A free operant analysis of programed insturction ... Assessing Instructional Programs began publishing

Embed Size (px)

Text of A free operant analysis of programed insturction ... Assessing Instructional Programs began...












  • Copyright, 1969 by

    Dennis L. Edinger

  • This dissertation is dcdicv.t.ed to

    Ogden R. Lindsley


    This dissertation is the first installment on

    my debt to Professor Ogden R. Lindsley .

    Carl Koenig and ,:Tolin Nicho l of the B1.:".!havior

    Bank (P.O. Box 3351 , Kansas City , Kansas ) deserve enthusiastic

    applause from the Florida group for their brilliant efforts

    in educational science . The data in this c1.isrrnrtnt.i.on could

    no t have been presented without their invaluable assistance .

    My gratitude , thanks , and profound regard to

    my chairman , nc. W. D. Walking, and to rrty minor c ircctor ,

    Dr.. H. s . Pennypacker, for the superb direction nr.n

    1.eadershipo.f 1ny infa nt ile gropings for a precise 3Cie ~;.ce of

    human behavior.. If I 2.m indeed a scientist, I am of their


  • TABLE OF CON'l'EN'.rs

    ACKl'lOhTJ .• r:;ocr~r .. 1ENTS • •••••••••••••••••••••••

  • LIST OP 'r.i\BLF:S

    TAilLE Page I A Simple Analysis of Variance for Differences

    netwr.:0.n Pre-P l acement Test Scores , last Cor.\pl cted Programmed Reader Book.let Number , and Post - lj lacec ,:,ent. Test Scores. .. ....... .. . . . .... . ... .... . . 2 3

    II A Li:idquist Type I AnalysiG cf Variance f or Differences Between Correct and Incorrect Progr a.·nmed Rec:\dcr Response Rates. • • • • • • • • • • • • • • • • • 2 5

    III The Direction and Magnitude of Di fference s Bc:?twocn Correct c:.nd Inco:i:.·rect Programmed neader Response Rates . . ... • • • • • • • • • • • • • • • • • • • • • • • • 26

    IV A Lin

  • LIST OF TABLES (continued)

    TABLE XI A 1-lilcoxon Signed Rank ·rest for Dif forences

    Between Before Phase Incorrect Progrru!l.mcd Reader Response R.J.tes an


    FIGURES Page I 'l\ schematic illustration of the within-subject

    design, with replications, used in this study... 15

    II correct an

  • i\ppendix E TA!JULAR CODE

    SRP-BC - Prograr.1cd Reader , Before Phase Correc t SRP-D C - Programed Reuder , During Phase Correc t SRP-A C - Progri:lmed Reader , After Phase Correct SRP-B I - Programed neadcr , Before Phase Incorre ct SRP-DI - Programed Reader 1 Dur i ng Phase In corre ct SRP-A I - Programed Reader , After Phase Incorrec t

    DT-I3C - Di

  • Chapter I


    Education is exclus i vcly co ncerned wi. th behav ior

    change for the purpose of developing and maintaining comple x

    repertoires of culturally-valued human behavior . In orde r

    to eva luate the effectiveness of its procedures, educat ion

    r,n.1st have a reliable and sensitive method for c~escr ibing

    an d rnGasur ing tile behavior changes it produces. Cm~r~n tly,

    such evaluat jon is performed almost total ly by psychometry

    psyche-educational tests and rating scales. Despite its

    undoubt:cd importance historically, psycho!r..ct·cy has nc~-, been

    shown to have se:::ious shortcomin

  • of cu:i:ricul.::. and teaching methods on behavior . i-.·ith the

    partial e:{.ception of achievement tests , psychometri c

    procedures never directly me

  • The first of these is the checklist.. Al thou

  • instruction. Bloom• s Taxonomy of F.dnca tionctl Ohjccti ves - - .. ------· . ---- (1956) has been repeatedly invoked as the evaluative

    / standard form a "cognitivcn point of view. 'l'he chief

    spokesman for this type of evaluation i.s Louise Tyler

    (1966). Not satisfied with theory only, she has reported

    some data on its use by teachers already frn1iliar with the

    taxonomy. Newman (1965) has u::;ed it to evaluate programed

    instruction in the Social Studies.

    The second method of evaluation that can be

    distinguished might more properly be called the comparison

    procedure. It generally takes the fom of comparing

    programed instruction with traditional teaching methods ,

    or programed texts with standard texts.

    Schramm (1964) indicates that much cf the

    evaluative research done in progrruned instruction is of


    this nature. An examination of Educational Abstract s

    confirms this statement for the subsequent years. ·rhe

    difficulty, however, is that comparison, as a researc h

    method , reveals little or nothing about the pro9ramed

    instructional material~.::!!:.• Nevertheless, the Joint

    Commit tee recom!!l.ends comparison as a method for the external

    validation of programed material , and it may be an acceptab.le

    procedure for this task.

    Another difficulty with comparison studies lies

    in the nature of the research dcoign employe d. It is almost

    irnpossible, ir:. this type of l·cscarch, to contro l for

    individual differences in children and in teachers, an d

    its even mor~ di ff icul t to control for. dif:f.et·enccs ln content .

  • 5

    'l'he majority of research of the comparative typ~ , as Schramm

    notes, is so poorly done that little faith can be placed

    in the validity of the results . There arc fortunate exceptions ,

    particularly with regard to exceptional child populations.

    Blackman and Copobianco (1965 ) , for example , report on

    the use of a specif ic programed material with retarded

    children . Likewise , Rainey and Kelly {19G.l) report.

    the use of a time-·telling program with educable retarda tcs ,

    and Streng (1964 ) reports evaluating a program with deaf

    opulations . This research is child-oriented and involves

    the determination of the utility of a specific program for

    developing a defined behavior in a given exceptional chil d

    population . The utility of this type of research for the?

    practici11g clussroom teacher should not be under-estimated.

    Closely rel~ted to the programed instruction with

    "other" comparison, is the programed instruction with

    achievement test comparison. Normally , the test used is

    one of the standard ~chievemcnt tests such as the Wide Range

    Achievement 'l'est or the Metropolitan Achievement Test . In

    this case , the research questions are directed to differenc e

    scores on the specific test before and after the administration

    of the programed instruction material . The discontinuous

    (before and after ) nature of this procedure is a seriou s

    shortcoming because it does not. allow ,'l point to poin t

    analysis of the relationship between the program and the

    child ' s behavior . This method, like the progrMicd instruction

    with "other" comparison, is recorr ,mended by the Joint

    Committee for the ex.tcrn~1l valldation of the material .

  • 6

    The careful reader of research is quick to not e

    that the dependent variable in these studie.s is not prograr:1.ed

    instruction performance, but achievement test performance .

    Programed instruction pcrfor.mance is then inferred f r om

    the test performanc e.

    In his text on evaluating programed .instru c tio n

    Jacobs , ct a l . , (1966 ) , mentions the Denver Stud y as the c lass ic

    mode l for eval uation . This study , reported int~~ by Jacob s,

    util i zed both the progr

  • 7

    test and attitude scale scores, no direct statement may

    be made regarding the behavior change on the program itself.

    Also mentioned in the evaluation literature ,

    but clearly not research , is a caution to the progra.i.l

    user to check the credentials of the program author c.nd

    the publisher. The Joint Committee advises all publishers

    to include with each program sold, comple tc develop!

  • 8

    experimental analysis of human behavior was presented . Sound

    experimental data were not forthcoming until Lindsley's classic

    study with chronic psychotics (1956 ). Ski11ncr {1958), reporting

    on his research with programed instruction technology , an

    extension of free operant techniques with animals , excited

    much interest in the educational community. Bijou (1957, 1958)

    developed observation techniques for young Ghildren patterned

    closely after those used by Lindsley.

    The marriage between the educator and free opera11t methods

    wa~; not long in coming. Birnbrauer, Bijou, Wolf , and Kidder

    (1965) demonstrated the application of free operant tcchnlque3

    in a classroom s.i. tu.=ttion using pro~rramed instruction as a

    cun:icular core. Zim.iuermr'in and Zimmerman ( 1962) also applied

    free operant techniques in a classroom with much less structure

    than Birnbrauer's classroom .


  • 9

    o. R. Lindsley (19