JOURNAL OF APPLIED BEHAVIOR ANALYSIS

"GOOD-BYE, TEACHER.. .'.

FRED S. KELLER

ARIZONA STATE UNIVERSITY2

When I was a boy, and school "let out" forthe summer, we used to celebrate our freedomfrom educational control by chanting:

Good-bye scholars, good-bye school;Good-bye teacher, darned old fool!

We really didn't think of our teacher as de-ficient in judgment, or as a clown or jester.We were simply escaping from restraint, din-ner pail in one hand and shoes in the other,with all the delights of summer before us. Atthat moment, we might even have been well-disposed toward our teacher and might havefelt a touch of compassion as we completed therhyme.

"Teacher" was usually a woman, not alwaysyoung and not always pretty. She was fre-quently demanding and sometimes sharp oftongue, ever ready to pounce when we got outof line. But, occasionally, if one did especiallywell in home-work or in recitation, he coulddetect a flicker of approval or affection thatmade the hour in class worthwhile. At suchtimes, we loved our teacher and felt thatschool was fun.

It was not fun enough, however, to keep methere when I grew older. Then I turned to an-other kind of education, in which the rein-forcements were sometimes just as scarce as inthe schoolroom. I became a Western Unionmessenger boy and, between deliveries of tele-grams, I learned Morse code by memorizingdots and dashes from a sheet of paper and lis-tening to a relay on the wall. As I look backon those days, I conclude that I am the only

'President's Invited Address, Division 2, Amer. Psy-chol. Ass., Washington, D.C., Sept., 1967.

2Currently on leave of absence at the Institute forBehavioral Research, 2426 Linden Lane, Silver Spring,Maryland. Reprints may be obtained from the author,3229 Park View Road, Chevy Chase, Maryland.

79

living reinforcement theorist who ever learnedMorse code in the absence of reinforcement.

It was a long, frustrating job. It taught methat drop-out learning could be just as difficultas in-school learning and it led me to wonderabout easier possible ways of mastering a skill.Years later, after returning to school and fin-ishing my formal education, I came back tothis classical learning problem, with the aimof making International Morse code less pain-ful for beginners than American Morse hadbeen for me (Keller, 1943).During World War II, with the aid of a

number of students and colleagues, I tried toapply the principle of immediate reinforce-ment to the early training of Signal Corps per-sonnel in the reception of Morse-code signals.At the same time, I had a chance to observe,at close hand and for many months, the opera-tion of a military training center. I learnedsomething from both experiences, but Ishould have learned more. I should have seenmany things that I didn't see at all, or sawvery dimly.

I could have noted, for example, that in-struction in such a center was highly individ-ualized, in spite of large classes, sometimespermitting students to advance at their ownspeed throughout a course of study. I couldhave seen the clear specification of terminalskills for each course, together with the care-fully graded steps leading to this end. I couldhave seen the demand for perfection at everylevel of training and for every student; theemployment of classroom instructors whowere little more than the successful graduatesof earlier classes; the minimizing of the lectureas a teaching device and the maximizing ofstudent participation. I could have seen, es-pecially, an interesting division of labor in theeducational process, wherein the non-commis-sioned, classroom teacher was restricted toduties of guiding, clarifying, demonstrating,

1968. 1, 79-89 NUMBER I (SPRING, 1968)

80 FRED S. KELLER

testing, grading, and the like, while the com-missioned teacher, the training officer, dealtwith matters of course logistics, the interpre-tation of training manuals, the construction oflesson plans and guides, the evaluation of stu-dent progress, the selection of non-commis-sioned cadre, and the writing of reports for hissuperiors.

I did see these things, of course, in a sense,but they were embedded deeply within a spe-cial context, one of "training" rather than"education". I did not then appreciate that aset of reinforcement contingencies which wereuseful in building simple skills like those ofthe radio operator might also be useful in de-veloping the verbal repertories, the conceptualbehaviors, and the laboratory techniques ofuniversity education. It was not until a longtime later, by a very different route, that Icame to such a realization.That story began in 1962, with the attempt

on the part of two Brazilian and two NorthAmerican psychologists, to establish a Depart-ment of Psychology at the University of Bra-silia. The question of teaching method arosefrom the very practical problem of getting afirst course ready by a certain date for a cer-tain number of students in the new university.We had almost complete freedom of action;we were dissatisfied with the conventional ap-proaches; and we knew something about pro-grammed instruction. We were also of thesame theoretical persuasion. It was quite nat-ural, I suppose, that we should look for freshapplications of reinforcement thinking to theteaching process (Keller, 1966).The method that resulted from this collabo-

rative effort was first used in a short-term lab-oratory courseS at Columbia University in thewinter of 1963, and the basic procedure of thispilot study was employed at Brasilia duringthe following year, by Professors RodolfoAzzi and Carolina Martuscelli Bori, with 50students in a one-term introductory course.Professor Azzi's report on this, at the 1965meetings of the American Psychological Asso-ciation and in personal correspondence, indi-cated a highly satisfactory outcome. The newprocedure was received enthusiastically by thestudents and by the university administration.

sWith the aid of (Dr.) Lanny Fields and the membersof a senior seminar at Columbia College, during thefall term of 1963-64.

Mastery of the course material was judgedexcellent for all who completed the course.Objections were minor, centering around therelative absence of opportunity for discussionbetween students and staff.

Unfortunately, the Brasilia venture came toan abrupt end during the second semester ofits operation, due to a general upheavalwithin the university that involved the resig-nation or dismissal of more than 200 teachers.Members of the original psychology staff havesince taken positions elsewhere, and have re-portedly begun to use the new method again,but I am unable at this time to report in detailon their efforts.

Concurrently with the early Brazilian devel-opment, Professor J. G. Sherman and I, in thespring of 1965, began a series of more or lessindependent applications of the same generalmethod at Arizona State University. With var-ious minor changes, this work has now beentried through five semesters with an increasingnumber of students per term (Keller, in press[a], in press [b], 1967; Sherman, 1967). Theresults have been more gratifying with eachsuccessive class, and there has been as yet nothought of a return to more conventional pro-cedures. In addition, we have had the satisfac-tion of seeing our system used by a few othercolleagues, in other courses and at other in-stitutions.4

In describing this method to you, I will startwith a quotation (Keller, 1967). It is from ahand-out given to all the students enrolled inthe first-semester course in General Psychology(one of two introductions offered at ArizonaState University) during the past year, and itdescribes the teaching method to which theywill be exposed unless they elect to withdrawfrom the course.

"This is a course through which youmay move, from start to finish, at yourown pace. You will not be held back byother students or forced to go ahead untilyou are ready. At best, you may meet allthe course requirements in less than onesemester; at worst, you may not complete

4For example, by J. L. Michael with high-school jun-iors on a National Science Foundation project at Grin-nell College (Iowa), in 1965; and by J. Farmer and B.Cole at Queens College (New York) in a course similarto the one described here.

80

"GOOD-BYE, TEACHER . . .8

the job within that time. How fast you gois up to you."The work of this course will be divided

into 30 units of content, which corre-spond roughly to a series of home-workassignments and laboratory exercises.These units will come in a definite nu-merical order, and you must show yourmastery of each unit (by passing a "readi-ness" test or carrying out an experiment)before moving on to the next."A good share of your reading for this

course may be done in the classroom, atthose times when no lectures, demonstra-tions, or other activities are taking place.Your classroom, that is, will sometimes bea study hall."The lectures and demonstrations in

this course will have a different relationto the rest of your work than is usuallythe rule. They will be provided onlywhen you have demonstrated your readi-ness to appreciate them; no examinationwill be based upon them; and you neednot attend them if you do not wish. Whena certain percentage of the class hasreached a certain point in the course, alecture or demonstration will be availableat a stated time, but it will not be com-pulsory."The teaching staff of your course will

include proctors, assistants, and an in-structor. A proctor is an undergraduatewho has been chosen for his mastery ofthe course content and orientation, forhis maturity of judgment, for his under-standing of the special problems thatconfront you as a beginner, and for hiswillingness to assist. He will provide youwith all your study materials except yourtextbooks. He will pass upon your readi-ness tests as satisfactory or unsatisfactory.His judgment will ordinarily be law, butif he is ever in serious doubt, he can ap-peal to the classroom assistant, or eventhe instructor, for a ruling. Failure to passa test on the first try, the second, thethird, or even later, will not be heldagainst you. It is better that you get toomuch testing than not enough, if yourfinal success in the course is to be assured.

"Your work in the laboratory will becarried out under the direct supervisionof a graduate laboratory assistant, whose

detailed duties cannot be listed here..There will also be a graduate classroomassistant, upon whom your proctor willdepend for various course materials (as-signments, study questions, special read-ings, and so on), and who will keep up todate all progress records for course mem-bers. The classroom assistant will conferwith the instructor daily, aid the proctorson occasion, and act in a variety of waysto further the smooth operation of thecourse machinery."The instructor will have as his princi-

pal responsibilities: (a) the selection of allstudy material used in the course; (b) theorganization and the mode of presentingthis material; (c) the construction of testsand examinations; and (d) the final eval-uation of each student's progress. It willbe his duty, also, to provide lectures, dem-onstrations, and discussion opportunitiesfor all students who have earned the priv-ilege; to act as a clearing-house for re-quests and complaints; and to arbitrate inany case of disagreement between stu-dents and proctors or assistants....

"All students in the course are expectedto take a final examination, in which theentire term's work will be represented.With certain exceptions, this examinationwill come at the same time for all stu-dents, at the end of the term.... The ex-amination will consist of questions which,in large part, you have already answeredon your readiness tests. Twenty-five per-cent of your course grade will be based onthis examination; the remaining 75% willbe based on the number of units of read-ing and laboratory work that you havesuccessfully completed during the term."

(In my own sections of the course, these per-centages were altered, during the last term, toa 30% weighting of the final examination, a20% weighting of the 10 laboratory exercises,and a 50% weighting of the reading units.)A picture of the way this method operates

can best be obtained, perhaps, by sampling theactivities of a hypothetical average student ashe moves through the course. John Pilgrim isa freshman, drawn from the upper 75% of hishigh-school class. He has enrolled in PY 112for unknown reasons and has been assigned toa section of about 100 students, men and

81

82 FRED S. KELLER

women, most of whom are also in their begin-ning year. The class is scheduled to meet onTuesdays and Thursdays, from 9:15 to 10:30a.m., with a laboratory session to be arranged.Together with the description from which I

quoted a moment ago, John receives a fewmimeographed instructions and some wordsof advice from his professor. He is told thathe should cover two units of laboratory workor reading per week in order to be sure of tak-ing an A-grade into his final examination; thathe should withdraw from the course if hedoesn't pass at least one readiness test withinthe first two weeks; and that a grade of In-complete will not be given except in specialcases. He is also advised that, in addition tothe regular classroom hours on Tuesday andThursday, readiness tests may be taken on Sat-urday forenoons and Wednesday afternoons ofeach week-periods in which he can catch upwith, or move ahead of, the rest of the class.He then receives his first assignment: an in-

troductory chapter from a standard textbookand two "sets" from a programmed version ofsimilar material. With this assignment, he re-ceives a mimeographed list of "study ques-tions", about 30 in number. He is told to seekout the answers to these questions in his read-ing, so as to prepare himself for the questionshe will be asked in his readiness tests. He isfree to study wherever he pleases, but he isstrongly encouraged to use the study hall forat least part of the time. Conditions for workare optimal there, with other students doingthe same thing and with an assistant or proc-tor on hand to clarify a confusing passage ora difficult concept.This is on Tuesday. On Thursday, John

comes to class again, having gone through thesets of programmed material and having de-cided to finish his study in the classroom,where he cannot but feel that the instructorreally expects him. An assistant is in charge,about half the class is there, and some late reg-istrants are reading the course description.John tries to study his regular text, but finds itdifficult to concentrate and ends by decidingto work in his room. The assistant pays no at-tention when he leaves.On the following Tuesday, he appears in

study hall again, ready for testing, but anx-ious, since a whole week of the course haspassed. He reports to the assistant, who sendshim across the hall, without his books and

notes, to the testing room, where the proctorin charge gives him a blue-book and one ofthe test forms for Unit 1. He takes a seatamong about 20 other students and startswork. The test is composed of 10 fill-in ques-tions and one short-answer essay question. Itdoesn't seem particularly difficult and, inabout 10 min John returns his question sheetand is sent, with his blue-book, to the proctor'sroom for grading.

In the proctor's room, in one of 10 smallcubicles, John finds his special proctor, AnneMerit. Anne is a psychology major who passedthe same course earlier with a grade of A. Shereceives two points of credit for about 4 hr ofproctoring per week, 2 hr of required attend-ance at a weekly proctors' meeting, and occa-sional extra duty in the study hall or testroom. She has nine other students besidesJohn to look after, so she will not as a rule beable to spend much more than 5 or 10 min ofclass time with each.Anne runs through John's answers quickly,

checking two of them as incorrect and placinga question mark after his answer to the essayquestion. Then she asks him why he answeredthese three as he did. His replies show twomisinterpretations of the question and onefailure in written expression. A restatementof the fill-in questions and some probing withrespect to the essay leads Anne to write anO.K. alongside each challenged answer. Shecongratulates John upon his performance andwarns him that later units may be a littleharder to master than the first.

John's success is then recorded on the wall-chart in the proctors' room, he is given hisnext assignment and set of study questions,and sent happily on his way. The blue-bookremains with Anne, to be given later to theassistant or the instructor for inspection, andused again when John is ready for testing onUnit 2. As he leaves the room, John noticesthe announcement of a 20-min lecture by hisinstructor, for all students who have passedUnit 3 by the following Friday, and he re-solves that he will be there.

If John had failed in the defense of one ortwo of his answers, he would have been sentback for a minimal period of 30 min for fur-ther study, with advice as to material mostneeding attention. If he had made more thanfour errors on his test, the answers would nothave been considered individually; he would

82

"GOOD-BYE, TEACHER . . .'

simply have been told that he was not readyfor examination. And, if he had made no er-rors at all, he would probably have been askedto explain one or two of his correct answers,as a way of getting acquainted and to makesure that he was the one who had really donethe work.John did fail his first test on Unit 2, and his

first two tests on Unit 4 (which gave trouble tonearly everyone). He missed the first lecture,too, but qualified for the second. (There wereseven such "shows" during the term, each at-tended by perhaps half of the students entitledto be there.) After getting through his first fiveunits, he failed on one review test before earn-ing the right to move on to Unit 6. On theaverage, for the remainder of the course, herequired nearly two readiness tests per unit.Failing a test, of course, was not an unmixedevil, since it permitted more discussion withthe proctor and often served to sharpen theconcepts involved.

In spite of more than a week's absence fromschool, John was able, by using the Wednes-day and Saturday testing sessions, to completehis course units successfully about a week be-fore the final examination. Because of hiscramming for other courses during this lastweek, he did not review for his psychologyand received only a B on his final examina-tion. His A for the course was not affected bythis, but his pride was hurt.Sometime before the term ended, John was

asked to comment on certain aspects of thecourse, without revealing his identity. (Re-member, John is a mythical figure.) Amongother things, he said that, in comparison withcourses taught more conventionally, this onedemanded a much greater mastery of the workassignments, it required greater memorizationof detail and much greater understanding ofbasic concepts, it generated a greater feelingof achievement, it gave much greater recogni-tion of the student as a person, and it was en-joyed to a much greater extent (Keller, inpress).He mentioned also that his study habits had

improved during the term, that his attitudetowards testing had become more positive,that his worry about final grades had dimin-ished, and that there had been an increase inhis desire to hear lectures (this in spite of thefact that he attended only half of those forwhich he was qualified). When asked specifi-

cally about the use of proctors, he said that thediscussions with his proctors had been veryhelpful, that the proctor's non-academic, per-sonal relation was also important to him, andthat the use of proctors generally in gradingand discussing tests was highly desirable.Anne Merit, when asked to comment on her

own reactions to the system, had many thingsto say, mostly positive. She referred especiallyto the satisfaction of having the respect of herproctees, of seeing them do well, and of ce-menting the material of the course for herself.She noted that the method was one of "mutualreinforcement" for student, proctor, assistant,and instructor. She suggested that it ought tobe used in other courses and at other levels ofinstruction. She wondered why it would notbe possible for a student to enroll in a secondcourse immediately upon completion of thefirst, if it were taught by the same method. Shealso listed several changes that might improvethe efficiency of the course machinery, espe-cially in the area of testing and grading, wheredelay may sometimes occur.

In an earlier account of this teachingmethod (Keller, 1967), I summarized thosefeatures which seem to distinguish it mostclearly from conventional teaching proce-dures. They include the following;

"(1) The go-at-your-own-pace feature,which permits a student to move throughthe course at a speed commensurate withhis ability and other demands upon histime.

"(2) The unit-perfection requirementfor advance, which lets the student goahead to new material only after demon-strating mastery of that which preceded.

"(3) The use of lectures and demon-strations as vehicles of motivation, ratherthan sources of critical information.

"(4) The related stress upon the writ-ten word in teacher-student communica-tion; and, finally:

"(5) The use of proctors, which permitsrepeated testing, immediate scoring, al-most unavoidable tutoring, and a markedenhancement of the personal-social aspectof the educational process."

The similarity of our learning paradigm tothat provided in the field of programmed in-struction is obvious. There is the same stress

83

FRED S. KELLER

upon analysis of the task, the same concernwith terminal performance, the same oppor-tunity for individualized progression, and soon. But the sphere of action here is different.The principal steps of advance are not"frames" in a "set", but are more like the con-ventional home-work assignment or laboratoryexercise. "The 'response' is not simply thecompletion of a prepared statement throughthe insertion of a word or phrase. Rather, itmay be thought of as the resultant of manysuch responses, better described as the under-standing of a principle, a formula, or a con-cept, or the ability to use an experimentaltechnique. Advance within the program de-pends on something more than the appear-ance of a confirming word or the presentationof a new frame; it involves a personal interac-tion between a student and his peer, or hisbetter, in what may be a lively verbal inter-change, of interest and importance to eachparticipant. The use of a programmed text, ateaching machine, or some sort of computeraid within such a course is entirely possibleand may be quite desirable, but it is not to beequated with the course itself." (Keller, 1967.)

Failure to recognize that our teaching unitsare not as simple as the response words in aprogrammed text, or the letter reactions toMorse-code signals, or other comparable atomsof behavior, can lead to confusion concerningour procedure. A well-known critic of educa-tion in America, after reading an account ofour method, sent me a note confessing to "agrave apprehension about the effect of break-ing up the subject matter into little packages.""I should suppose," he said, "it would preventall but the strongest minds from ever possess-ing a synoptic view of a field, and I imaginethat the coaching, and testing, and passing inbits would amount to efficient training ratherthan effectual teaching."Our "little packages" or "bits" are no

smaller than the basic conceptions of a scienceof behavior and cannot be delivered all atonce in one large synoptic parcel. As for theteaching-training distinction, one needs onlyto note that it is always the instructor whodecides what is to be taught, and to what de-gree, thus determining whether he will becalled a trainer or a teacher. The method heuses, the basic reinforcement contingencies heemploys, may be turned to either purpose.Many things occur, some of them rather

strange, when a student is taught by a methodsuch as ours. With respect to everyday studentbehavior, even a casual visit to a class willprovide some novel items. For example, all thestudents seated in the study hall may be seenstudying, undistracted by the presence ormovements of others. In the test room, a stu-dent will rarely be seen chewing on his pencil,looking at a neighbor's blue-book, or staringout the window. In the crowded proctors'room, 10 pairs of students can be found con-currently engaged in academic interaction,with no couple bothered by the conversationof another, no matter how close by. Uponpassing his assistant or instructor, in the cor-ridors or elsewhere, a student will typically beseen to react in a friendly and respectful man-ner-enough to excite a mild alarm.More interesting than this is the fact that a

student may be tested 40 or 50 times in thecourse of one semester, often standing in linefor the privilege, without a complaint. In oneextreme instance, a student required nearlytwo terms to complete the work of one (afterwhich he applied for, and got, permission toserve as a proctor for the following year).Another unusual feature of our testing and

grading is the opportunity given to defend an"incorrect" answer. This defense, as I notedearlier, may sometimes produce changes inthe proctor's evaluation, changes that are reg-ularly checked by the assistant or the instruc-tor. Occasionally, a proctor's O.K. will be re-jected, compelling the student to take anothertest, and sensitizing the proctor to the dangersof leniency; more often, it produces a note ofwarning, a correction, or a query written bythe instructor in the student's blue-book; butalways it provides the instructor with feedbackon the adequacy of the question he has con-structed.

Especially important, in a course taught bysuch a method, is the fact that any differencesin social, economic, cultural, and ethnic back-ground are completely and repeatedly subor-dinated to a friendly intellectual relationshipbetween two human beings throughout a pe-riod of 15 weeks or more. Also, in such acourse, a lonesome, ill-favored underprivil-eged, badly schooled, or otherwise handi-capped boy or girl can be assured at least amodicum of individual attention, approval,encouragement, and a chance to succeed. Theonly prerequisite for such treatment is a well-

84

"GOOD-BYE, TEACHER.

defined amount and quality of academicachievement.Another oddity of the system is the produc-

tion of a grade distribution that is upsidedown. In Fig. 1, are the results from a class of208 students at Arizona State University dur-ing the past semester. Note the diminishingrelative frequency as one moves from A to D.The category of E, indicating failure, is swol-len by the presence of 18 students who failedto take up their option ofW (withdrawal fromthe course). Grades of C and D were due tothe failure of students to complete all theunits of reading or laboratory before goinginto the final examination.

.40Uja J3

.20 * .

.10

Fig. 1

tory course at Queens College (N. Y.) duringthe second semester of the past school year.The same method of teaching was employedas at Arizona State, but the work requirementwas somewhat greater in amount. The distinc-tive feature here is the relative infrequency oflow grades. Only four students received lessthan a B rating. Professor John Farmer, whoprovided me with these data, reports that thetwo students receiving F had dropped out ofthe course, for unknown reasons, after sevenand eight units respectively.

.40

.5.0

Uj .20

.10 ~ AdF D C B A

Fig. 3

Figure 2 shows data from the class 1 yr ear-lier. Essentially the same distribution holds,except for the category of Incomplete, whichwas then too easily obtainable. Discouragingthe use of the Incomplete, together with theprovision of more testing hours, apparentlyhas the effect of regularizing study habits andequalizing the number of tests taken per weekthroughout the term.

.50

.40

C.)

.30

, . 20

10

.00W INC E D

Fig. 2C B A

In Fig. 3 (filled bars), the grade distributionis for a section of 25 students in an introduc-

With this teaching method, students whoare presumably inferior may show up betterupon examination than presumably superiorstudents taught by more conventional proce-dures. Figure 4 shows two distributions ofgrades on a mid-term examination. The emptybars represent the achievement of 161 studentsof an Ivy League College, mainly sophomores,in the first semester of a one-year lecture-and-laboratory course in elementary psychology.The filled bars represent the achievement of66 Arizona State University students, mainlyfreshman, on an unannounced mid-term quizprepared by the Ivy League instructor andfrom which 13% of the questions had to beeliminated on the grounds of differentialcourse coverage.

Relevant to this comparison is that picturedin Fig. 3. The grade distribution obtained byProfessor Farmer (and his associate, BrettCole) is here compared with one obtainedfrom a section of 46 students in the samecourse, taught in the conventional manner bya colleague who is described as "a very goodinstructor". The filled bars show the Farmer-

85

FRED S. KELLER

40

0aU

mg

as

w

10

r D C- C CI' B- a SI AGRADESFig. 4

Cole results; the empty ones are those fromProfessor Brandex.Such comparisons are of some interest and

may relieve the tedium of a lecture, but theyraise many questions of interpretation, andtheir importance should not be over-empha-sized. The kind of change needed in educa-tion today is not one that will be evaluated interms of the percentage of A's in a grade dis-tribution or of differences at the 0.01 (or0.001) level of confidence. It is one that willproduce a reinforcing state of affairs for every-one involved-a state of affairs that has here-tofore been reached so rarely as to be the sub-

ject of eulogy in the world's literature, andwhich, unfortunately, has led to the mystiqueof the "great teacher" rather than a soberanalysis of the critical contingencies in opera-tion.Our method has not yet required a grant-

in-aid to keep it going. On one occasion wetried to get such help, in order to pay formimeograph paper, the services of a clerk, andone or two additional assistants. Our requestwas rejected, quite properly, on the groundsthat our project was "purely operational". Al-most any member of a present-day fund-grant-ing agency can recognize "research" when he

86

"GOOD-BYE, TEACHER ... 7.

sees it. I do think, however, that one shouldbe freed, as I was, from other university de-mands while introducing a system like ours.And he should not be asked to teach more

than two such courses regularly, each serving100 students or less, unless he has highly qual-ified assistants upon whom he can depend.

Neither does the method require equipmentand supplies that are not already available toalmost every teacher in the country. Teachingmachines, tape recorders, and computerscould readily be fitted into the picture. Mov-ing pictures and television could also be usedin one or two ways without detriment to thebasic educational process. But these are luxu-ries, based on only partial recognition of our

problem, and they could divert us from more

important considerations. (Proctors, like com-

puters, may go wrong or break down, but theycan often be repaired and they are easily re-

placed, at very little expense.)The need for individualized instruction is

widely recognized, and the most commonlysuggested way of filling this need is automa-tion. I think that this solution is incomplete,especially when applied to the young; and I'dlike to mention a personal experience thatbears upon the matter.

In the summer of 1966, 1 made numerous

visits to a center for the care and treatment ofautistic children.5 One day, as I stood at thedoor of a classroom, I saw a boy get up fromhis chair at the end of a class period and givea soft pat to the object on the desk in front ofhim. At the same time, he said, with a slightsmile, "Good-bye, Teaching Machine!"This pseudo-social behavior in this funda-

mentally asocial child amused me at the time.It reminded me of Professor Moore's descrip-tion of the three-year-old who became irritatedwhen his "talking typewriter" made a mistake,called the device a "big bambam", requestedits name, and ended by asking, "Who is yourmother?" Today, however, I am not so sure

that this is funny. It does suggest that affectionmay be generated within a child for an electro-mechanical instrument that has been essentialto educational reinforcement. Unfortunately,such a machine, in its present form, is unlikelyto generalize with human beings in the boy'sworld, giving to them a highly desirable rein-

5At the Linwood Children's Center, Ellicott City,Maryland.

forcing property. In fact, the growth of thistype of student-machine relation, if it were theonly one, would be a poor substitute for adirectly social interaction.

In an earlier report upon our method, Imentioned that it had been anticipated, par-tially or in toto, in earlier studies and I de-scribed one of these in some detail. As for cur-rent developments by other workers in ourfield, I have not made any systematic attemptto examine the offerings, even those that dealwith college or university instruction. How-ever, I have been impressed by several of themwhich seem to have points in common withours, which have met with some success, andwhich will probably be increasingly heardfrom in the future.One of these is the Audio-Tutorial Ap-

proach to the teaching of botany, developedby S. N. Postlethwait at Purdue University(Postlethwait and Novak, 1967). Another isthe Socratic-Type Programming of generalpsychology by Harry C. Mahan (1967) and hisassociates at Palomar College, in California;and a third is the Interview Technique re-cently applied by C. B. Ferster and M. C. Per-rott (1968) in teaching principles of behaviorto graduate students in education at the Uni-versity of Maryland.

Professor Postlethwait's method places greatemphasis upon "independent study sessions"in which students carry out each individualwork assignment in the course at their ownpace, by means of the extensive use of tapesand films. Teaching assistants provide for oralquizzing on major concepts and help the stu-dents with difficult assignments. Weekly"small assembly sessions" are used primarilyfor recitation and the discussion of problemsor small research projects; and "general as-sembly sessions" deal mainly with motiva-tional materials. Postlethwait reports highstudent interest and greatly improved per-formance with the use of this technique."Grades have risen from 6% A's under theconventional system to as high as 25% A's insome semesters. Failures have decreased from20%0 in the conventional system to as few as4%."

"Socratic-Type Programming" is describedby Professor Mahan as "a philosophy and tech-nology of instruction which emphasizes stu-dent response rather than presentations by theteacher. Its basic media consist of exercises

87

FRED S. KELLER

made up of questions and short answers cov-ering the content of a standard text, the textitself, tapes for recording the questions in theexercises, a classroom tape recorder for ad-ministering tests, tape duplicating facilities, alistening center in the college library, andstudent owned tape recorders for home usewhenever possible. Classroom time is devotedlargely to the discussion of points covered bythe questions. All examinations are the short-answer type and are presented aurally ontape." Students must pass three periodic testswith a score of 85% or better before they arepermitted to take a comprehensive final exam-ination. The method does not yet permit"multiple exit" from the course, but Mahansays it is "tending very much in that direc-tion." (1967.)The Interview Technique, as described by

Ferster and Perrott, does permit students tocomplete the course at different times, and italso approximates the student-and-proctor fea-ture. Progress through the course is possibleby verbalizing successive units of course con-tent in a lengthy series of short interviews.The interviews are conducted mainly betweenstudents currently enrolled in the course, andany student is free to leave the course whenall of his reading assignments have been ade-quately covered. The interviewer may some-times be a staff member, as at the beginning ofthe course, but generally he is a student whohas already been interviewed by someone elseon the topic in question. The interviews arehighly formalized, with the interviewer play-ing the role of the listener, checker, appraiser,and summarizer. Each interview is an open-book affair, but of such short and sharply-defined duration (10 min, as a rule) that thestudent can do no more than cue himself byreference to the printed page.The goal of this method is nothing less than

fluency with respect to each main feature ofthe course. Lectures, group discussions, anddemonstrations are available at certain times,contingent upon a given stage of advance. In-adequate interviews are rejected, in whole orpart, without prejudice, and with suggestionsfor further study. A product of high quality isguaranteed through staff participation at crit-ical points. A modification of this procedure,which is to include written tests and the em-ployment of advanced-student proctors, isplanned by Professor Ferster for the introduc-

tory course in psychology at Georgetown Uni-versity during the coming semester.

In systems like these, and in the one I havecentered on, the work of a teacher is at vari-ance with that which has predominated in ourtime. His public appearances as classroom en-tertainer, expositor, critic, and debater nolonger seem important. His principal job, asFrank Finger (1962) once defined it, is truly"the facilitation of learning in others." Hebecomes an educational engineer, a contin-gency manager, with the responsibility of serv-ing the great majority, rather than the smallminority, of young men and women who cometo him for schooling in the area of his compe-tence. The teacher of tomorrow will not, Ithink, continue to be satisfied with a 10% ef-ficiency (at best) which makes him an object ofcontempt by some, commiseration by others,indifference by many, and love by a few. Nolonger will he need to hold his position by theexercise of functions that neither transmit cul-ture, dignify his status, nor encourage respectfor learning in others. No longer will he needto live, like Ichabod Crane, in a world thatincreasingly begrudges providing him roomand lodging for a doubtful service to itsyoung. A new kind of teacher is in the mak-ing. To the old kind, I, for one, will be gladto say, "Good-bye!"

I started this paper on a personal note andI would like to end it on one. Twenty-oddyears ago, when white rats were first used aslaboratory subjects in the introductory course,a student would sometimes complain abouthis animal's behavior. The beast couldn'tlearn, he was asleep, he wasn't hungry, he wassick, and so forth. With a little time and ahandful of pellets, we could usually show thatthis was wrong. All that one needed to do wasfollow the rules. "The rat," we used to say,"is always right."My days of teaching are over. After what I

have said about efficiency, I cannot lay claimto any great success, but my schedule of re-wards was enough to maintain my behavior,and I learned one very important thing: thestudent is always right. He is not asleep, notunmotivated, not sick, and he can learn agreat deal if we provide the right contingen-cies of reinforcement. But if we don't providethem, and provide them soon, he too maybe inspired to say, "Good-bye!" to formaleducation.

88

"GOOD-BYE, TEACHER ... ." 89

REFERENCES

Ferster, C. B. and Perrott, M. C. Behavior principles.New York: Appleton-Century-Crofts, 1968. Pp. 542.

Finger, F. W. Psychologists in colleges and universi-ties. In W. B. Webb (Ed.), The profession of psy-chology. New York: Holt, Reinhart and Winston,1962. Pp. 50-73.

Keller, F. S. Studies in international morse code: 1.a new method of teaching code reception. Journalof Applied Psychology, 1943, 27, 407-415.

Keller, F. S. A personal course in psychology. In R.Ulrich, T. Stachnik, and J. Mabry (Eds.), The con-trol of behavior. Glenview, Ill.: Scott, Foresman,1966. Pp. 91-93.

Keller, F. S. Neglected rewards in the educationalprocess. Proc. 23rd Amer. Conf. Acad. Deans, LosAngeles, Jan., 1967. Pp. 9-22.

Keller, F. S. New reinforcement contingencies in the

classroom. In Programmiertes lernen, Wissenschaft-liche Buchgesellschaft, Darmstadt, in press.

Keller, F. S. Engineering personalized instruction inthe classroom. Rev. Interamer de Psicol., 1967, 1,189-197.

Keller, F. S. and Schoenfeld, W. N. The psychologycurriculum at Columbia College. American Psychol-ogist, 1949, 4, 165-172.

Mahan, H. C. The use of Socratic type programmedinstruction in college courses in psychology. Paperread at West. Psychol. Ass., San Francisco, May,1967.

Postlethwait, S. N. and Novak, J. D. The use of 8-mmloop films in individualized instruction. AnnalsN. Y. Acad. Sci., Vol. 142, Art. 2, 464-470.

Sherman, J. G. Application of reinforcement princi-ples to a college course. Paper read at Amer. Educ.Res. Ass., New York, Feb., 1967.

Received 27 October 1967.