Issuu on Google+

American Educational Research Association

Teaching Problem Solving Through Computer Simulations Author(s): John Woodward, Douglas Carnine, Russell Gersten Source: American Educational Research Journal, Vol. 25, No. 1 (Spring, 1988), pp. 72-86 Published by: American Educational Research Association Stable URL: Accessed: 01/02/2010 13:01 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. 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

American Educational Research Association is collaborating with JSTOR to digitize, preserve and extend access to American Educational Research Journal.

AmericanEducationalResearch Journal Spring 1988, Vol. 25, No. 1, pp. 72-86

Teaching Problem Solving ThroughComputer Simulations John Woodward Douglas Carnine and Russell Gersten University of Oregon Thepurposeof this study was to investigatethe effectivenessof a computer simulation in enhancingstudentlearningin a unit of health. This study involved30 mildly handicappedstudentswho wererandomlyassignedto one of two instructionalgroups:(a) structuredteachingand (b) a combination of structuredteachingand a computersimulation.Studentswere taughtfor 12 consecutivedays. Thefirst half of each day's lessonfollowed effectiveteachingpractices.Thesecondhalf was (a) a computersimulation or (b) traditionalenrichmentand applicationexercises.Followingthe last day of instruction,studentsweretestedon basicfacts, concepts,and health problem-solvingskills. Posttestresultsindicatedsignificantdifferenceson basicfacts and conceptsthat werereinforcedby the simulation(p < .01). Thesedifferenceswereretainedon a maintenancetest given 2 weeksafter theposttest.Themostsignificantdifferenceswereon the testthatmeasured problem-solvingskills (p < .001). The resultsshow that the combination of structuredteachingand a computersimulationwas effectivein teaching not onlyfactual-levelknowledge,but highercognitiveskills as well. Secondary students spend a considerable amount of their time completing application-oriented activities. In performing these tasks, students are asked to make a variety of inferences about a subject area by prudently using facts, concepts, and strategies or problem-solving skills. Unfortunately, it is easier just to teach students rote memory information and procedural knowledge (i.e., the literal algorithms used in solving a problem) than the comprehension and problem-solving skills called for in many application items (Doyle, 1983). Some writers (Budoff, Thormann, & Gras, 1984; Cherryholmes, 1966; Cruickshank & Tefler, 1980; Greenblat & Duke, 1975) have suggested that 72

Teaching Problem Solving

one way to enhance these kinds of cognitive skills is througheducational simulations. Simulations are thought to increase student participation (Boocock & Schild, 1968; Farran, 1968; Stembler, 1975) and allow lowachievingstudentsmuch-neededpracticein applyingwhat they'velearned to new situations(Cohen & Bradley, 1978). Yet the results of researchon educational simulations, on the whole, have been discouraging.After an extensivereviewof studiesconductedin the early 1960s, Cherryholmes(1966) found that the effectsof simulations wereno greaterthan for conventionalinstruction.Studentsneitherlearned more facts and concepts than they did in conventional instruction nor showed the anticipatedincreasesin criticalthinkingand problemsolving. A more recent review of simulation researchby Pierfy (1977) reached conclusions similar to those of Cherryholmes.At best, the results have been mixed regardingthe effects of educationalsimulations(Bredemeier & Greenblat, 1981; DeNike, 1976; Greenblat, 1973; Jackson, 1979; Livingston, 1970;McHenry, 1969;McKenzie& Padilla,1984;Pierfy).Current researchin computer simulationsand other computerenrichmentactivities, although limited, also indicates little support for these techniques (Bangert-Drowns,Kulik, & Kulik, 1985;Waugh, 1986). Much of the researchon simulationshas been plaguedby fundamental weaknessesin researchdesign. Severalinterventionshave been much too brief, usually with only one play of the simulation (Boocock & Schild, 1968;Fletcher, 1971; Pierfy, 1977). Quite a few studies(Brenestuhl,1975; Emery & Enger, 1972; Fennessey, Livingston, Edwards, Kidder, & Nafziger, 1975; Livingston, 1970) used rather crude quasi-experimental designsinvolving intact classes'being assignedto treatmentson a nonrandom basis. In some cases, the intent of the simulationgames, and hence the researchhypotheses,were poorly formulated(Williams, 1980). Finally, there are problemswith criterionmeasuresused in many simulation studies.Of the 22 studiesreviewedby Pierfy(1977), virtuallyall of them wereinvestigator-developed criterionmeasures,with verylittle detail about the constructionof the measuresor their reliabilityand validity.By failing to analyze thoroughlythe instructionalgoals of simulations,some researchersdid not design measuresto capture everythingtaught by the games (Megarry,1979). The problems with simulation research,however, go well beyond research design and instrumentationproblems. Fletcher(1971) noted that many of the simulationsused in the researchhad never been field tested and thus were of unknown quality. He cited the great variation in the quality of the games used (in terms of complexity,levels of sophistication, and interactionamong participants)as a major source of the weak results reported. The studyreportedhere attemptedto remedymany of the problemsand issues cited above. The obvious problemsregardingthe length of interven73

Woodward,Carnine, and Gersten

tion and random assignmentwere relativelyeasy to avoid. Also, validity and reliabilityissues of the measuresused are addressed.We attemptedto create an instrument that reflected the problem-solvingskills actually taught in the simulation. The intent and instructionalgoals of the simulation Health Ways were formalized into a set of specific strategies,all monitored by a meta-strategy.Figure 1 representsthe problem-solving strategiesused in the simulation. The simulationitself was field testedin severaljunior high school classes with students of varied abilities. Problematicfeatures of the simulation were refinedaftereach field test. Othercomponentsof the intervention,to be discussedin the "Methods"section, were pilot tested beforethe study. The present study also attemptedto addressa curious featureof most past simulation studies. Virtuallyall studies have directlycontrastedsim74

Teaching Problem Solving

ulations to conventionalteaching methods. On only a few occasions have the combination of conventional instructionand a simulationbeen comparedto a conventional method alone. This is understandablewith noncomputersimulations,as time considerations(i.e., the number of days or weeks needed to implement a simulation fully) usually prohibiteda combined intervention. With computer simulations, the situation is different. The effects of different variablescan be demonstratedquickly (Doob, 1972), and less instructional time is required to demonstrate causal relationshipsand essentialconcepts. Unfortunately,very few experimentalstudiesinvolving educationalcomputersimulationshavebeen documented.None compared conventionalinstructionand a computersimulationwithjust conventional instruction. A final aspectsets the presentstudy apartfrom earlierresearch.Conventional instructionwas deliveredaccordingto principlesderivedfrom recent teachereffectivenessresearch(cf. Brophy& Good, 1986). Teachermodeling, high rates of teacher-studentinteraction,guided practive,and structured seatwork were used in teaching basic facts and concepts to both groupsin this study. This enabledus to bettergaugethe additionaleffects the simulationhad on basic facts and concepts. A secondarypurposeof the study was to determinethe extent to which a simulation could assist secondarylearningdisabled students in special education classroomsin acquiringfactual informationand problem-solving skills.In additionto evaluatingthese students'performancein absolute terms, their performancewas comparedto that of their nonhandicapped age-mateswho were enrolledin health educationcourses. Method Subjects All of the subjectswere learningdisabled high school students eligible for special education services by federal and local standards.Because of the reading requirementsof the simulation, students who scored lower than the sixth grade reading level, as measured by the Metropolitan AchievementTest (1970), were excludedas subjects.The 30 studentswho met these criteriaparticipatedin the study; an equal number of students in three classes were randomlyassignedto either the conventionalor the simulationcondition. Materials Health Ways is a commercialsoftwareprogramdevelopedby Carnine, Lang, and Wong (1985) for the Apple II and IBM PC computers.As a simulation,Health Waysrequireslearnersto manipulateseveralvariables (e.g., life-threateningdiseases,stresslevels) in orderto achieve an optimal life expectancy.Learnersare presentedwith a basic healthprofilethat can 75

Woodward, Carnine, and Gersten

be examined furtherby selecting various options. An example of such a profileis presentedin Appendix 1. For example,learnerscan inquireabout eating habitsby selectingthe nutritionoption. By doing this, they can see how much cholesterol, sugar, salts, and so forth, the profile character consumes.Learnersmust make changesin the most life-threateninghealth habits(e.g., heavy smoking)and control other variablesto be successfulat a Health Waysgame. We also developed the Health Ways SupplementaryCurriculum,an accompanyingwritten curriculum(Woodward& Gurney, 1985) that extended information presented in Health Ways and the original Health Waysteachersguide. Informationwas taken directlyfrom two widelyused junior high school health textbooks.All the informationwas rewrittento control for vocabularyand amount of new information.Clippingsfrom newspapers,news magazines,journal articles,and health pamphletswere also used in the supplementarycurriculum.Even though the supplementary materials relied heavily on texts and other sources, considerable preparationtime was required.Researchersestimatethat it took approximately 70 hours to create the modified curriculum.The readinglevel of the curriculum,as determined by the Fry (1977) ReadabilityTest, was approximatelysixth grade. Procedures All students were instructedfor 40 minutes per day for 12 days. The lesson consisted of two parts. The first part, called "structuredteaching," wasidenticalfor subjectsin both conditions.A structuredteachingmethod, following the model proposed by Rosenshine and Stevens (1984) and Brophyand Good (1986), was used for this segment. Structuredteaching.Instructionwas conductedin a largegroupof 12 to 15 students for this part of each lesson. Instructionbegan by reviewing essentialinformationfrom previouslessons. Studentswere then presented with a list of vocabularywords, which were essential to the day's lesson (e.g., cholesterol,diabetes).For the next 15 minutes, each student independently read the two to three pages of text for that day's lesson and answeredwritten comprehension questions. The teacher then discussed the answers with the group and presented a series of review questions coveringthe main points of the lesson. At the end of the initial instruction, students separatedinto two groups, one that worked with the computer simulation (the Simulation Group) and one that worked on traditional applicationactivities(the ConventionalGroup). The Conventional Group worked in the resource room under the supervisionof the resource room teacher, who presentedthese students with a variety of applicationor review activities. These exercises,typical of a highschool healtheducationclass,werereviewedfor representativeness by two health teachersat the high school wherethe study took place. For 76

Teaching Problem Solving

example, the ConventionalGroup studentskept track of their diets for 3 days and analyzedtheir cholesterollevels. Otherexercisesincludedanalyzing one-paragraphprofiles of different individuals and diagnosing poor health habits. Studentscompletedthe exercisesduringthe last 20 minutes of the period. Simulation Group students were taught in a computer lab, with each student workingindividuallyat a microcomputer.Studentsin this condition worked on the Health Ways simulation for 20 minutes each day. Instructionover the 12-day period was broken into three phases: initial modeling of the simulation tutorial and one simulation game (3 days), guided practice on three simulation games (2 days), and independent practicewith individualfeedbackfrom the instructor(7 days). In the initial modelingphase, the instructormodeled each component of an effective strategyfor playing the Health Ways games. The teacher modeled workingon the most importanthealth problemsfirst,to control stressimmediatelyif it rose to a level that was too high, and to be sure to maintain changes by using the maintenanceoption. Each component of this strategywas modeled in isolation and then integratedin laterinstruction. The most importantinstructionin the initialmodelingphaseinvolved the prioritizingof health problems. Students were first taught to look at current disease and hereditaryinformation. A profile's indication of a heredity of hypertensionmeant that the student should first look at the individual'sdiet and check the level of salt consumption. If it was high, this led to a correlatedchange (i.e., a heredityof hypertensionimplies the need for a low salt diet, thus the student must change the character'ssalt intake).This habit was the firstto be changed.The remaininghabitswere identifiedin their orderof priority(i.e., the second and third most important habits). In the guidedpracticephase of instruction,studentswere brieflytaught in a groupwith one microcomputer.They were shown the initial screenin a Health Ways game profile, and individual students were asked to prioritize the first three health problems or habits and their correlated changes.After a correctidentificationwas made, the studentswere shown anotherprofile.Afterthree or four profiles,each studentwent to his or her computerin the lab and playedthe Health Waysgames. The independentpractice phase allowed the students to play Health Ways games continuously for 20 minutes. The teachercirculatedamong the students, observingand commenting on a student's"play"of a game (i.e., how well he or she employed the strategytaught in the modeling phase of instruction). All teachingwas done by the researcherand a certifiedspecialeducation teacher. Assignment of teachersto treatmentwas counterbalanced,with the researcherand the teacherchanginggroupshalfwaythroughthe experiment. This was done to control for the effects of the teacher,a common 77

Woodward,Carnine,and Gersten

problem that has flawed many computer-basedinstructionstudies (Bangert-Drownset al., 1985). The amount of total instructionaltime was controlledin this study;both groupsreceivedthe same amount of teaching and independentwork. Measures Studentswere assessed 1 day, 2 days, and 2 weeks followinginstruction. On the first day, acquisition of basic facts and concepts taught in the curriculumwas measuredby the Nutritionand DiseaseTest. The Nutrition and DiseaseTest was a 30-itemtest designedto measurestudents'retention of the importantinformationcontained in the writtencurriculum.Questions on the test were fill-in-the-blank,usuallyrequiringonly one- or twoword answers.The first 20 questions were solely from the written curriculum. The remaining 10 were questions on materialthat appearsin both the writtencurriculumand the Health Ways simulation.Internalconsistency reliability(coefficientalpha)of this measureis .84 basedon a sample of 42 students. The Nutrition and Disease Test was given again 2 weeks afterthe instructionas a retentionmeasure. The second measure, the Health Diagnosis Test, was a set of three written profilesadministered2 days after the instruction.This test measured the student's ability to detect important health problems facing an individual,identify and change relatedhealth habits,and control stressas it increasedas a resultof the healthchanges.Centralto the HealthDiagnosis Test was prioritizinghealth problems.For example, the test measuredthe student'sability not only to identify but also to orderthe health problems in terms of their importance to the individual's longevity. The Health Diagnosis Test has a test-retestreliabilityof .81. Appendix 2 presentsan example of one of the test's three profiles. Scoringprocedures.For the Nutrition and Disease Test, only answers containedin the writtencurriculumor acceptablesynonymswere consideredcorrect.Subscalescoreswereobtainedfor two sections:(a) those items reinforcedand (b) those not reinforcedby the Health Wayssimulation. Specialproceduresfor scoringthe HealthDiagnosisTest weredeveloped. Three differentareaswere assessedin the measure:(a) identifyingimportant health problems, regardlessof order, and making the appropriate correlatedchange, (b) the ability to prioritize health problems, and (c) attendingto stresswhen it was at a high level. Current health facts and statistics were used to develop criteria for correct prioritizing.The criteria state that the learner should attend to current disease and hereditaryinformation in determiningwhich health habitsare most detrimental,hence, which habits need to be changedfirst. A strictcriterionand a moderatecriterionwere used to measurestudents' ability to prioritize.To score at the strictcriterion,a studentmust change the three most importanthealthproblemsin a specificorder(i.e., the habit 78

TeachingProblem Solving

associated with the current disease first, the one associated with the hereditarydisease second, and the remaining detrimental habit third). These criteriawere establishedby a committee consisting of the experimenter,a professionalhealtheducator,and two specialeducationresearchers. Studentswho scoredat the moderatecriterionsimply changed,in any order,the habits associatedwith the currentdisease and hereditaryproblems within the firstthree changes. In addition, both tests were also given to a randomsample of nonhandicappedhigh school studentsfrom healthclasses.Tests weregiven to 10th, 11th, and 12th gradersin introductoryand advancedhealthclasses.Their scoreswere comparedwith those of the two groupsthat participatedin the study. Results The Nutritionand Disease Test A 2 x 2 (treatmentby time of test) analysis of variancewith repeated measures on one factor was performedon the total number of correct responseson the Nutrition and Disease Test. Table 1 providesthe descriptive statisticsfor the correctnumber of responsesfor the post- and maintenance tests for each group. Means were also convertedto a percentage correct. The analysis shows a significantmain effect for instructionalmethod; 5.30, p < .03. For both treatment groups there was a significant in scores from post to maintenance test; F(1,26) = 16.23, p < .001. No drop F(l,28) =

significantinteractionwas found. The simulation had a significanteffect on masteryof key concepts in the unit; this effect was maintainedover a 2-week period. Subscalesanalyses. The 30-item test was brokeninto two subscales:(a) items reinforcedby the Health Ways simulation and (b) items taught in the curriculum and not reinforced by the simulation. Separate 2 x 2 ANOVAs with repeatedmeasureswere performedon each subscale.The effect on items reinforced by Health Ways was significant; F(1,28)= 40.02, TABLE 1 Means (M), medians(Mdn), and standarddeviations(SD) of numberof total correctanswerson the Nutritionand Disease Test Posttest Instructionalgroup

Mean Mdn SD %cor- N M Mdn SD rect 15 22.00 21.5 3.72 73.3 15 19.97 20.5 5.08 15 17.93 18.5 5.86 59.7 15 15.47 16.5 5.44 N

Simulation Conventional



Mean %correct 66.5 51.6 79

Woodward,Carnine,and Gersten p < .01. The effect for items not reinforced, however, was nonsignificant; F(1,28) = 3.73, p < .06. This demonstrates that the simulation was an effective vehicle for reviewing material that had already been presented in the written curriculum. Health Diagnosis Test Scores for the conventional and simulation groups were compared on (a) the total test score and (b) the total test score without stress as a factor. The reason for this was that the simulation group was explicitly taught the relationship between a health change and an increase in stress through Health Ways. Students in the conventional group were never taught about this relationship, thus it would be unlikely that these students would immediately control the stress level in the Diagnosis Test. Therefore, the factor of stress was removed from the analysis of the total test scores, which appears at the top of Table 2. The three essential problem-solving skills for the Health Diagnosis Test were independently compared: (a) prioritizing health habits, (b) stress management, and (c) identifying health problems and making correlated changes. The t tests demonstrate a significant difference between the two groups according to all analyses demonstrating that the intervention had a consistent, significant effect on problem-solving skills of the simulation students. The correlation between Metropolitan Achievement Test (MAT) Reading scores and total scores on the Health Diagnosis Test was nonsignificant (. 12), much weaker than the correlation of .44 between MAT Reading and the Nutrition and Disease Test, a more conventional academic measure of facts and concepts. The nonsignificant correlation suggests no relationship between traditional academic measures (such as a standardized achievement test) and the problem-solving skills measured by the Diagnosis Test. TABLE 2 Summaryoft testsfor the Diagnosis Test Simulation



M SD SD Total test score 27.7 6.2 12.47 4.9 Componentproblem-solvingskillsinvolved in the test

t 7.52

df 28

p <.001

Prioritizing alone








Stressmanagement Identifyinghealth problemsand making correlated changes
















TeachingProblem Solving

In addition to the generally superior performanceby the simulation students on the Diagnosis Test, there is some indication of a moderately strongrelationshipbetweentheirscoreson this test and theirunderstanding of the simulation.At the end of the third profile,the examinerasked each student to state his or her reasonsfor making the first, second, and third changeson the profile.In otherwords,the examineraskedwhy the changes weremade in the orderspecifiedby the student.Responsesto this question werecategorizedas (a) the studentguessed(didn'tknow, didn'tcare,didn't know why), (b) the student was working on health problems but in no apparentorder (i.e., no prioritizingstrategywas used), or (c) the student worked on the most important health problem first (i.e., some kind of prioritizingstrategywas used). The correlationbetween a student'sscore for this responseand his or her total test score on the DiagnosisTest was .69. This suggestsa moderatelystrong relationshipbetween the strategies that students thought they were using in the test and those that they actuallyused. SecondaryAnalyses:ComparisonWithNonhandicappedHigh School Students In a supplementalanalysis, a one-way analysis of variance (ANOVA) was used to comparethe test performanceof the conventionaland simulation groups with nonhandicappedstudents from regularhealth classes who did not participatein the study. The purposeof this quasi-experimental comparisonwas to extend the posttest analysis to students of a comparableage groupwho were also receivinghealthinstruction.Again,scores from each section of the Health Ways Nutrition and Disease Test and the Health Ways Diagnosis Test were analyzed.These resultsare presentedin Table 3. Total scoreon the DiagnosisTest showedsignificantdifferencesbetween the groups (F(2,42)= 27.36, p < .001). A Tukey post-hoc comparison indicatedsignificantdifferencesfavoringthe handicappedsimulationstudents in comparisonto regularclassroomstudents(p < .01). Therewas an equallysignificantdifferencefavoringthe regularclassroomstudentsover the learning disabled students in the conventional group (p < .01). The learning disabled students in the simulation group had problem-solving skills on the health profilessuperiorto those of nonhandicappedstudents in regularhealth classes. The nonhandicappedstudents, in turn, outperformedthe handicappedstudentsin the conventionalgroup. A significantdifferencealso appearedbetween the groups on the reinforced items on the Nutrition and Disease Test, F(2,42) = 5.35, p < .01. Tukey post-hoc comparison showed a significantdifferencebetween the special education simulation group and the two other groups (p < .05), favoringthe handicappedstudentstaughtby Health Ways.Differenceson the nonreinforcedsubscaleitems were nonsignificant. 81

Woodward, Carnine, and Gersten


andnonhandicapped Means(M)andstandarddeviations(SD)forhandicapped studentson thetwoacademicmeasures N M SD Nutritionand DiseaseTest:Total score Mildly handicappedstudentstaughtby: 15 Simulation 15 Conventional 15 students Nonhandicapped Nutritionand diseasetest: Items reinforcedby HealthWays Mildly handicappedstudentstaughtby: 15 Simulation 15 Conventional 15 students Nonhandicapped HealthWays DiagnosisTest: Total score Mildly handicappedstudents 15 Simulation 15 Conventional 15 Nonhandicappedstudents:

22.00 17.93 19.47

3.72 5.86 4.94

7.33 5.60 5.53

1.35 2.20 1.46

27.73 12.47 18.07

5.89 4.88 6.03

Discussion The Health Ways simulationwas an effectivetool in teachingmaterial not easily taught by traditional means. In this study, we used direct instructiontechniquesin the initial phase of instructionto teach material rewrittenfrom widely used health textbooks. The resultsindicate the use of computer simulations can effectively complement traditionalinstruction. In addition to providingthe experimentalstudents with a simulation, we taughtan explicit strategythat enabledthem to be successful(see Figure 1). The resultssupportthe view that a structuredapproachin simulations, one where learners'tactics are specifiedand guided, does have significant educationaleffects. The analyses indicated that the students in the simulation group performedat a significantlyhigherlevel on health fact and conceptitems that the simulation reviewed. The simulation also had a significanteffect in developing problem-solvingskills in health. Simulation students were significantlybetter at prioritizingspecific health habits, ones that needed to change in the game's profile characterin order to improve his health and longevity. The superior performance by the learning disabled students in the simulationgroupover nonhandicappedstudentsfromregularhealthclasses suggeststhe extent to which explicit strategyinstructioncan be successful in teachingproblem-solvingskills.The two nonhandicappedstudentswho 82

TeachingProblem Solving

had the highest scores on the Diagnosis Test articulateda prioritizing strategy comparable to that given by several of the special education students. Thus, many of the special education students in the simulation group showed a conscious awarenessof the strategiesthat they were using, as did the two untaught,nonhandicappedstudents,who may haveachieved their awarenessin a more intuitive manner. Both the instruction in basic health concepts and the explicit strategy for the simulationwere based on instructionaldesign principlesdescribed in the Theory of Instruction(Engelmann & Carnine, 1982). Instruction began with models by the teacher of both successful and unsuccessful strategiesthat could be used with Health Ways. Next, studentspracticed the strategyover a range of profiles with feedbackfrom the researchers. Gradually,the explicit remindersor prompts about steps in the strategy were removed, and the simulation students worked on Health Ways profilesindependently. The simulation itself was also designed to foster the acquisition of a strategy.The Health Wayssimulationwas precededby a tutorialcontaining threesimplerversionsof the simulationprofiles,each one slightlymore complex than the precedingone. This gradualprogressionfrom simple to complex allowed aspects of the overall strategy to be introduced and practiced one at a time. The contribution of detailed, explicit strategy instruction(i.e., the methods describedby Engelmann& Carine, 1982, and Palincsar& Brown, 1982) has been investigatedwith handicapped students in reading comprehension (Carnine & Kinder, 1985), content area instruction in science (Darch, Carnine, & Kameenui, in press) and logical reasoning(Collins, 1984). Some researchon computersimulations suggeststhat when contraryproceduresare followed(e.g., informalinstruction, particularlywhereclearstrategiesand correctivefeedbackare absent), studentlearningis insignificant(Waugh, 1986). The presentresultsindicatethat simulationscombined with instruction in strategies for successful use of the simulation can contribute to a student'slearningof both factual informationand problem-solvingskills. However, the results say nothing about the use of computer simulations as "stand alone" activities. Since the two treatments differ in several respects,we cannot isolatea specificvariablethat accountedfor the results. That 15 learning disabled students exhibited problem-solvingskills and that this was true for only two nonhandicappedstudentsunderscoresthe potentialof combininginstructionand simulations.Futureresearchcould addressthe simulation alone by comparingHealth Waysto Health Ways accompaniedby explicit strategyinstruction.These findings would help articulatethe context in which computer simulationscan be of the most benefit to students. A separateline of researchwill continue to evaluate the various components of strategyinstruction, whether mediated by a teacheror by a computer. 83

Woodward,Carnine, and Gersten APPENDIX 1 Health Wayssimulationprofile Ages: Year 50

Today's Week 00

Day 0





Will power = 180 Stress = 35

Name: James Heredity:lung cancer Diseases:no currentailments 1. Weight:20 pounds overweight 2. Tobacco:moderatesmoker 3. Alcohol:nondrinker 4. Exercise:moderate5 times a week 5. Nutrition:see submenu 6. Lifestyle:see submenu 7. Maintenancemenu 8. Stressreductionmenu 9. Help APPENDIX2 Healthfacts on Julie Hill Age: 34 Heredity:diabetes Currentdiseases:liver disease 1. Weightand diet: 7 pounds overweight a. Eats breakfasts b. Eats a lot of food with cholesterol c. Eatsa lot of empty-caloriesweets d. Eats very little food with sodium e. Drinksvery few beverageswith caffeine f. Eatsa lot of food with fiber g. Eatsbalancedmeals 2. Tobacco:nonsmoker 3. Alcohol:light drinker(3 drinksa week) 4. Exercise:exercises5 times each week 5. Stress:averagestressin her life References & Bangert-Drowns,R., Kulik,J., Kulik,C. (1985). Effectivenessof computer-based education in secondaryschools. Journalof ComputerBased Instruction,12(3), 59-68. 84

TeachingProblem Solving Boocock, S., & Schild, E. (Eds.). (1968). Simulationgames in learning.Beverly Hills, CA: Sage. Bredemeier,M., & Greenblat,C. (1981). The educationaleffectivenessof simulation games:A synthesisof findings.Simulation& Games, 12, 307-332. Brenestuhl,D. (1975). Cognitivevs. affectivegains in computersimulation.Simulation& Games, 6, 303-311. Brophy,J., & Good, T. (1986). Teacherbehaviorand studentachievement.In M. Wittrock(Ed.), Third handbookof researchon teaching (pp. 328-375). New York:Macmillan. Budoff,M., Thormann,J., & Gras,A. (1984). Microcomputersand special education. Cambridge,MA: Brookline. Carnine, D., & Kinder, D. (1985). Teaching low performingstudents to apply generativeand schemastrategiesto narrativeand expositorymaterial.Remedial and SpecialEducation,6(1), 20-30. Carine, D., Lang, D., & Wong, L. (1985). Health Ways.Unpublishedcomputer program. Cherryholmes,C. H. (1966). Some currentresearchon effectivenessof educational simulations:Implicationsfor alternativestrategies.AmericanBehavioralScientist, 10, 4-7. Cohen, R., & Bradley,R. (1978). The use of simulationgamesto createrealworlds for specialeducationstudents.ReadingImprovement,15, 275-278. Collins, M. (1984). The effectivenessof computer-deliveredcorrectionprocedures on low-performingsecondarystudents'reasoningskills. Unpublisheddoctoral dissertation,Universityof Oregon,Eugene. Cruickshank,D., & Tefler, R. (1980). Classroomgames and simulations.Theory Into Practice,19, 75-80. Darch,C., Carnine,D., & Kameenui,E. (1986). The role of graphicorganizersand social structurein content area instruction.Journalof ReadingBehavior,18(4), 275-295. DeNike, L. (1976). An exploratorystudyof the relationshipof educationalcognitive style to learningfrom simulationgames.Simulation& Games, 7, 65-74. Doob, P. (1972). Prospectsfor simulationgamingin healthplanningand consumer healtheducation.(ERICDocument ReproductionServiceNo. ED 070 596) Doyle, W. (1983). Academicwork.Reviewof EducationalResearch,53, 159-199. Emery,E., & Enger,T. (1972). Computergamingand learningin an introductory economics course.Journalof EconomicsEducation,3, 77-85. Engelmann,S., & Carine, D. (1982). Theoryof instruction:Principlesand applications.New York:Irvington. Farran,D. (1968). Competitionand learningfor underachievers.In S. Boocock & E. Schild(Eds.),Simulationgames in learning(pp. 191-204). BeverlyHills, CA: Sage. Fennessey, G., Livingston, S., Edwards,K., Kidder, S., & Nafziger, A. (1975). Simulation,gaming,and conventionalinstruction:An experimentalcomparison. Simulation& Games,6, 288-302. Fletcher,J. (1971). The effectivenessof simulationgamesas learningenvironments. Simulation& Games,2, 425-554. Fry, E. (1977). Fry's ReadabilityScale. Providence,RI: Jamestown. 85

Woodward,Carnine,and Gersten Greenblat,C. (1973). Teaching with simulation games: A review of claims and evidence. TeachingSociology,1, 62-83. Greenblat,C., & Duke, R. (1975). Gaming simulation:Rationale, design, and application.New York:Halsted. Jackson,M. (1979). An antipodeanevaluationof simulationin teaching.Simulation & Games, 10, 99-138. Livingston,S. (1970). Simulationgames and attitudechanges.Attitudestowardthe poor (ReportNo. 63). Baltimore,MD: Johns HopkinsUniversity,Centerfor the Studyof Social Organizationof Schools. McHenry,W. (1969). A study of the use of the Life CareerGuidance Game in junior high school group guidance. Unpublisheddoctoral dissertation,George WashingtonUniversity,Washington,DC. McKenzie, D., & Padilla, M. (1984, April). Effects of laboratoryactivities and writtensimulationson the acquisitionof graphingskillsby eighthgradestudents. Paper presentedat the annual meeting of the National Associationof Science Teaching,New Orleans. Megarry,J. (1979). Monitoring.Simulation/Gamesfor Learning,9, 170-177. Palincsar,B., & Brown,A. (1982). Inducingstrategiclearningfrom texts by means of informedself-controltraining.Topicsin Learningand LearningDisabilities, 2(1), 1-17. Pierfy, D. (1977). Comparativesimulationgame research:Stumblingblocks and steppingstones. Simulation& Games,8, 255-268. Rosenshine, B., & Stevens, R. (1984). Classroominstructionin reading. In D. Pearson (Ed.), Handbook of researchon teaching (pp. 745-798). New York: Longman. Stembler,W. (1975). Cognitiveeffectsof a programmedsimulation.Simulation& Games, 6, 392-403. Waugh,M. (1986, June).The effectof teacherinvolvementon studentperformance in a computer-basedscience simulation.Paperpresentedat the National Education ComputingConference,San Diego. Williams,R. (1980). Attitudechangeand simulationgames.Simulation& Games, 11, 177-196. Woodward,J., & Gurey, D. (1985) Health Ways supplementarycurriculum. Unpublishedcurriculum,Universityof Oregon,Eugene. Authors JOHN P. WOODWARD, Research Associate, University of Oregon, Project Follow Through, 1751 Alder,Eugene,OR 97403. Specialization:specialeducation technology. DOUGLAS W. CARNINE, Associate Professor,University of Oregon, Project Follow Through, 1751 Alder, Eugene, OR 97403. Specialization:instructional design. RUSSELL M. GERSTEN, Associate Professor,University of Oregon, Project FollowThrough,1751 Alder,Eugene,OR 97403. Specialization:researchdesign.


Teaching Problem Solving Through Computer Simulations