwas significantly higheramongstudents in the non-codesamplethanamongthoseinthehonorcodesample(t = 26.073, means were2.62 in the honor p < 0.0001). The academicdishonesty code sampleand 2.77 at thenon-codeschools.(The meansfortheuntransformed measureof academic dishonesty were 14.00 and 16.56, respectively.) was significantly As detailedin table 1, academicdishonesty correlatedwith:(1) theunderstanding/ acceptanceofacademicintegrity policies (hypothesistwo); (2) the perceivedcertaintyof being reported (hypothesisthree);(3) the perceivedseverity of penalties(hypothesis of peers'behavior(hypothesis four);and (4) theperceptions five),supbivariaterelationships. portingall ofthehypothesized The moreinteresting policyquestionis therelativeimpactofeach of theseindependent variableson cheatingbehavior.To examinethisquesTABLE I of StudyVariables Intercorrelations N Measure Code of beingcaught Certainty of policy Understanding of penalties Severity Peers'behavior Academicdishonesty M S.D. Intercorrelations 4 2 3 1 6079 0.50 0.50 6036 2.05 0.73 0.46 5784 11.40 2.47 0.42 0.43 5884 3.00 0.79 0.31 0.31 0.51 5993 9.99 2.43 -0.43 -0.42 -0.49 -0.31 5904 2.70 0.23 -0.32 -0.27 -0.25 -0.12 aresignificant NOTE: Allcorrelations atp < 0.001. 5 0.51 6 Journalof HigherEducation 532 tion,a multipleregression modelwas constructed withacademicdishonestyas thedependentvariableand existenceof an explicithonorcode (enteredas a dummyvariable),the understanding/acceptance of academicintegrity of beingreported, theseverity of policies,thecertainty penalties,and peers'behavioras theindependent variables.The model was significant and,withtheexceptionoftheunderstanding/ acceptance of policymeasure,all of the independent variablesmade statistically to thefinalmodel.Peers'behaviorhad by far significant contributions the strongestinfluenceon academicdishonesty. These resultsare reportedin table2. Post hoc analysis.Giventheimportance ofpeers'behavior,a crucial issue may be to understandhow the widerinstitutional contextinfluencesthisvariable.Althoughwe didnothypothesize theserelationships, a post-hocmultipleregression analysisoftherelationbetweentheperceptionof peers'behaviorand othercontextualfactorsprovidessome insightto guidefutureresearchand policyconsiderations. As shownin table3, thisanalysisindicatesthattheunderstanding and acceptanceof academicintegrity policieshas the strongest associationwithstudents' perceptions oftheirpeers'behavior. Discussion The correlational data (table 1) supportthehypotheses thatacademic is positivelyassociatedwithperceptions of peers'academic dishonesty and negatively associatedwiththeunderstanding dishonesty, ofacademic integrity policies,the certainty of beingreported,and the severity of penalties.Further,a t-testfoundsignificantly lowerself-reported academicdishonesty amongstudentsat honorcode institutions. TABLE 2 of BeingReported, Regressionof Peers'Behavior,Code, Severityof Penalties,Certainty and Understanding of Policyon AcademicDishonesty Variable Explanatory Constant Peers'behavior Code of penalties Severity of beingreported Certainty of policy Understanding NOTE: N = b 2.234 0.045 -0.006 0.020 -0.017 0.002 Beta 0.468 -0.126 0.067 -0.052 0.020 = 435.37, 5466.R2= 0.2850,Adjusted R2= 0.2844,F(5,5461) p < 0.0001. p 0.0001 0.0001 0.0001 0.0001 0.1895 AcademicDishonesty 533 TABLE 3 of Policy,Code, Certainty Regressionof Understanding of BeingReported,and Severity of Penaltieson Peers' Behavior ExplanatoryVariable Constant of policy Understanding Code of beingreported Certainty of penalties Severity b 15.489 -0.296 -1.098 -0.625 -0.095 Beta -0.299 -0.225 -0.187 -0.031 p 0.0001 0.0001 0.0001 0.0155 NOTE: N = 5606. R2 = 0.3383,AdjustedR2 = 0.3378, F(4,5602)= 716.03,p < 0.0001. The multivariate ofpeers'behavior analysisfoundthattheperception contextual thatsociallearnwas themostinfluential variable,suggesting ingtheorymaybe particularly usefulforunderstanding academicdisThe stronginfluence ofpeers' honestybehavioramongcollegestudents. behaviormaysuggestthatacademicdishonesty notonlyis learnedfrom thebehaviorofpeers,butthatpeers'behaviorprovidesa kind observing ofnormative supportforcheating.The factthatothersarecheatingmay feelsleftat a disadalso suggestthat,in sucha climate,thenon-cheater vantage.Thus,cheatingmaycometo be viewedas an acceptablewayof gettingand stayingahead [18]. Considerthefollowingcommentsfrom is rampantat, . . . so much studyrespondents: "[Academicdishonesty] seemsto be everybody so thattheattitude does it- I'l be at a disadvantageifI don't";"If othersdo it,you'rebeingleftbehindbynotparticipating";"It'sthe'90's;yousnooze,youlose";"Whenmostoftheclassis exam and theywillruinthe curve,it influences cheatingon a difficult to cheat so you yourgradewon'tbe affected." An important ofthisstudyforfutureresearchand forthe implication managementof academicdishonestymay be that any movementto as thesolesolution adopthonorcodes is ill conceivedifitis undertaken to theacademicdishonesty is a complex problem.Academicdishonesty behaviorinfluenced bymultiplevariablesbeyondthemereexistenceof an honorcode. The post-hocanalysissuggested thatan institution's abilityto develop a sharedunderstanding and acceptanceofitsacademicintegrity policies has a significant and substantive impacton studentperceptions oftheir peers'behavior,themostpowerfulinfluenceon self-reported cheating. of thesepoliciesmaybe extremely Strivingformutualunderstanding important. Thus,programsaimedat distributing, explaining, and gainingstudentand faculty acceptanceofacademicintegrity policiesmaybe 534 Journalof HigherEducation mayin[33],theseprograms useful.As suggestedelsewhere particularly abouttheseriousness and cludethesigningofhonorpledges,reminders in coursesyllabi and statements consequencesof academicdishonesty, aboutacademicmisconduct. in thepresentresearchone of the lowestratesof selfInterestingly, codeinstitution. was foundat a non-honor reported academicdishonesty to the conceptof acacommitted is strongly However,thisinstitution demichonor,makingit a majortopicofdiscussionin itsstudenthandwhereitgoesto sessionsforincomingstudents, book and at orientation greatlengthsto ensurethatitspolicyis understoodand thatacademic At honoris theobligationofeverymemberofthecampuscommunity. the otherextreme,one of the higherincidencesof unethicalbehavior honorcode. However,stuwas foundat a schoolwitha long-standing and acceptanceof the dentsreporteda low level of understanding foradminisresponsibility school'spolicyand theofficialwithprimary this that thehonorcode supported finding bysuggesting theinstitering in communicating and implementing its itsefforts tutionhas diminished code in recentyears. In thefinalanalysis,themostimportant questionto ask concerning can createan environacademicdishonesty maybe how an institution is sociallyunacceptable,thatis,where mentwhereacademicdishonesty areclearlyunderstoodand wherestudents perinstitutional expectations Althoughthere ceivethattheirpeersareadheringto theseexpectations. are no simpleanswers,one alternative maybe Kohlberg's[27, 28] sugSchoolsorgangestionthatschoolsshouldbecome"justcommunities." ized aroundthe"justcommunity" conceptaregoverneddemocratically, in thedevelopment ofa socialcontractthatdeand students participate The unfinesnorms,values,and members'rightsand responsibilities. the institutional in is that climatecreated these"just assumption derlying thatarenecessary willprovidetheconditions formoraldecommunities" velopmentand behavior[23, 36]. Althoughmoreresearchis neededto can actuallydiscourageacademic whether determine "justcommunities" in heresug[24],thedata reported dishonesty highereducationsettings study. gestthatthisconceptdeservesfurther itis forthisposition, In additionto themoreobjectivesupportoffered to look at studentcommentsfromhonorcode schoolsthat instructive foracademic seemto have achievedsomesharedsenseofresponsibility For example,in responseto a question amongtheirstudents. dishonesty fornotcheating, thefollowing whichprobedmotivations responseswere and wouldn'tdo notunusual:"I liketherespectI getat [theinstitution] to jeopardizethat";"Peerpressure- youwouldfeelveryemanything AcademicDishonesty 535 saw it";"as forcheatingon a test,it'ssocially barrassedifotherstudents thingsin high "I didmanyofthese'academicdishonesty' unacceptable"; is at [theinstitution] theatmosphere school butnotsincearriving one of respectforthestudent- and so I haverespectforthesystem"; "Respectforothers openness/freedom"; "Respectfor[theinstitution's] in the class - cheatingmay penalizethemas well."One respondent pointedout,"We do not'report'peopleliketheHonorBoardwas some on a student-to-student levelis muchmore hitsquad - confrontation common& alwaysthefirstand usuallytheonlysteptaken."We believe thatAmitaiEtzioni[16] would see thesecommentsas evidencethat of any is intact,thereis littleneedforgovernment "whena community kind." cheatstudents at schoolswithhighlevelsofself-reported Bycontrast, in termsofa 'we'veringoftendiscusstheissueofacademicdishonesty is acceptablebecausethey Cheatingbyus (students) sus'they'mentality. 'deserve'itforanynumberofreasons(facultyand/oradministrators) thepoorqualityofteaching,and unclearinunreasonableassignments, on majorassignments werementioned frequently bystudents. structions in lightof Of course,theconclusionsstatedheremustbe interpreted limitations of thisstudy.For example,becausethis themethodological surveydata, the observedlinkages studyreliedupon cross-sectional arecorrelavariablesand academicdishonesty betweentheindependent tional and not necessarilycausal. In addition,althoughwe insured ofthesurveydata,bias due to and confidentiality completeanonymity theresults.For and social desirability mayhave influenced self-report at a code institution would be example,it is possiblethatrespondents on a survey.Further,given less willingto admitacademicdishonesty our sample,we can notgeneralizebeyondsmall,selectiveinstitutions. Thereis some researchevidenceto suggestthatless cheatingoccursin forresearchers moreselective institutions [6], anditis important smaller, to eithermeasureor controlfor(as was done here)theseextraneous in absolute sourcesof variance.However,despitepossibledifferences reasonto believethattheunderlycheatinglevels,thereis no theoretical variablesand the betweentheindependent ingnatureoftherelationships in larger, lessselectiveinstitutions. variablewouldbe different dependent in comparingacademicdishonesty interest Finally,givenourprimary in code versusnon-codeinstitutions, we believethatthe largemultiequal sample used here,that includesan approximately institution ofthisstudy. numberofcode and non-codeschools,is a majorstrength The sampleis particularly becausethesmaller,moreselective interesting of a institutions institutions studiedhererepresentthe undergraduate 536 Journalof HigherEducation shareof our nation'sleaders.For example,almostone disproportionate in five(18.5 percent)of Business Week's 1991 CorporateElite (the CEOs ofthenation'stop 1000firms)earneda degree[14] at one of the institutions. participating Thus,thisstudyprovidesa uniqueopportunthathelpto shapetheleadityto understand theacademicenvironments ersofthisnation. contributes to a smallbutgrowingempiricalliteraThisinvestigation tureon academicdishonesty and is one offewempiricalstudiesto examine the role of honorcodes in combinationwithothertheoretically These importantcontextualfactorsin reducingacademicdishonesty. relevantbecause theyare open to contextualvariablesare particularly administrative influence.In responseto the keyquestionabouthonor ofan honcodeeffectiveness, thefindings suggestthatthemereexistence or code is not as important as othersocial contextfactors.Academic is most stronglyassociatedwiththe perceptionsof peers' dishonesty to a lesserdegreeby the existenceof a code, behavior.It is influenced of penaltiesand, incertainty of beingreported, the perceivedseverity and acceptanceof academic integrity directly, by the understanding how policies.Althoughadditionalresearchwillbe neededto understand thesefactorsoperateand howtheycan be influenced, clearlythefindings of thisresearchare hopeful.Althoughtheyleave manyquestionsunanswered,theydemonstrate thatthevisionoftheCarnegieFoundation accepttheirobligathatourcampusesbecomeplaces"whereindividuals tionsto thegroupand wherewell-defined governance procedures guide Albehaviorforthe commongood" [10, p. 7] maynot be unrealistic. thoughthetask willbe difficult, thesuccessachievedby severalof the schoolsstudiedin thissample,code and non-codeschoolsalike,attest thatit is possible. References 1. Baird,J. S. CurrentTrendsin College Cheating.Psychologyin the Schools, 17 (1980),515-22. 2. Bandura,A. Social Foundationsof Thoughtand Action.EnglewoodCliffs,N.J.: Prentice-Hall. 1986. "ConcernsofCollegeStudents and G. Kozmetsky. 3. Beltramini, R. F., R. A. Peterson, RegardingBusinessEthics."Journalof BusinessEthics,3 (1984), 195-200. 4. Berger,J. "Honor Code: Rewardsand Pitfallsof an Ideal." New York Times,9 March 1988,p. B9. 5. Bok, D. Universities and theFutureof America.Durham,N.C.: Duke University Press,1990. 6. Bowers,W. J. StudentDishonestyand Its Controlin College.New York:Bureauof 1964. AppliedSocial Research,ColumbiaUniversity, AcademicDishonesty 537 7. Brooks,C. M., et al. "StudentAttitudestowarda Medical School HonorCode." Journalof Medical Education,56 (1981),669-71. Investigation ofStudentsunderan HonorSystem 8. Campbell,W. G. A Comparative Los Angeles:University ofSouthern and a ProctorSystemin theSame University, CaliforniaPress,1935. 9. Canning,R. "Does an HonorSystemReduceClassroomCheating?An ExperimenEducation,24 (1956),291-96. tal Answer."Journalof Experimental ofTeaching.CampusLife:In Searchof 10. CarnegieFoundationfortheAdvancement Princeton,N.J.: The CarnegieFoundationforthe Advancementof Community. Teaching,1990. Analysisfor 11. Cohen,J., and P. Cohen. Applied MultipleRegression/Correlation theBehavioralSciences.2nded. Hillsdale,N.J.:Erlbaum,1983. 12. Collision,M. "Surveyat RutgersSuggeststhatCheatingMay Be on the Rise at Chronicle ofHigherEducation,24 October1990,pp. A31- A32. LargeUniversities." in Worried."The . 13. "ApparentRise Students'CheatingHas CollegeOfficials Chronicleof HigherEducation,17 January1990,p. A33. . "The CorporateElite."BusinessWeek,25 November1991,185-216. 14. TrainingAfR., and D. M. Shank."PersonalWorkEthicand Effort 15. Eisenberger, 49 (1985),520-28. and Social Psychology, fectCheating."Journalof Personality of Thinkers, BothLiberaland Conservative." The 16. Etzioni,A. "A New Community WallStreetJournal,8 October1991,p. A22. "ScholasticDishonestyamongCollegeUndergrad17. Eve, R. A., and D. G. Bromley. uates: Parallel Tests of Two SociologicalExplanations."Youthand Society,13 (1981),3-22. EducationalRe18. Fass, R. A. "By HonorBound: EncouragingAcademicHonesty." cord,67 (1986), 32-35. and Deterrence. Amsterdam: Elsevier,1975. 19. Gibbs,J. P. Crime,Punishment, and the Lack of Commitment 20. Haines,V. J.,et al. "CollegeCheating:Immaturity, Attitude." Researchin HigherEducation,25 (1986), 342-54. Neutralizing and Social Structure:A Studyof 21. Harp, J., and P. Taietz. "Academic Integrity Social Problems,13 (1966),365-73. CheatingamongCollegeStudents." E. M., and S. E. Feldman."CollegeCheatingas a Functionof Sub22. Hetherington, ject and SituationalVariables."Journalof EducationalPsychology,55 (1964), 212-18. 23. Higgins,A. "Researchand MeasurementIssuesin MoralEducationInterventions." In Moral Education:A FirstGenerationof Researchand Development,editedby R. L. Mosher,pp. 92-112. New York:Praeger,1980. inTwo 24. Higgins,A., and F. Gordon."WorkClimateand Socio-MoralDevelopment Companies."In Moral Education:Theoryand Application,edited Worker-owned byM. W. Berkowitzand F. Oser,pp. 241-68. Hillsdale,N.J.:Erlbaum,1985. 25. Jayna,M. R. "Ah, Fall: Class, Exams - and Cheating."APA Monitor,22 (1991), 28. Journalof College 26. Jendrek,M. P. "FacultyReactionsto AcademicDishonesty." StudentDevelopment,30 (1989),401-6. 27. Kohlberg,L. "HighSchool Democracyand Educatingfora JustSociety."In Moral 538 Journalof HigherEducation Education:A First Generationof Researchand Development,editedby R. L. Mosher,pp. 20-57. New York:Praeger,1980. . "The JustCommunity 28. Approachto MoralEducationin Theoryand Practice."In Moral Education:Theoryand Application, editedbyM. W. Berkowitz and F. Oser,pp. 27-88. Hillsdale,N.J.:Erlbaum,1985. 29. Lane, M. S., and D. Schaupp."Ethicsin Education:A ComparativeStudy."Journal of BusinessEthics,8 (1989),943-49. 30. Leming,J. S. "CheatingBehavior,Subject Variables,and Componentsof the Internal-External Scale UnderHighand Low RiskConditions." Journalof EducationalResearch,74 (1980),83-87. 31. Liska,A. "DeviantInvolvement, Associationsand Attitudes: theUnderSpecifying lyingCausal Structures." Sociologyand Social Research,63 (1978),73-88. 32. Melendez,B. Honor Code Study.Cambridge,Mass.: HarvardUniversity, 1985. 33. Michaels,J. W., and T. D. Miethe."ApplyingTheoriesof Devianceto Academic 70 (1989),872-85. Cheating."Social ScienceQuarterly, 34. Nuss, E. M. "AcademicIntegrity: ComparingFacultyand StudentAttitudes." ImprovingCollegeand University Teaching,32 (1984), 140-44. 35. Perry,A. R., et al. "Type A Behavior,CompetitiveAchievement-Striving, and CheatingamongCollegeStudents." PsychologicalReports,66 (1990),459-65. 36. Power,F. C., A. Higgins,and L. Kohlberg.LawrenceKohlberg'sApproach to Moral Education.New York:ColumbiaUniversity Press,1989. 37. Rosenhan,D. L., B. S. Moore, and B. Underwood."The Social Psychologyof Moral Behavior."In Moral Developmentand Behavior,editedbyT. Lickona,pp. 241-52. New York:Holt,Rinehart& Winston.1976. S. "Words,Deeds,and thePerceptions 38. Schwartz, ofConsequencesand Responsibilityin ActionSituations." Journalof Personality and Social Psychology, 10 (1968), 232-42. 39. Stannard,C. I., and W. J. Bowers,"The College Fraternity as an Opportunity forMeetingAcademicDemands."Social Problems,17 (1970),371-90. Structure 40. Stevens,G. E. "EthicalInclinations ofTomorrow's Citizens:ActionsSpeak Louder?" Journalof BusinessEducation,59 (1984), 147-52. 41. Tabor,M. B. W. "HonorCodes Geta SecondLook."Christian ScienceMonitor,30 January1987,p. B3. 42. Tittle,C. R., and A. R. Rowe."Moral Appeal,SanctionThreat,and Deviance:An Test."Social Problems,20 (1973),488-97. Experimental 43. Tom,G., and N. Borin."Cheatingin Academe."JournalofEducationfor Business, 63 (1988), 153-57. 44. Ward,D. A. "Self-Esteem and DishonestBehaviorRevisited."TheJournalof Social Psychology,126(1986),709- 13. 45. Ward,D. A., and W. L. Beck. "Genderand Dishonesty."The Journalof Social 130(1990),333- 39. Psychology, 46. Wellborn,S. N. "Cheatingin CollegeBecomesan Epidemic."US. News & World Report,89 (20 October1980),39-42. 47. Zimring, F. E., and G. J. Hawkins.Deterrence:TheLegal Threatin CrimeControl. ofChicagoPress. 1973. Chicago:The University