THE LEARNING PYRAMID:DOES IT POINT TEACHERS IN THE RIGHT DIRECTION?

JAMES P. LALLEYROBERT H . MILLER

Education DepartmentD'Youville College

This paper raises serious questions about the reliability of thelearning pyramid as a guide to retention among students. Thepyramid suggests that certain teaching methods are connectedwith a corresponding hierarchy of student retention. No specificcredible research was uncovered to support the pyramid, whichis loosely associated with the theory proposed by the well-respected researcher, Edgar Dale. Dale is credited with creatingthe Cone of Experience in 1946. The Cone was designed to rep-resent the importance of altering teaching methods in relation tostudent background knowledge: it suggests a continuum of meth-ods not a hierarchy. While no credible research was uncovered tosupport the pyramid, clear research on retention was discoveredregarding the importance of each of the pyramid levels: each ofthe methods identified by the pyramid re.sulted in retention, withnone being consistently superior to the others and all being effec-tive in certain contexts. A key conclusion from the literaturereviewed rests with the critical importance of the teacher as aknowledgeable decision maker for choosing instructional meth-ods.

The Learning Pyramid (see Figure 2) isan often-cited guide for teachers to use fordesigning effective instruction. It can befound in books (Sousa, 2001; Danielson,2002; Drewes & Milligan, 2003), articlesin refereed journals {Wood. 2004; Brueck-ner & MacPherson, 2004; Darmer et al.,2004; DeKanter. 2005) and in teacherresources {Boulmetis, 2003; Hershman &McDonald, 2003). Further, a recent inter-net search using Google® returned 12,200hits, of which 452 were from cites withthe generic top level domain .org, 313 from.edu, and 30 from .gov. While it appears anintuitive model and an implied compre-hensive research summary, the logic of themodel, as well as its research base, have

been questioned (Thalheimer, 2005).Therefore, it is our intention to examinethe following: the source of the generalstructure of the pyratnid. Dale's Cone ofExperience; available research on reten-tion from the methods identified by thepyramid; and consider the relationship(s)among the methods.

Dale's Cone of ExperienceEdgar Dale was a protninent educator

and nationally known scholar regardingthe use of media in instruction. He had alaudable career at the Ohio State Univer-sity and its media center is named in hishonor. In 1946 Dale published the first ofthree editions of his infiuential

64

Visual Methods in Teaching. The puiposeof the text was to delineate:

the use of audio-visual materials inteaching - materials that do notdepend primarily upon reading toconvey their meaning. It is basedupon the principle that all teachingcan he greatly improved by the useof such materials because they canhelp make the learning experiencememorable...this central idea has,of course, certain limits. We do notmean that sensory materials must beintroduced into every teaching situ-

Examining the Learning .../ 65

ation. Nor do we suggest that teach-ers scrap all procedures that do notinvolve a variety of audio-visualmethods (Dale, 1954, p. 3, italics inoriginal)

The 1946 edition included the debut ofhis Cone of Experience (See Figure 1). Thecone was Dale's attempt to organize vari-ous types of experience according to theirlevels of abstractness, with Direct, Pur-poseful Experiences being the at leastabstract (or most concrete) end ofthe con-tinuutn, and Verbal Symbols, the most

DALE'S CONE OF EXPERIENCE

i

^^^^^^^^^^^^^m RocordlnBt - Radio ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ |^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ V Still Plcturai ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ H

^^^^^^^^^^^M Motion Picture! ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ H

^^^^^^^^m TelevlalDn ^ ^ ^ ^ ^ ^ ^ ^ H

^ ^ ^ ^ ^ V Exhibits ^ ^ ^ ^ ^ H

^ ^ ^ V Fiold Trtp> ^ ^ ^ ^ 1

^ ^ ^ V OBmonstraHons ^ ^ ^ ^ |

^ V Dramatized Experlencas ^ ^ |

F Contrived Expenances V

Direci • Purposeful Exp«rienc«s \

Figure I. Dale\s Cone of Experience, 1954

66 / Education Vol. 128No.1

abstract, at the other. However, he notesthat there will clearly be overlap amongthe levels. Before proceeding, we shouldnote that that there were modifications tothe cone to accommodate emerging tech-nologies (e.g.., television was added in1954) in the three editions of his text (Dale,1946; 1954; 1969)). However, the gener-al construct of the cone, progressing fromthe concrete to the abstract, remainedintact.

Dale describes Direct, Purposeful Expe-riences as those in which "you have directparticipation, with responsibility in theoutcome" {{954, p. 42, italics in original).He asserts that our most vivid memories areoften the result of direct experiences. Asan example. Dale offers performing a lab-oratory experiment as direct experience.However, he goes on to note that:

Life cannot, of course, be livedexclusively on this direct, concrete,sensory level. Whenever we remem-ber something we have experienced,we have begun to abstract. Even ourearliest experiences involve somedegree of abstraction. As very youngchildren we learn to talk about thedoll or the cat or the man that is notphysically present, and thus ourdirect, concrete experience becomesassociated with abstractions. (1954,p. 44, italics in original)

Dale prefaced his presentation of thecone with the reminder that the gapbetween the least and the most abstract isoften quite narrow, noting that even thewords of small children, for whom we oftenadvocate concrete learning experiences.

are abstractions. In Dale's words, the coneserves as a "visual metaphor of learningexperiences" and that "you will make adangerous mistake, however, if you regardthe bands on the cone as rigid, inflexibledivisions" (1954, p.42). Also, he notes that"increasing abstractness does not meanincreasing difficulty" (1954, p.42). Final-ly, in presenting a notion that still appliesto schools today, instruction would typi-cally improve if it included opportunitiesfor Direct Experience. To provide balance,however, he goes on to cite Dewey (1916)who reminds us that while direct experi-ences are important to learning they arealso, by their very nature, limiting: learn-ing about a geographic region in sensorystimulating manner (i.e., being there)would indeed be a valuable experience.However, that experience would be veryrestricted in terms of developing an under-standing of geography, its multiplecomponents, and numerous regions (bet-ter learned from a map)—although it islikely a good beginning. Clearly, Dale rec-ognized the importance of having a varietyof learning experiences and the comple-mentary nature of those experiences.

The Learning PyramidAt some point. Dale's original cone of

experience was transformed into the ubiq-uitous Learning Pyramid (Figure 2). To bemore accurate, we should say LearningPyramids, given that it has taken on morethan one form over the years (it is alsocommon to find the pyramid labeled asDale's Cone). The pyramid replaces levelsof experience with instructional methods,with lecture occupying the peak of thepyramid, followed by reading, etc. with

Examining the Learning .../ 67

the lowest level being teaching others, andindicates various levels of retention asso-ciated with each method. There are.however, more levels in the cone than aretypically found in the pyramid (i.e., thereis not a one-to-one correspondencebetween the levels of the cone and the lev-els of the pyramid). As noted, there arevariations of the pyramid that include adifferent hierarchy of methods or differ-

ent methods, as well as different levels ofretention ascribed to these methods. Forexample, the most common pyramid (e.g.,Sousa, 2001) identifies lecture at the topof the pyramid and indicates that its cor-responding level of retention at fivepercent, with reading falling below it witha retention level often percent. In contrast,the National Training Laboratories, for-merly of Bethel Maine and currently

LearningPyramid

(Average LearningRetention Rates)

National Training Laboratories, Bethel Maine

Figure 2: A Typical Presentation of the Learning Pyramid

68 / Education VoL 128 No.l

located in Alexandria Virginia, provides apyramid that locates reading at the top witha retention level of ten percent, followedby hearing with a retention level of 20 per-cent.

Additional information regarding thepyramid raises questions of credibility. TheNational Training Laboratories, in responseto an email request from a member of theAcademic Computing Department at theCollege of Charleston in South Carolina,stated the following about the pyramid:

It was developed and used by NTLInstitute at our Bethel, Maine cam-pus in the early sixties when we werestill part of the National EducationAssociation's Adult Education Divi-

sion. Yes, we believe it to be accu-rate - but no, we no longer have -nor can we find - the originalresearch that supports the numbers.We get many inquiries every monthabout this - and many, many peoplehave searched for the originalresearch and have come up emptyhanded. We know that in 1954 a sim-ilar pyramid with slightly differentnumbers appeared on p. 43 of a bookcalled Audio-Visual Methods inTeaching, published by the EdgarDale Dryden Press in New York. Yetthe Learning Pyramid as sucb seemsto have been modified and alwayshas been attributed to NTL Institute.

PacifiCnrp CSS Train the Trainer

PEOPLE GEISERALLY REMEMBER

An important teaming principle, supported by extensive research, is thatpeopte teambest when they are activety involved in tht teaming process. The "iower down the cone'you go, the more you team and retain.

10% of whatthey read I Read

20H of whatthey hear / Hear a lecture

30% of whatthey see i

Looti at exhibits, mock-ups,diagrams, displays

50% of what they / Watch live demonstrations, videos orHEAR Ai\D SBE I movie, go on a site visit

70% of what theySAY or WRITE

9(1% of what they l~SAY AS THEY DO /ANACTIVITY /

Complete ivorksheets, manuals,discussion guides

Simulate a reai experience(practice, with coaching)

Do the reai thing

Adapted from .iudio-Visual Methods in TeachingEdgar Dale DryJen Pre.is, N.Y., 1954 p. 43

Figure 3. Pyramid Provided hy the National Training LaboratoriesNTL Institute for Applied Behavitirai Science, iOO N. Lee Street, Suite 300, Aiexandria,VA 22314. 1-800-777-5227

Examining the Learning .../ 69

To summarize the numbers {whichsometimes get cited differently)learners retain approximately:

90% of what they learn whenthey teach someone else/useimmediately.75% of what they learn whenthey practice what they learned.50% of what they learn whenengaged in a group discussion.30% of what they learn whenthey see a demonstration.20% of what they learn fromaudio-visual.10% of what they learn whenthey've learned from reading.5% of what they learn whenthey've learned from lecture.(February, 2003)

This same correspondence containedthe pyramid shown in Figure 3, indicatingdifferent levels of retention for differentactivities. While this inconsistency may bethe result of two separate models or sim-ple typographical errors, the source oftheresearch raises serious questions regard-ing the credibility of the pyramid. Thelaboratories' statement that the pyramidmight have "slightly different numbers"than Dale's original work is unfounded.An examination of all three editions ofDale's texts reveals that he never uses theword retention to describe the outcome ofa particular learning experience and, moreimportantly, he presents no numbers refer-ring to empirical research. His conclusionswere based on theory and personal obser-vations, not on research. Further, hecautioned that there is overlap among thelevels and we should not treat them as being

discrete. For example, one might encountervisual and verbal symbols when watchingtelevision, on a field trip, or observing ademonstration. Thus, these numbers appearto be speculation.

In addition to these inconsistencies frompublication to publication, the pyramid andits comparison of instructional methodsand respective levels of retention raise anumber of empirical issues. There is animplied assumption that tbese methodsbave been compared to one another in asystematic manner employing soundresearch methodologies. At a minimum,these empirical issues would include:

• That each of the methods, employed asan experimental treatment, was of thesame duration (e.g., a student's readingsession would last as long as an indi-vidual teaching or discussion session).

• That each of the methods would havebe conducted or supervised by the sameteacher or that multiple teachers wouldhave been matched in terms of educa-tion, teaching experience and subjectarea (e.g., the lecture being given bythe same teacher as the one leading thediscussion). Further, the teacher(s)should have been well versed in bothcontent and method.

• That the content to be learned with eachmethod would be the same, regardlessofthe method being employed.

• That the outcome measure(s), or depen-dent variable(s), was one measuringretention, the ability to recall or dosomething after a time delay (e.g., days,weeks or months), rather one tbat iscompleted immediately after treatment.

70 / Education Vol. 128 No.1

To conduct research in this manner withseven equivalent treatment groups wouldbe a daunting task indeed, and if such issueswere not addressed the assumptions regard-ing the relative effectiveness ofthe methodsare erroneous. This alone makes us skep-tical that such research was done. Further,inconsistencies in referencing, as well asvariation in purported percentages, height-en our skepticism. Finally, although Sousa(2001) reports that retention was measuredafter a 24-hour period (for which Sousaprovides no reference), many who promotethe value ofthe pyramid make no mentionofthe time between leaming and measuredretention.

Beyond these inconsistencies, we areconcerned with the overlap among theapproaches. Isn't reading typically donebefore one engages in an academic dis-cussion? Are there many, if any, instanceswhere one is ready to teach about some-thing before leaming about it? Isn't priorknowledge of the content necessary tounderstand what is observed during ademonstration? Aren't most audiovisualpresentations preceded by an introduction,or an orienting lecture, and then followedby a discussion. In sum. the instructionaldesign process would be far more compli-cated and thoughtful than simply relyingon the pyramid.

Thus, it is our thesis that isolating thevarious methods is nearly impossible inthe process of good teaching: a balancedcombination, based on content, teacherbackground knowledge, resources, and stu-dent characteristics, is likely to be mosteffective. Further, we wish to go beyondquestioning the validity of the learningperiod to examine the research on retention

for the various methods identified. In thenext section we wiil present some of theresearch results for these methods. How-ever, we wiil not attempt to arrange theseresults in any type of a hierarchy, identi-fying one as better than another, as is donein the pyramid. To do this would be high-ly speculative based on the researchmethodology issues raised earlier.

Teaching Methods and RetentionThe research presented here is not

intended to be a comprehensive literaturereview for each method, as that would gobeyond the scope and purpose of this dis-cussion. Instead we wish to determine if theidentified methods do result in retentionof what was leamed after some period oftime. In some instances we will also reporteffects on leaming to provide a more well-rounded discussion. There were two mainchallenges in examining this literature.First, many empirical studies focus onleaming but do not examine retention, soavailable data are limited in some cases.Second, it is difficult to determine per-centage of retention from a given treatmentbecause percentage of retention is not typ-ically reported: and, for us to calculate it,the researchers must have reported pretestand retention scores, as well as the totalnumber of items on the retention test. Thiswas also not typical. Further, as will beseen, the length of treatment and the delaysbetween treatment and retention vary fromstudy to study which makes attributingretention percentages to one method rela-tive to others impracticable. Therefore, wewill not attempt to ascribe percentages togiven methods.

Examining the Learning .../ 71

Lecture or Direct InstructionDirect instruction is a highly structured

teaching plan often associated withHunter's Mastery Teaching model (1985).It empbasizes teacher direction and stu-dent teacher interaction. Here, the teacherprovides explicit experiences to assist stu-dent attainment of lesson objectives (Eby,Henell, & Jordan, 2006). Direct instructionis the most researched teaching strategyand the one strategy, more than others, tbathas improved student achievement (Kim& Axelrod. 2005: Stein, Carnine. & Dixon.1998). Clearly, direct instruction relies onsome form of teacher lecture. According tothe assertions of the learning pyramid, lec-ture/direct instruction would be least likelystrategy to produce retention.

However, direct instruction is a usefulteaching strategy for students throughoutthe grades, students with exceptionalities(Eby et al., 2006; Rosenshine. 1976;Adams & Engelmann, 1996). and. chil-dren from low socio-economicbackgrounds who typically come to schoolwith less background knowledge than thosefrom more affluent backgrounds (Kim &Axelrod, 2(X)5). Further, direct instructionhas been shown to have a significant affecton retention (Randier & Bogner, 2002).

ReadingIt is not new to assert that a primary

mission of schools is to increase studentreading comprehension. Eurther, a sub-stantial number of studies have beendevoted to analyzing the impact variouslearning to read strategies might have onincreasing the chances students willremember what they bave read. Eor exam-ple, in a study by Wbite (2005) tifteen

grade two regular education teachersagreed to follow a structured set of lessonplans throughout a school year. Theseteachers explicitly aimed at increasing stu-dent capacity to use analogies to help gainmeaning of unknown words througb under-standing of known words. Results showedsignificant positive relationship betweenthe number of lessons the students partic-ipated in and students' gains onstandardized tests in reading comprehen-sion. Results supported the use of specificcomprehension strategy interventions as ameans of improving student understand-ing of text.

Additionally. Van Keer and Verhaegje(2005) also investigated specific strategyinterventions in their study of second andfifth graders. However, in their study theyincluded two types of strategy practice:teacher directed whole class practice andpeer assisted practice. Results for bothgrade level groups indicated significantachievement gains for both groups. In turn,their findings supported a persistent lineof reading comprehension research(McGill-Eranzen & Allington, 1999; Press-ley & Allington, 1999; Snow. Burns, &Griffin, 1999) that indicated that compre-hension strategy practice is related toimproved reading comprehension. Themission and research initiatives have takenon a new sense of urgency with the myri-ad of high-stakes tests relying on astudent's ability to read and interpret infor-mation in each of the content areas. In fact,test taking is increasingly being thoughtof as a new reading genre.

Audio-Visual

72 / Education Vol. 128No.1

Audio-Visual materials represent themost ill-defined teaching methodology. Insome ways, they might not be considereda method of instruction but tools tbat canbe incorporated when using the methodsdiscussed here. Their many forms caninclude videos, sound, pictures, and graphs.They can be used individually or present-ed in various combinations. They can alsobe presented using a variety of media. Athorough examination of each of these isbeyond the scope and purpose of this arti-cle. Further, preliminary literature searchesidentified few, if any, empirical researchthat addressed retention. However, we willexamine some research that seems ger-mane to this discussion.

Increasing attention has been given tothe usefulness of a visual experience toenhance learning. A correspondingresponse to that interest has been the imple-mentation of graphic organizers in lessons.A graphic organizer is a visual represen-tation of facts and concepts and therelationship between and among them (Ebyetal., 2006).

Kim et al. (2004) provide a synthesis ofresearch on the effect of graphic organiz-ers on the reading comprehension ofstudents classified as learning disabledacross grade levels. These researchersexamined twenty one studies and deter-mined all but two yielded overall beneficialoutcomes of graphic organizer use and stu-dent achievement. They reportedsignificant effect sizes with mean posttestscores ranging form 48 to 83% correct onposttests and follow-up. Although they didnot find significant results related to trans-fer, the researchers concluded that the useof a visual enhanced student understand-

ing of what was read.In research on other forms of visual

instruction. Van Hell, Bosman and Bartel-ings (2003) studied the use of visualmaterials by children with spelling prob-lems. They found that "visual dictation"using cue cards improved their spellingwhen it was measured one month afterinstruction. In a study that used actualvideo, Hodges. Chua. and Franks (2003),found that students who saw videotapedfeedback of their performances while learn-ing motor movements had better retentionfollowing a four to seven day delay thanthose who did not receive feedback. Also,when video feedback was provided dur-ing computer assisted instruction, itresulted in superior retention when com-pared to textual feedback (Lalley, 1998).

As technology continues to evolve,teachers will have many options for includ-ing audio-visual materials in instruction.These will include the use of computertechnology to allow students to engage insimulations, a form of discovery learning,which has shown some initial promise(Shaffer, Squire, Halverson, & Gee, 2005).A similar but more immersive and inter-active environment is virtual reality whichis predicted to come to the classroom with-in the next ten years. The integration ofsuch technologies is sure to present chal-lenges while also offering great promisefor both teachers and students.

DemonstrationDemonstration typically involves an

expert (i.e.. teacher) performing a learn-ing task while students observe. The intentis to model the correct behavior in anattempt to minimize ambiguity in instruc-

Examining the Learning .../ 73

tion, and therefore, limit the potential forstudent errors and misconceptions. It isthought to be particularly useful when poorunderstanding of the task to be performedcan lead to students experiencing physicalharm (e.g.. in industrial technology cours-es) or where costly learning materials (e.g..dissection specimens) may be wasted.Research on this method is relativelysparse, which may be directly related tothe safety and cost issues raised, whereresearch on teaching without demonstra-tion would be neither practical norinformative. However, there has been someresearch on using demonstration in moretraditional content areas. One such studywas conducted by Polhemus. Dambe,Moorad, and Dambe (1985) that examinedstudents learning about the concept oflength and how it is not dependent uponcontext (e.g.. spatial orientation). Studentswho were provided with demonstrationsusing manipulatives showed significantpretest (2.6) to retention test (12.2) gainsafter a seven week delay following instruc-tion. However, these were essentiallyidentical to the gains made by studentswho took turns observing each other doingsimilar demonstrations and those who hadtheir own set of manipulatives to learn witbfollowing teacher demonstrations. Allgroups had greater change from pretest toretention than the control group thatreceived no instruction (2.6 on the pretestand 2.1 on the retention test). Similarly, ina study by Pigford (1974) in which pre-service teachers were given instructionabout the metric system, equivalent gainswere made from pretest to retention bydoing hands-on learning in a laboratoryand those receiving lecture/demonstration.

This raises yet another issue regardingoverlap and integration of the methodsidentified by the pyramid: in the classroom,similar to the Pigford study, lecture anddemonstration are likely interwoven. Thiswould make determining the isolated effec-tiveness of the methods impractical and,more importantly, would risk makingexperimental treatments artificial. Thus,teachers would be well advised to providestudents with demonstrations when war-ranted by safety concerns, effective use ofmaterials, and/or need for clarity.

Cooperative Learning and DiscussionGroups

Research support for cooperative learn-ing and achievement has been outlined inJohnson, Johnson and Stane's (20(K)) metaanalysis. Their data indicate higher levelsof student achievement with cooperativelearning as a teaching strategy as comparedto competitive and individualisticapproaches. The cooperative learningmethods included studies comparing John-son and Johnson's Learning Together,Slavin's (1997) Students-Teams-Achieve-tnent-Division (STAD) and the TeamAssisted Individualizcition method to indi-vidualistic methods. According to the datafrom the meta analysis, each provided high-er achievement than competitive andindividualistic methods.

Widespread agreement with the John-son, et al. meta analysis can be found in theliterature (Eby et al., 2006; Mayer, 2002;Omstein & Lasley, 2004). In a study ofmiddle-class students. Yager, Johnson,Johnson, & Snider (1986) compared coop-erative learning with group processing (5minutes of reflection and discussion at the

74 / Education Vol. 128 No.1

end of each session), cooperative learningwithout planned group processing, andindividualistic learning where studentsused learning materials and only soughtassistance from the teacher (essentially adiscovery method). Prior to instruction,students in the sample were compared andidentified as high, middle or low ability.Retention scores, regardless of ability level,showed that cooperative learning withgroup processing was more effective thancooperative learning without group pro-cessing and individualistic instruction, andthat cooperative learning without groupprocessing was more effective than indi-vidualistic learning.

Discussion Groups are intended to stim-ulate student thinking and articulation ofideas related to a topic (Jacobsen, Eggen,& Kauchak, 2005). The teacher's role isto set the conditions in the classroom toincrease the chances students will partic-ipate in the discussion. The teacher canserve as leader, equal participant, or as amonitor who doesn't directly participatebut acts to keep the group on task.

Research on the impact of discussion isoften associated with descriptive studiesof strategies teachers believe have the mostlikelihood of creating an atmosphere toproduce student involvement in a discus-sion (Wilen, 2004). Three studies werefound dealing with retention and discus-sion. In two of these (Drane, Smith, Light,Pinto, & Swarat, 2005; Morgan, Whorton,& Gunsalus, 2000), subjects were collegestudents and a combination of lecture anddiscussion was compared to students whoexperienced lecture only or cooperativelearning only. The results indicated a slightshort-term retention benefit for students in

favor of lecture discussion but no long-term benefit. In the third study, Riesenmy,Mitchell. Hudgins & Ebel (1991) fourth-and fifth graders were taught critical think-ing and problem-solving skills using thediscussion method. They retained theseskills after four and eight weeks while acontrol group that received no instructiondid not.

Practice by DoingOne of tbe most highly advocated types

of practice by doing is Discovery Learn-ing. It is thought to be effective because itencourages students to work like a pro-fessional in the filed (e.g., a scientist).Students work on their own or in groupsto discover principles and relationships ina given content area to develop a person-al understanding of concepts andrelationships that are more meaningful andbetter understood than if they were sim-ply told about them. It can be open-ended,such as taking students on a field trip to anaquarium to learn about marine life, or,can be structured (i.e., guided) in a waythat is intended to help students meet spe-cific objectives. As early as 1960, JeromeBruner advocated the use of discoverylearning in math and the physical sciences,as well as in social sciences.

Ivins (1985) found that eighth-gradescience students retained more informa-tion when concepts were introduced in alaboratory context followed bylecture/reading, than when introducedthrough lecture/reading and followed upby laboratory exercises. It should be noted,based on Ivins' description of the proce-dure, tbat this is not so much a matter ofcomparing methods as it a comparison of

sequencing. Further. Ivins describes thelaboratory activities as "guided discovery".

Research has. at times, shown that dis-covery learning, when compared to directinstruction, results in superior learning.Hillocks" (1984) meta-analysis of teach-ing composition found two mainapproaches. He compared "presentational",lecture and teacher-led discussion to "envi-ronmental", which was guided discoverywith structured activities. It was found thatenvironmental resulted in superior learn-ing. However, retention was not examined.

In a study that did examine retention.Bay et al. (1992) found that discovery wasmore effective than direct instruction withlearning disabled and non-learning dis-abled science students. Students workedin triads in a one-to-two ratio, respective-ly. Using the same learning materials toteach about displacement of water, a dis-covery group worked with materials tocomplete well-structured and specificgoals, while the direct instruction groupobserved the teacher demonstrate conceptsand then completed worksheets to prac-tice the concepts. Retention was tested twoweeks following treatment. Regarding theeffectivenessof the discovery method, theyconclude "whereas direct instruction issuccessful, particularly for basic skillinstruction in reading and mathematics,such teaching may be less beneficial forscience" (p. 567).

Thus, it may be the case that discoverylearning is not appropriate for all circum-stances. Recall that in the Pigford (1974)study cited earlier, the same retentionresults were achieved through hands-onlaboratory learning as were achievedthrough lecture/demonstration. Also. Ran-

Examining the Learning .,. / 75

dler and Bogner (2002) found significantincreases from pretest to retention after asix-week delay for both direct instructionand discovery but found no differences inretention between the two groups.

Given that discovery can be an effectivemethod but not always more effective thanothers, Bruner (1960) reminds us that dis-covery would be too time-consuming forall teaching and that one determinant ofits effectiveness is the teacher's familiari-ty with the content being taught. Thus,teachers must consider their own contentknowledge, and their students' prior knowl-edge and abilities, to choose whendiscovery exercises will be effective.

Teaching OthersThe most common form of students

teaching others is Peer Tutoring. In termsof the effect of teaching on retention, peertutoring research (e.g., Fasko, 1994) oftenprovides little insight because it oftenfocuses on its impact on students beingtutored (tutee) and not the impact on thetutor. However, in good teaching practice,peer tutoring should only be done when itis in the best interest of the lutor becauseit provides opportunities to overleam thematerial and engage in higher levels ofthinking, and/or develop certain socialskills. If the method only helps the tutee,an alternate method should be employed.While research has demonstrated that peertutoring improves achievement, it ofteninvolves reciprocal peer tutoring (Heller& Fantuzzo. 1993: Rekrut. 1994) wherestudents take turns being the tutor, so deter-mining the effects of teaching versus beingtaught is not possible, but there is someresearch that has examined its affect on

76 / Education Vol. 128 No.1

the tutor. Using a story telling strategy asthe content to be learned. Reknit (1994)found that students who were tutors didshow superior retention following onemonth delay when compared to studentswho did not tutor. However, it is paramountto note that both groups of students weregiven training on the strategy, so tutoringwas not the initial learning experience forthe tutors. Instead, it was an opportunity tooverlearn the strategy by teaching it.

SummaryThe research reviewed here demon-

strates that use of each of the methodsidentified by the pyramid resulted in reten-tion, with none being consistently superiorto the others and all being effective in cer-tain contexts. A paramount concern, givenconventional wisdom and the researchcited, is the effectiveness and importanceof reading and direct instruction, which inmany ways are undermined by their posi-tions on the pyramid. Reading is not onlyan effective teaching/learning method, itis also the main foundation for becominga "life-long learner". Its fundamentalimportance was noted early on by Dalewho in 1946 stated he "would give muchmore attention to effective reading in allcurricula" ( p. 6). Further, direct instruc-tion has been shown to be effective,especially for children from low socio-eco-nomic backgrounds.

It is widely accepted that teachersshould facilitate active learning in theirclassrooms. According to Stinson and Mil-ter (1996). "there has been a paradigm shiftexpressed as moving from being the 'Sageon the Stage" to serving as the 'Guide onthe Side"" and teachers will need to have

the following skills: active listening, coach-ing, mentoring, and facilitation. Wesuggest, based on the literature reviewed,that a capable teacher needs to be both"sage" and "guide", and all things inbetween. Thus, the skills identified can beascribed to no particular teaching method,instead tbey are skills that will make teach-ers effective when using most any of themethods discussed here.

If we were to draw any conclusionbased on the pyramid, it would be that themethods be thought of as on a continuumas opposed to in a hierarchy. Therefore,the less prior knowledge students have themore likely it is that effective methodswould be found toward the direct instruc-tion end of the continuum, and as students'knowledge increased, they would be morecapable of learning with methods involv-ing discussion and teaching. However,because learning is an ongoing process,this will not preclude that further learningwill take place with more direct methods.Thus, even the most experienced learners,such as successful heart surgeons, couldlearn from a more experienced learner, asurgeon with a new technique, and the bestinitial methods would likely belecture/demonstration which would leadto practice by doing, and possibly teach-ing others. Not surprisingly, this retumsus to the assertions of Dale (1946) andDewey (1916) that for successful learningexperiences, students need to experience avariety of instructional methods and thatdirect instruction needs to be accompaniedby methods that further student under-standing and recognize why what they arelearning is useful.

Examining the Learning ...ill

ReferencesAdam.s, G. L.. & Engelmann. S. (1996). Research

on Direct Instruction: 25 Years beyond DIS-TAR. Seattle: Educational AchievementSystems.

Bay. M.. & et al. (1992). Science Instruction forthe Mildly Handicapped: Direct Instructionversus Discovery Teaching. Journal ofResearch in Science Teaching, 29(6), 555-57().

Boulmetis, J. (2003). Learning Pyramid. Instruc-tor, 113(3), 9.

Brueckner. J. K., & MacPherson, B. R. (2004).Benefits from peer teaching in the dental grossanatomy laboratory. European Journal of Den-tal Education. 8{2\ 72-77.

Brunei, J. S. (1960). Process of education. NewYork: Vintage Books.

Dale, E. (1946). Audio-visual methods in teach-ing. New York: The Dryden Pres.s.

Dale, E. (1954). Audio-visual methods in teaching(2 ed.). New York: The Dryden Press.

Dale. E. (1969). Audio-visual methods in teaching(3 ed.). New York: The Dryden Press.

Danielson, C. (2002). Enhancing Studetit Achieve-ment: A Framework for School Improvement.Alexndria. Virginia: Association for Supervi-sion and Curriculum Development.

Darmer, M. R.. Ankersen. L.. Nielsen, B. G.. Land-berger. G., Lippert. E.. & Egerod. f. (2004).The effect of a VIPS implementation pro-gramme on nurses knowledge and attiludestowards documentation. Scandinavian Journalof Caring Sciences, IHO). 325-332.

DeKanter, N. (2005). Gaming Redefines Interac-tivity for Learning. TechTrends: LinkingResearch & Practice to Improve Learning,49(3), 26-31.

Dewey. J. (1916). Democracy and education: anintroduction to the philosophy of education.New York: The Macmillan Company.

Drane. D.. Smith, H.. Light. G., Pinto, L., &Swarat. S. (2005), The Gateway Science Work-shop Program: Enhancing StudentPerformance and Retention in the SciencesThrough Peer-Facilitated Discussion. Journalof Science Education & Technologv, 74(3),337-352.

Drewes, F., & MilU^an, K. L. D. (2003). How ToStudy Science (4 ed.). Boston: McGraw-Hill.

Eby, J., HerrelL A., & Jordan, M. (2006). Teach-ing in K-12 Schools: A Reflective ActionApproach. Columbus, Ohio: PearsoB-Merrill-Prentice Hall.

Fasko, S. N. (1994). The Effects of a Peer Tutor-ing Program on Math Fact Recall andGeneralization. Paper presented at the Paperpresented at the Annual Meeting of the Amer-ican Psychological Association (1994). LosAngeles. CA.

Heller, L. K.. & Fantuzzo. J. W, (1993). Recipro-cal peer tutoring and parent partnership: Doesparent involvement make a difference? SchoolPsychology Review. 22(3), 517-534.

Hershman, D. M., & McDonald, E. S. (2003). Sur-vival Kit for New Secondary Teachers.Garland, TX: Inspiring Teachers Publishing,Inc.

Hillocks, G. (1984). What works in teaching com-position: A mcta-analysis of experimentaltreatment studies. American Journal of Edu-cation, 93. 133-170.

Hodges. N. J., Chua. R.. & Franks, I. M. (2003).The role of video in facilitating perception andaction of a novel coordination movement.Journal of Motor Behavior, 35{3), 247-260.

Hunter. M. (1985). Mastery Teaching. ThousandOaks. California: Corwin Press,

Ivins, J. E, (1985). A Comparison ofthe Effectsof Two Instructional Sequences Involving Sci-ence Laboratory Activities. The AmericanBiology Teacher. ERIC Document Service(ED259953).

Jacobsen, P A.. Eggen, P, & Kauchak, D, (2005).Methods for Teaching: Promoting StudentLearning in K-12 Classrooms (7 ed.). Boston:Prentice Hall.

78 / Education Vol. 128 No.1

Johnson, D. W.. Johnson, R. T,. & Stanne. M. B.(2000), Cooperative Learning Methods: AMeta-Analysis. December 12. 2005. fromhttp://www.co-operation.org/pages/cl-meth-ods.html

Kim, T., & Axelrod, S. (2005), Direct instruction:an educators' guide and a plea for action.Behavior Analyst Today. 6(2), 111-120.

Lalley, J. P. (1998). A comparison of video andtext as forms of feedback during computer-assisted instruction. Journal of EducationalComputing Research. 75(4). 323-338.

McGill-Franzen. A,, & Allington. R. L. (1999).Putting books in the classroom seems neces-sary but not sufficient. Journal of EducationalRe.search. 93(2). 61-14.

Morgan, R, L.. Whorton, J. E.. & Gunsalus, C.(2000), A Comparison of Short Term and LongTerm Retention: Lecture Combined with Dis-cussion Versus Cooperative Learning. Journalof instructional Psychology. 27(1). 53-57,

Pigford. V. D. (1974). A Comparison of an Indi-vidual Laboratory Method with a GroupTeacher -Demonstration Method in TeachingMeasurement and Estimation in Metric Unitsto Preservice Elementary Teachers: The Flori-da State University. ERIC Document Service(ED 108928).

Polhemus, S., Dambe. M., Moorad. F.. & Dambe,M. (1985). Group training of conservationskills with children from a non-technological.International Journal of Psychology. 2011), I.

Pressley, M.. & Allington, R. (1999), What shouldreading instructional research be the researchof? Issues in Education, 5(1). 1-35.

Randier. C, & Bogner, F. (2002). Comparingmethods of instruction using bird species iden-tification skills as indicators. Journal ofBiological Education, 36(4). 181-188.

Rekrut, M, D, (1994), Teaching to leam: Strategyutilization through peer iiitoring. High SchoolJournal, 77(4), 304-314.

Riesenmy. M. R,, Mitchell, S.. Hudgins, B. B,. &Ebel. D. (1991). Retention and transfer of chil-dren's self-directed and critical thinking skills.Journal of Educational Research, 85(1), 14-25.

Rosenshine, B. (1976), Recent Research on Teach-ing Behaviors and Student Achievement.Journal of Teacher Education, 27(1), 61-64.

Shaffer, D. W., Squire. K. R,, Halverson, R., &Gee, J. P. (2005), Video Games and The Futureof Leaming. (Cover story). Phi Delta Kappan,87(2). 105-111.

Snow. C. E.. Burns. M. S,. & Griffin, P. (1999).Preventing reading difficulties in young chil-dren. In Reading re.search: Anthology: Thewhy? of reading instruction, (pp. 148-155).Novato. CA: Arena Press,

Sousa. D. A. (200\).How the brain learns: a class-room teacher's guide (2" ed.). Thousand Oaks,Calif: Corwin Press.

Slavin, R. E. (1997). Educational Psychology:Theory and Practice (5' ed.), Boston: Allynand Bacon:

Stein. M.. Caniine. D.. & Dixon. R. (1998). Directinstruction: integrating curriculum design andeffective teaching practice. Intervention inSchool & Clinic, .•?.?(4). 227-234.

Stinson. J.. & Milter, R. (1996). Problem-basedLearning In Business Education: CurriculumDesign and Implementation Issues, In L. a. G.Wilkerson & Wim (Eds.). Bringing Problem-Based Learning to Higher Education: Theoryand Practice. San Francisco, CA.: Jossey-BassPublishers.

Thalheimer, W. (2005). Bogus Research Uncov-ered Work -Learning Research Somerville,Massachusetts, USA (617) 718-0067.http://www.work-leaming.com/chigraph.htm

Van Hell. J.G., Bosman.A. M,T., & Bartelings,M. C. G. (2003). Visual dictation improves thespelling performance of three groups of dutchstudents with spelling disabilities. LearningDisabiliry Quarterly. 26(4). 239-255.

Van Keer. H.. & Verhaeghe, J, P. (2005). Effectsof Explicit Reading Strategies Instruction andPeer Tutoring on Second and Fifth Graders'Reading Comprehension and Self-EfficacyPerceptions. Journal of Experimental Educa-tion. 7 3(4), 291-319.

Examining the Learning .../ 79

White. T. G. (2005). Effects of systematic andstrategic analogy-based phonics on grade 2 stu-dents' word reading and readingcomprehension. Reading Research Quarierlv,400-), 234.

Wilen. W. W. (2004). Refuting MisconceptionsAbout Classroom Discussions. Social Studies,95(1), 33-39.

Wood. E. J. (2004). Problem-Based Learning:Exploiting Knowledge of how People Leam toPromote Effective Learning. Bioscience Edu-cation e-joiinmt(BEE-j) (Vol. 3, pp. 3-13): TheHigher Education Academy.

Yager. S.. Johnson. R. T, Johnson, D. W.. &Snider, B. (1986). The Impact of Group Pro-cessing on Achievement in CooperativeLearning Groups. Journal of Social Psvcholo-gy. 1260). 389-397.