Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=vjeb20
Journal of Education for Business
ISSN: 0883-2323 (Print) 1940-3356 (Online) Journal homepage: http://www.tandfonline.com/loi/vjeb20
The Effect of Using Case Studies in Business
Statistics
Susan E. Pariseau & Boualem Kezim
To cite this article: Susan E. Pariseau & Boualem Kezim (2007) The Effect of Using Case Studies in Business Statistics, Journal of Education for Business, 83:1, 27-31, DOI: 10.3200/ JOEB.83.1.27-31
To link to this article: http://dx.doi.org/10.3200/JOEB.83.1.27-31
Published online: 07 Aug 2010.
Submit your article to this journal
Article views: 55
View related articles
esearchersineducationhaveclearly demonstrated that students’ active participation in the educational process increases their learning and retention. In an active, collaborative, or coopera-tivelearningenvironment,studentstake moreresponsibilityfortheirowneduca-tion,andteachingeffectivenessincreases astheteacherbecomesafacilitatorand coachratherthanalecturer.
The basic elements of an active learning environment include interde- pendence,face-to-faceinteraction,indi-vidual accountability, the use of inter-personal skills, and group processing (Johnson, Johnson, & Holubec, 1990). Methodssuchassmallgroupactivities, case studies, and cooperative student projectsgivestudentstheopportunityto activelyparticipateinthelearningpro- cessbytalking,listening,reading,writ-ing, and reflecting (Meyers & Jones, 1993). Many authors have argued that learning increases when students make use of multiple senses (Kvam, 2000). Thus, augmenting lectures with group work,inwhichstudentsdiscussmate-rial, solve case studies, write written reports,andmakepresentations,should enhance the learning experience in a businessstatisticscourse.
Different students learn in different ways(Kolb,1984).Somestudentsprefer tolistenandtakenotes,whereasothers needtobeactivelyinvolvedinthelearn-ing process by speakneedtobeactivelyinvolvedinthelearn-ing and sharneedtobeactivelyinvolvedinthelearn-ing
their own experiences. Some students prefer individual work, whereas others prefer collaborative work. Considering thesediverseneeds,itseemsreasonable for educators to conclude that using a variety of strategies will help to keep students interested and involved while enhancingtheirlearning.
Rosenthal (1995) used in-class small groupexercisesandout-of-classwriting assignmentsinupper-divisionprobability theory courses with mixed results. Stu-dentsgavefavorableratingstothegroup exercises, which typically lasted one or twoclasshours,andgaveneutralratings to the writing assignments. The use of peerreviewwiththewritingassignments enhanced the level of satisfaction that participantsreported.Rosenthalreported thattheuseofin-classexercisesseemed toincreasestudentdiscussionofcourse materialoutsideofclass.
Kvam (2000) studied the effect of group projects on student retention in introductoryengineeringstatisticscours-es.Groupsof4–6studentscompleteda totaloffourprojectsina10-weekclass whilethecontrolgrouplearnedbyatra-ditionallectureformat.Kvammeasured retention immediately after students completedthecourseand8monthslater. The results suggested that collaborative workincreasedretentionforaveragestu-dentsinthecourse.
As early as 1970, Allan Birnbaum (1971) expressed dissatisfaction with
TheEffectofUsingCaseStudiesin
BusinessStatistics
SUSANE.PARISEAU BOUALEMKEZIM MERRIMACKCOLLEGE
NORTHANDOVER,MASSACHUSETTS
R
ABSTRACT.Theauthorsevaluatedthe effectonlearningofusingcasestudies inbusinessstatisticscourses.Theauthors dividedstudentsinto3groups:acontrol group,agroupthatcompleted1casestudy, andagroupthatcompleted3casestudies. Resultsevidencedthat,onaverage,students whomtheauthorsrequiredtocompletea caseanalysisreceivedsignificantlyhigher gradesontheircomprehensivefinalexami-nation.Theirperceptionsoftheirlearning experiencewerealsosignificantlyhigher withrespecttocommunication,software use,andtheabilitytoapplystatisticsto businessdecisions.
Keywords:activelearning,casestudies, experientiallearning,pedagogicalresearch, statistics
Copyright©2007HeldrefPublications
the teaching of statistics and urged the involvementofeachstudentofstatistics in research representing a discipline of interest to that student. He encouraged thedevelopmentofcasestudiesthatedu-catorscoulduseinteachingstatistics.
Statistics instructors face the simul-taneous challenges of motivating stu-dents to study material that they think isuninterestingandofdealingwithper-formance extremes and attempting to increase retention (Conners, Mccown, & Roskos-Ewoldsen, 1998). Jamison (1998)foundthatstudentprojectspro-motedcriticalthinkingandthatstudents displayed enthusiasm for the research projects.Ameta-analysisofresearchon the effects of cooperative learning that Johnson,Johnson,andHolubec(1990) completedin1989indicatedthat“coop-erative learning situations are more effective in promoting achievement thanindividualisticortraditionallearn-ing situations” (p. 30). Those authors found benefits in cooperative learning situationsthatincludedanincreaseduse ofcriticalthinkingcompetencies;more positiveattitudesaboutthesubject;and more helping, encouraging, and assist-ingamongstudents.
Harrington and Schibik (2004) sug-gested ways for educators to create a collaborative environment in teach-ing business statistics. Harrington and Schibik suggested alternatives includ-ing use of statistics software, writinclud-ing assignments, case study methods, and teamprojects.
Hakeem (2001) studied the effect of using a semester-long active-learning project in business statistics. Hakeem reported that students completing the project performed better on exami-nations. More students in the project groupreceivedagradeofA,andfewer studentsintheprojectgroupreceiveda gradebelowC.
Although Hakeem (2001) focused on the use of a semester-long project, in our study, we used team case-study analysis as an active-learning device. Thecasestudiesusedrealdataandwere open-ended,thusrequiringeachstudent to develop critical thinking skills both individuallyandasamemberofateam. We required students to complete all analysesandtomakerecommendations inwritingtogainexperienceinmaking
appropriatedecisionsinabusinessenvi-ronment. This methodology highlights the relevance of the study of statistics forbusinessstudents.
METHOD
To test the effect of case studies on students’ learning in business statistics courses, we studied eight sections of the same undergraduate course during twofallsemesters.Thecourseisafour-credit introductory course in business statistics that primarily first-semester sophomore students take. All sections weretaughtbythesameprofessorand extensivelyusedcomputerapplications such as Excel. All students used their ownlaptopcomputersduringeachclass meeting.The8sectionsmet4hoursper weekovera14-weeksemester.Allsec-tions used the same syllabus and were assigned the same problems. The final comprehensive examination was the sameforallclasses.Ofthefinalexami- nation,80%consistedofsolvingprob-lemsand20%involvedtheanalysisofa minicase.Tominimizepotentialbiasin grading the minicase, the authors used a common grading rubric with clearly definedexpectationsandscoresforeach componentoftheminicase.
Duringthefirstsemesterofthestudy, we randomly selected two of the four sections of statistics to participate in theone-casegrouptotesttheeffectof usingcasestudiesinbusinessstatistics courses. We required students in these twosectionstocompleteonecasestudy butdidnotrequirestudentsintheother two sections to complete a case study (zero-casegroup).Becausetwoclasses metinthemorning(8:00and9:30a.m.) and two classes met in the afternoon (12:00 and 1:30 p.m.), we randomly selected one morning class and one afternoonclasstoparticipateintheone-case group. We used this procedure to eliminatethepotentialforatime-of-day effectonstudents’performance.During the second semester of the study, we required all four sections to complete threecases(multiple-casegroup).
For the sections that would com-plete cases, we randomly divided the class into teams of 4 or 5. Each team completedworkonallcasesoutsideof class. We required each team to
com-pleteawrittenreportandmakeanoral presentation of their case analysis. We evaluatedteamsonthebasisofcontent, presentation, and the ability to answer questions. We encouraged members of nonpresentingteamsintheclasstoask questions,andstudentsreceivedabonus for participation if they asked a ques-tion.
Early in each of the two semesters, wegaveapretesttoallstudentstodeter-minewhetherthestatisticalbackground ofthestudentswasthesame.Thepre-testwasamultiplechoiceexamination that covered all topics in introductory businessstatisticscourses.Weusedthe pretesttodeterminewhetheronegroup hadmorebackgroundinstatistics,per-hapsduetocourseworkinhighschool. A difference in statistical background betweenthe3groupscouldpotentially havebiasedthestudy.
At the end of the first semester, we surveyed students to determine their perceptions of the statistics class. We used a 7-point Likert-type scale, with students selecting a number between 1 (strongly disagree) and 7 (strongly agree) to express their perceptions of theirbusinessstatisticsexperiences.
RESULTS
We tested results from the pretest, whichwehadgivenearlyinthesemes-ters, by using analyses of variance (ANOVA).As we expected, the results indicatedthattherewerenosignificant differences in mean scores among the three groups (Table 1).Thus, the three groups started the course with similar statisticalbackgrounds.
At the end of the course, we admin-istered the same comprehensive final examinationtoallstudents.Theexami-nationconsistedofproblems(80%)and a minicase (20%).We separated grades on the exams into two portions so that we could examine the results for the minicase separately from those of the problemportionoftheexam.Weexpect-ed students in the one-case group and the multiple-case group to outperform thoseinthecontrolgroupinbothareas. Weexpectedthemultiple-casegroupto outperformbothothergroups.
Figure1displaysabox-and-whisker plotforgradesontheminicaseportion
of the examination. Regarding the box plots for the three groups, it is clear thatthemediancasegradewashighest for the group that completed multiple cases (multiple-case group) and lowest forthegroupwithzerocaseexperience (zero-case group). The distribution of casegradesfortheone-casegroupand that of the multiple-case group appear quitesimilar,althoughthefirstandthird
quartiles are higher for the multiple-casegroup.
The second quartile or median score for the zero-case group is identical to the first quartile score for the one-case group. Of the students in the zero-case group, 50% scored below 11; whereas oftheone-casegroup,only25%scored below11.Thebeststudentsinthezero-case group appeared to be capable of
performingwellontheminicaseanalysis even without benefit of a required case study during the course. In all groups, some participants received low scores. Itappearsthatforthepoorerperforming students,theclasscaseanalysismaynot have prepared them better for the mini-caseintheexam.Thecasestudyseems tohaveenhancedlearningfortheaverage andbetterstudentsinbothcasegroups.
Althoughtheboxplotclearlyexhibits adifferenceinscores,itisnecessaryto perform a statistical test to verify the hypothesis that the mean scores for the groupswouldbedifferent.First,weper-formed a test for equality of variances, whichdidnotrejectthenullhypothesis of equality of variances. Next, assum-ingequalityofvariances,weperformed ANOVA, and the results (Table 2) sup-ported the hypothesis that the mean scoresonthecaseportionofthecompre-hensivefinalexaminationwereunequal. Next, we completed an LSD post hoc pair-wise comparison. Table 3 clearly displays significant differences between the zero-case group and the one-case group and between the zero-casegroupandthemultiple-casegroup. Thehypothesisthatthemeanscorefor the multiple-case group is greater than the mean score for the one-case group issignificantatthe0.1level.Foraone-tailedpair-wisecomparison,thepvalue is0.085,showingmildsignificance.
Bothcasegroups(one-andmultiple- casegroups)hadtheopportunitytocom-pleteoneormorecasestudiesforwhich theyreceivedfacultyandpeerfeedback. Theirmeangradesonthecaseportionof thefinalexamweresignificantlyhigher than the mean grade of the zero-case group.
We completed a second box-and-whisker plot for the final examination scoreminusthecaseportion(Figure2). As the interquartile range evinced, the scores for the average students in the multiple-case group were more consis-tentthanthescoresoftheothergroups. Thezero-casegroupexhibitedfargreat-er dispersion in scores. It appears that activelearningenhancedthelearningof statistics and not merely the ability to completeacaseanalysis.
Although it appeared obvious that the scores on the final examination minus those of the case portion were
TABLE1.AnalysisofVarianceforStatisticalKnowledgePretest
Factor n M SD F(2,151) p
Zerocases 43 37.23 11.81 1.55 .217 Onecase 43 37.56 10.15
Multiplecases 68 40.34 9.46
TABLE2.AnalysisofVarianceforCasePortion(20%)ofFinalExamination
Item n M SD F(2,161) p
Zerocases 43 11.71 3.28 10.96 .000 Onecase 43 13.89 3.63
Multiplecases 78 14.81 3.44
TABLE3.MultipleComparisonforCasePortion(20%)ofFinalExamination
Factor(I) Factor(J) MeanDifference(I–J) SE p
Zero One –2.1861 .75448 .004 Multiple –3.1048 .66448 .000 One Zero 2.1861 .75448 .004 Multiple –0.9188 .66448 .169 Multiple Zero 3.1048 .66448 .000 One 0.9188 .66448 .169
FIGURE1.Box-and-whiskerplotforcaseportionoffinalexamination.
Multiplecases
Onecase
Zerocases
0 5 10 15 20
Scoreoncaseportionoffinalexam(20%maximum)
higherforbothcasegroups,wecom-pleted an analysis. We performed a test for equality of variances, which
did not reject the null hypothesis of equalityofvariances.Next,assuming equality of variances, we performed
ANOVAs, and the results (Table 4) supported the hypothesis that there was a difference in the mean scores on the problem portion of the final examination among the three groups ofstudents.
Ananalysisoftheposthocpair-wise comparisons(seeTable5)providedevi-denceofsignificantdifferencesbetween the zero-case group and the one-case groupandbetweenthezero-casegroup and the multiple-case group. However, we found no differences between the one-case group and the multiple-case group.Apparently,theuseofmorethan one case did not improve performance ontheproblemportionofthecompre-hensivefinalexamination.
Attheendofthefirstsemesterofthe study,weaskedeachstudenttocomplete a survey. We gave the following four statements regarding perceptions of the businessstatisticscourseexperience:
1.Mybusinessstatisticscoursedevelops theabilitytocharacterizeandanalyze numerical data through the use of descriptivesummarymeasures. 2.My business statistics course
devel-ops the ability to apply inferential statistics to make informed business decisions.
3.My business statistics course devel-ops the ability to actively communi-catetheresultsofstatisticalanalysis. 4.Mybusinessstatisticscoursedevelops
the ability to employ software to aid intheanalysesofbusinessproblems.
Weaskedstudentstothinkabouttheir currentbusinessstatisticscourseandto show the extent to which they thought theircoursepossessedthefeaturesthat each statement described by using a Likert-type scale that ranged from 1 (stronglydisagree)to7(stronglyagree). We hypothesized that the perceptions of the students in the one-case group would show a higher level of agree-mentwitheachofthesestatements.The seriesoftwoindependent-samplettests validatedthishypothesis(Table6).
DISCUSSION
Students in the one-case group and the multiple-case group were involved inmanyactivelearningtechniques.Each
TABLE4.AnalysisofVarianceforFinalExaminationScoreMinus Case(80%)
Factor n M SD F(2,161) p
Withoutcase 43 57.39 13.66 3.29 .04 Withonecase 43 62.98 11.64
Withmultiplecases 78 62.97 11.74
TABLE5.MultipleComparisonsforFinalExaminationScoreMinus CasePortion
Factor(I) Factor(J) MeanDifference(I–J) SE p
Zero One –5.58140 2.6420 .036 Multiple –5.57901 2.32612 .018 One Zero 5.58140 2.64120 .036 Multiple 0.00239 2.32612 .999 Multiple Zero 5.57901 2.32612 .018 One –0.00239 2.32612 .999
TABLE6.ttestforPerceptionsofBusinessStatisticsCourse
Question t(81) p
1.Mybusinessstatisticscoursedevelopstheabilityto characterizeandanalyzenumericaldatathroughtheuseof
descriptivesummarymeasures. –1.70 .046 2.Mybusinessstatisticscoursedevelopstheabilitytoapply
inferentialstatisticstomakeinformedbusinessdecisions. –2.26 .013 3.Mybusinessstatisticscoursedevelopstheabilitytoactively
communicatetheresultsofstatisticalanalysis. –2.73 .004 4.Mybusinessstatisticscoursedevelopstheabilitytoemploy
softwaretoaidintheanalysesofbusinessproblems. –1.71 .045
FIGURE2. Box-and-whisker plot for final examination score minus case portion.
Multiplecases
Onecase
Zerocases
30 40 50 60 70 80 Examscorewithoutcaseportion(80%maximum)
student was a member of a small team thatwerequiredtocompleteoneormore caseanalysesduringthesemester.Team-mates worked together outside of class, andwerequiredthemtosubmitwritten reports and to make one professional presentationoftheircaseanalysistotheir classmates. Thus, we exposed students tobothalectureenvironmentandacol-laborativelearningenvironmentinwhich theyactivelyparticipatedbytalking,lis-tening, and writing. The results of the use of case analysis and collaborative learning included an increase in criti-cal thinking competencies as the abil-ity to effectively complete the minicase onthefinalcomprehensiveexamination indicated. We found strong, significant differencesinperformancesonthemini-case between both differencesinperformancesonthemini-case groups and the zero-casegroup.Themeanperformance of the multiple-case group was mildly higherthanthatoftheone-casegroup.
Mean performances on the prob-lemportionofthefinalcomprehensive examination were significantly higher forbothcasegroups.Wefoundnodif-ferences between the one-case group andthemultiple-casegrouponthispor-tionoftheexamination.
Studentsintheone-casegroupevinced more positive attitudes about the sub-jectofstatistics,asstudentperceptions of their statistics class indicated.Their perceptionsoftheirlearningexperience were significantly higher with respect to communication, software use, and theabilitytoapplystatisticstomaking businessdecisions.
Wefounduseofasinglecasestudy tohavealmostassignificantanimpact onthelearningenvironmentastheuse of multiple cases.We suspect that the collaborative learning environment fostered more helping, encouraging, and assisting among students and that the collaborative learning environ-ment fostered these positive elements irrespective of the number of outside cases that we required. This seems to support the findings of Rosenthal in which in-class collaboration fostered moreout-of-classdiscussionofcourse materials.
LimitationsandFutureResearch Werecommendthatfutureresearch-ers try to determine why additional benefits did not arise when students completedmultiplecases.Wasoursup-positioncorrectthatonlyonecasealone (one-case group) fostered collaborative learning? Did the multiple-case group begintoseekefficiencythroughdelega-tion of tasks following complebegintoseekefficiencythroughdelega-tion of the first case? This efficiency would lead to less collaboration on the sec-ond and third cases and could account fortheinsignificantdifferencesthatwe found between the one-case group and themultiple-casegroup.
NOTES
Dr.SusanE.Pariseau isaprofessorofman-agement at Merrimack College, North Andover, MA. Her research interests are lean operations, experiential learning, just-in-time (JIT), and assessmentoflearningandquality.
Dr.BoualemKezimisanassociateprofessorof management at Merrimack College, NorthAndo- ver,MA.Dr.Kezim’sresearchinterestsareBayes-iantheory,statistics,andexperientiallearning.
Correspondence concerning this article should be addressed to Dr. Susan E. Pariseau, Girard SchoolofBusinessandInternationalCommerce, Merrimack College, 315 Turnpike Street, North Andover,MA01845.
E-mail:susan.pariseau@merrimack.edu
REFERENCES
Birnbaum,A.(1971).Aperspectiveforstrength-eningscholarshipinstatistics. AmericanStatis-tician,25,14–17.
Conners,F.,Mccown,S.,&Roskos-Ewoldsen,B. (1998). Unique challenges in teaching under-graduatestatistics.TeachingofPsychology,25, 40–42.
Hakeem,S.(2001).Effectofexperientiallearning instatistics.JournalofEducationforBusiness, 77,95–98.
Harrington,C.,&Schibik,T.(2004).Methodsfor maximizing student engagement in the intro-ductory business statistics course: A review.
Journal of American Academy of Business, Cambridge,4(1/2),360–364.
Jamison, M. (1998, November).Building con- ceptualunderstandingofresearchandstatisti-cal methods through student projects. Paper presented at the annual meeting of the Mid-SouthEducationalResearchAssociation,New Orleans,LA.AbstractretrievedSeptember24, 2004, from ERIC database. (ERIC Document ReproductionServiceNo.ED427067) Johnson,D.,Johnson,R.,&Holubec,E.(1990).
Circles of learning: Cooperation in the class-room.Edina,MN:InteractionBookCompany. Kolb,D.A.(1984).
Experientiallearning:Experi-enceasasourceoflearninganddevelopment. UpperSaddleRiver,NJ:Prentice-Hall. Kvam, P. (2000). The effect of active learning
methods on student retention in engineering statistics.AmericanStatistician,54,136–140. Meyers,C.,&Jones,T.(1993).Promotingactive
learning:Strategiesforthecollegeclassroom. SanFrancisco:Jossey-Bass.
Rosenthal, J. (1995). Active learning strategies in advanced mathematics classes.Studies in HigherEducation,20,223–228.