Guerin 1997

8/7/2019 Guerin 1997

1/23

69

JOURNAL OF APPLIED BEHAVIOR ANALYSIS 1997, 30, 6991 NUMBER 1 (SPRING 1997)

TEACHING CHILDREN TO SPELL:DECREASING CONSONANT CLUSTER ERRORS BYELIMINATING SELECTIVE STIMULUS CONTROL

BEVERLY BIRNIE-SELWYN AND BERNARD GUERIN

UNIVERSITY OF WAIKATO

Six elementary-aged children were taught to spell words containing initial consonantclusters (CCs). They were trained to select printed words in response to the correspondingspoken words using computerized matching-to-sample procedures. After each trainingsession, they were tested for spelling with a constructed-response transfer test. Based onprevious selective stimulus control research, we hypothesized that only the first letter ofan initial CC might control spelling when CC spelling errors are made. Thus, a critical-difference matching-to-sample training condition that required the children to respondto both letters of the CC to be correct was compared to a multiple-difference trainingcondition that required the children to respond to only one letter of the pair. Resultsshowed that children made fewer errors during the multiple-difference training conditionthan during the critical-difference training condition. On the constructed-response trans-fer tests, however, more overall errors and CC errors were made in the multiple-difference

condition than in the critical-difference condition, and the words trained in the multiple-difference condition required more training sessions to reach criterion. All children im-proved their spelling of novel CC words by the completion of training. If normal class-room or home reading was to be supplemented by computer tasks of the kind used here,some spelling problems could be circumvented without costly intervention by a teacheror a special trainer.

DESCRIPTORS: spelling, children, computers, education, stimulus overselectivity

Most research on spelling, and how toteach spelling, has been done by educationalresearchers based in cognitive psychology(Foorman, Francis, Novy, & Liberman,

1991; Griffith, 1991; Lewkowicz, 1980;Treiman & Zukowski, 1988, 1990). Withinbehavior analysis, Lee and colleagues (Lee &Pegler, 1982; Lee & Sanderson, 1987) havedeveloped a conceptual basis for spelling thatemerges from reading and suggested that re-searchers look more closely at the stimuluscontrol involved in specific types of spellingerrors. Dube, McDonald, McIlvane, andMackay (1991), Stevens, Blackhurst, andSlaton (1991), and Moxley and Warash

We thank Vicki Lee and Mary Foster for helpfulcomments about this manuscript, and Bill Dube, Rob-ert Stromer, and Harry Mackay for useful discussionsduring a visit to the Shriver Center.

Correspondence should be addressed to BernardGuerin, Department of Psychology, University of Wai-kato, Private Bag, Hamilton, New Zealand (E-mail:[email protected]).

(19901991) have developed useful com-puter technologies for teaching spelling,whereas other applied behavior analysts havedeveloped interventions that reinforce cor-

rect spelling (Gettinger, 1985; Greenwood etal., 1987; Neef, Iwata, & Page, 1980; Ollen-dick, Matson, Esveldt-Dawson, & Shapiro,1980) and have looked at variables that con-trol practice effects (Cuvo, Ashley, Marso,Zhang, & Fry, 1995; Gettinger, 1993).

Dube et al. (1991) taught 2 young menwith intellectual disabilities to spell using acomputer. Previously trained, arbitrarilymatched words and pictures were used as vi-sual stimuli. Participants were shown thestimuli and were required to select letters

presented at the bottom of the screen witha touch screen to construct the matchingword. The selection pool contained 10 let-ters, some of which, if selected in the rightorder, would spell the word correspondingto the picture. Dube et al. found that the

8/7/2019 Guerin 1997

2/23

8/7/2019 Guerin 1997

3/23

71STIMULUS CONTROL OF SPELLING ERRORS

dure for decreasing selective stimulus con-trol. Applying this procedure to spelling thecompound stimulus SH in the word SHOW,the use of PLAY as a comparison in match-ing-to-sample would be part of a multiple-difference condition, the use of SLOP wouldbe part of a minimal-difference condition,whereas the use of SLOW would be part ofa critical-difference condition. The presentstudy applied only multiple- and critical-dif-ference methods to CCs.

Interestingly, selective stimulus controlhas been found to be widespread in childrenwith autism and intellectual handicaps butis not usually found in nondisabled children(Allen & Fuqua, 1985; Lovaas, Koegel, &

Schreibman, 1979; Lovaas, Schreibman,Koegel, & Rehm, 1971). This might be afunction of the types of stimuli examined,however. Nondisabled children might nothave selectivity problems with the shapesand patterns commonly used to research se-lective stimulus control with people with in-tellectual disabilities or autism, but theymight exhibit selective stimulus control withmore complex stimuli. CC errors might beone example of selective stimulus control innondisabled children (Lee & Sanderson,

1987; Treiman, 1993).The aim of the current research was to

reduce the number of initial CC errors madeby nondisabled children by training worddiscriminations using multiple- and critical-difference methods and testing transfer toconstructing the words from individual let-ters. The critical-difference condition trainedthe children to identify both letters of theCCs, whereas in the multiple-difference con-dition, the children needed to identify onlythe first letter of the word. The matching-

to-sample task used to train the words wassimilar to Dube et al.s (1991), except thatthe sample stimuli were spoken words ratherthan pictures. The question was whether thecritical-difference and multiple-differencematching-to-sample discrimination training

procedures would produce differences inperformance on a constructed-responsetransfer test.

METHODPARTICIPANTS

Six parents volunteered their children asparticipants in this study after volunteerswere recruited for a computer experiment.The children were of normal intelligenceand were not selected because of specialspelling problems. They were between theages of 4 years 5 months (Fiona) and 7 years(Colin).

MATERIALS AND SETTING

The experiment was carried out in a stan-dard office, and the experimenter was pres-ent throughout each training session. Toyswere available during breaks between ses-sions. The children could ask for a breakfrom the experiment at any time during thesession. An Apple Macintosh IIsi computerwith a keyboard and mouse was used topresent the visual and auditory stimuli. Thevisual experimental stimuli were displayedon a color screen (20 cm by 15 cm), and

the auditory experimental stimuli were re-corded in Hypercard and attached as a re-source file to the main program. Five differ-ent screen events, intended to give positivereinforcement and consisting of differingcombinations of flashing colors, moving pic-tures, and sound, were constructed.

The child responded to the experimentalstimuli by positioning an arrow on thescreen with the mouse and then clicking themouse on the key he or she had chosen. Inthe training sessions, the child selected one

of three words presented on the screen. Inthe constructed-response tests, he or shechose letters to spell the stimulus word. Thetop portion of Figure 1 shows the trainingscreen for preliminary training with three-letter words. In the experimental training

8/7/2019 Guerin 1997

4/23

8/7/2019 Guerin 1997

5/23


tial CCs were common English words thathave been found to result in spelling errors(Treiman, 1993), but this was heavily con-strained by the availability of three critical-

difference comparison words. The compari-son words in the selection pool for the twotraining conditions were based on differentcriteria. In the critical-difference trainingsessions, the comparison words were mini-mally different from the correct word thatwas being trained. Only the second conso-nant in the first comparison word differedfrom the correct word, and the vowel andfinal consonant were different in the secondcomparison word. For example, when thecorrect word was SNOW, the first compar-

ison word was SLOW and the second com-parison word was SNAP. In order to adhereto these selection criteria, four of the selec-tion pool words were nonsense words, al-though they were phonetically correct in En-glish: SKED, SMEW, GWAT, and TWAY.In the multiple-difference training sessions,the first and second comparison words in theselection pool differed completely from thecorrect word. The comparison words allcontained four letters and could include ini-tial, final, or no CCs. For example, when thecorrect word was SNOW, the two compar-ison words were NICE and REST.

PROCEDURE

Each child was given an initial oral spell-ing test consisting of the 24 training words.No reinforcement was given for correct orincorrect responses. Words that were spelledcorrectly in this initial oral test were not in-cluded in the experimental conditions forthose children. Out of the 24 words, the

children answered only 5, 6, 3, 2, 9, and 6correctly (Anna to Fiona, respectively). Theymade 10, 13, 19, 26, 9, and 12 CC errors(Anna to Fiona, respectively). The Appendixshows the incorrect spellings of those wordschosen for training (P).

Preliminary Screen Training

The object of the preliminary training wasto teach the children to respond appropri-ately to the auditory stimuli by clicking the

mouse on screen words or letters. This train-ing used five three-letter words with no CCs,but was otherwise identical to the matching-to-sample discrimination training and con-structed-response test conditions describedbelow. The children were taught in threephases: first, to select the spoken letter froma selection of ten letters; second, to select aspoken three-letter word from a pool ofthree words; and last, to select the correctletters to spell those same three-letter words.On the first trial of each phase, only the

correct words or letters were available for se-lection by making the other two gray andinoperative. Therefore, the children got allfive correct. On the second trial, all wordsor letters were black and could be selected,but the correct one shook in an obviousmanner as a prompt. After five correct wordsor letters in a row, the third trial began withall words or letters available and with noprompts. When a child made five consecu-tive incorrect responses, the second trial pro-

cedure reintroduced with its obviousprompts. All children completed this entirepreliminary screen training in less than 50min, at which point they had been trainedto select the question mark for the presen-tation of the spoken stimulus words andthen to select words or letters by using themouse.

Experimental Conditions

At the beginning of the first session, con-structed-response tests were given (0 in Ap-

pendix) of the three critical-difference con-dition words and the three multiple-differ-ence condition words about to be trainedwords that the children had spelled incorrectlyon the initial oral test (P in Appendix).

Each session consisted of 30 training pre-

8/7/2019 Guerin 1997

6/23

74 BEVERLY BIRNIE-SELWYN and BERNARD GUERIN

sentations, five for each of the six words be-ing trained at any one time. At any time,there were usually three words being trainedusing the multiple-difference condition andthree words using the critical-difference con-dition. Each session of 30 training presen-tations was immediately followed by sixnonreinforced constructed-response tests,one for each of the six words being trainedin that session. Occasional probes on wordsto be trained in future sessions were also giv-en after the constructed-response tests (seeAppendix, lower case words). Usually, twosessions were completed each day, with abreak of an hour or more between sessions.The selection pool for training consisted of

whole words, whereas in the constructed-re-sponse test condition, the selection poolconsisted of 10 letters from which to con-struct (spell) the sample word.

Multiple-difference and critical-differencetraining presentations. The operating condi-tions, consequences for feedback, blackout,and the layout of the computer screens werethe same for both the multiple-differenceand the critical-difference conditions. Theonly differences between the two trainingmethods were the incorrect comparison

words that appeared for selection duringtraining.

The children were presented with thescreen shown at the top of Figure 1. Theyfirst clicked on the question mark (an ob-serving response) to produce a speech bub-ble from the face on the screen and theauditory stimulus. The auditory stimulusconsisted of an oral presentation of the wordbeing trained and a sentence containing thatword (e.g., Snow. [pause] Snow is white.).The child could repeat the observing re-

sponses before responding. The child thenselected the word from the selection poolthat matched the auditory stimulus. Onlyone word from the selection pool could bechosen, because all the words in the selectionpool became inoperable (gray) after the first

response. The word that was selected ap-peared in the box just above the selectionpool. The child then clicked on the ENDkey to indicate that he or she had finishedthat presentation. Three seconds of blackoutwere presented after an incorrect response; acorrect response received one of the five dif-ferent feedback screens. The next trial fol-lowed immediately.

Six words were presented to each childfive times each session. Three words weretrained with each of the two different train-ing conditions, and no word was trained inboth conditions for any 1 child. The train-ing condition used for a word was variedacross the 6 children. Three children (Bella,

David, and Fiona) received training onwords selected from the first 12 words inTable 1 by the multiple-difference methodand on words from the second 12 words bythe critical-difference method. The wordsfor the other 3 children (Anna, Colin, andElena) were the reverse: critical-differencetraining on the first 12 words and multiple-difference training on the second 12 words.The actual words used depended upon eachchilds initial oral tests. The training condi-tions were also alternated within each ses-

sion. For example, the first word might re-ceive multiple-difference training, the sec-ond critical-difference training, the thirdmultiple-difference training, and so on.

Constructed-response transfer test. Aftercompleting the 30 training presentations ofeach session, the six words being trained atthat time were each tested once. No feed-back or blackout conditions were presentedduring this test. The computer screen layoutfor the constructed-response test conditionis shown at the bottom of Figure 1. The

auditory stimuli were the same as in the crit-ical- and multiple-difference training condi-tions. The child was required to click on thequestion mark to produce the auditory stim-ulus and then to construct a word by select-ing up to six letters from the selection pool

8/7/2019 Guerin 1997

7/23


area. The letters appeared in the construct-ed-response box just above the selection poolafter they were selected. The END key wasagain in operation, and no more than six

letters could appear in the constructed-re-sponse box. The letters in the selection poolbecame inoperable (gray) when they hadbeen selected; in this way a child could notselect the same letter twice.

Data from the constructed-response testcondition were analyzed at the end of eachsession. When a child had met criterion byconstructing a word corrrectly for three con-secutive sessions, that word was dropped andanother word to be trained using the sametype of condition was introduced. All new

words were pretested again before training.If a word was spelled correctly, it did notundergo training, and pretesting continueduntil a new word was not correctly spelled.That word was then introduced in training.

Dependent Measures

The results that were of most interest were(a) the percentage of correctly selected wordsduring training, (b) the number of sessionsrequired to learn to construct the words cor-

rectly, (c) the percentage of words spelledcorrectly during the constructed-responsetests, and (d) the percentage of CC errorsduring the constructed-response tests. CCerrors were categorized as follows. A CC er-ror was the incorrect spelling or omission ofthe first or second consonant in each word.A first omission CC error was the completeomission of the first consonant in any givenword (e.g., NAP instead of SNAP). A sec-ond omission CC error was the completeomission of the second consonant in any

given word (e.g., WEN instead of WHEN).A reverse CC error was the reversal of thesecond consonant and third vowel in anygiven word (e.g., SONW instead of SNOW).Finally, an insert error occurred when a letterwas inserted between the two consonants of

the cluster (excluding reverse CC errors)(e.g., SENOW instead of SNOW).

RESULTS

MATCHING-TO-SAMPLEDISCRIMINATION TRAINING

Overall Percentage Correct in Training

The top panel of Figure 2 shows the per-centage of correct responses made by eachchild during the training presentations, withthe results separated for the critical-differ-ence and multiple-difference conditions.Children selected whole words during thistraining, and did not construct words fromindividual letters. The data illustrate that agreater percentage of correct responses wererecorded in the multiple-difference condi-tion than in the critical-difference conditionfor all 6 children.

CONSTRUCTED-RESPONSETRANSFER TESTS

Overall Percentage Correct

The middle panel of Figure 2 shows thepercentage of correct responses made in con-structed-response tests, separated by condi-

tion. The data illustrate that the percentageof correct responses was higher for all chil-dren in the critical-difference condition thanin the multiple-difference condition.

Number of Training Sessions to MeetConstructed Word Criterion

The Appendix depicts the complete in-dividual data for the constructed-responsetests over all sessions separately for the twoconditions. Words shown in capital letterswere being trained when a constructed-re-

sponse test was given. From this, the numberof training sessions required for each wordcan be found. For example, with the critical-difference training, Anna required threetraining sessions for the word SNOW beforemeeting criterion and 11 training sessions

8/7/2019 Guerin 1997

8/23


Figure 2. The top two panels show the percentage of words that were correctly selected during trainingand constructed-response tests for both the critical-difference (open bars) and multiple-difference (black bars)training conditions. The bottom panel shows the percentage of words spelled during constructed-responsetesting that contained CC errors for both the critical-difference and multiple-difference training conditions.

8/7/2019 Guerin 1997

9/23


for the word WHEN. Over the six wordstrained with critical-difference training, sheaveraged 5.3 training sessions to meet crite-rion.

These results show that all children tookfewer sessions to learn critical-differencewords than multiple-difference words. The6 children (Anna to Fiona, respectively) tookan average of 5.3, 5.4, 4.8, 4.8, 3.7, and 4.7sessions per word for the critical-differencetraining words and 6.0, 7.8, 11.7, 5.7, 5.3,and 6.4 sessions per word for the multiple-difference training words.

Fourteen of the 24 words were trained byboth methods with at least 1 child per meth-od, thus allowing a between-child compari-

son. For example, 5 children were trained onthe word SKIM, 2 in the critical-differencecondition and 3 in the multiple-differencecondition. The Appendix indicates that for9 of these 14 words, children took fewer ses-sions to complete training in the critical-dif-ference condition than in the multiple-dif-ference condition. Those words were SKIM(5, 4 vs. 5, 9, 9), STAY (3, 3, 7 vs. 11, 4,8), SCAB (3, 4 vs. 6, 7), WRIT (5, 3, 3 vs.9, 8, 5), BLEW (4, 10, 3 vs. 10, 4, 9),CHAP (4, 3 vs. 7, 6), FROG (6 vs. 10, 7),

GNAT (3, 3, 3 vs. 5, 3, 5), and CROP (3,6 vs. 12). Children took fewer sessions tocomplete training in the multiple-differencecondition for only 2 of the 14 words:WHEN (11 vs. 6, 6) and TWIN (9, 9, 7vs. 5, 5, 5). The remaining three words(SNOW, FLAY, and TRIG) did not show aclear difference in the number of sessions re-quired to complete training.

Another result included in the Appendixis that for both training conditions, childrenimproved their spelling of novel words by

the end of the experiment. For all children,after six or seven words had been trained,the novel words presented as untrainedprobes were constructed correctly. For ex-ample, Anna was trained on six words withcritical-difference training, and the next four

probes were correctly spelled without train-ing (SCAB, SHOW, PLAY, SKIM).

Consonant Cluster Errors

Of the 180 errors made in this experi-ment, 61% were CC errors. Of these 109CC errors, 30 (27%) were made in the crit-ical-difference condition and 79 (73%) weremade in the multiple-difference condition.

Consonant cluster errors for each word. Thebottom panel of Figure 2 shows the per-centage of words that contained CC errorsfor both training conditions. The resultsshow that all 6 children made more CC er-rors in the multiple-difference condition.

As can be seen in the Appendix, the chil-

dren frequently made no CC errors for aword in the critical-difference condition,whereas they made numerous CC errors inthe multiple-difference condition. Annasdata show that four of the six words trainedin the critical-difference condition recordedno CC errors compared to only one of thefive words in the multiple-difference condi-tion. Bellas data show that in the critical-difference condition, four of the seven wordsrecorded no CC errors, but all six words inthe multiple-difference condition recorded

CC errors. The results for Colin show thatfive of the nine words trained in the critical-difference condition recorded no CC errorscompared to one out of nine in the multiple-difference condition. Davids results showthat seven of the nine words in the critical-difference condition recorded no CC errorscompared to only three of the nine words inthe multiple-difference condition. For Elena,in the critical-difference condition, five ofthe seven words recorded no CC errors, butin the multiple-difference condition two of

the six words recorded no errors. Fionas re-sults show that six of the eight words in thecritical-difference condition recorded no CCerrors, whereas two of the seven words in themultiple-difference condition recorded noerrors.

8/7/2019 Guerin 1997

10/23


Table 2

The Number of Each Type of CC Errors for BothTraining Conditions

Type oferrors

Critical

differ-ence

Multiple

differ-ence

Total CCerrors

Percent-age of

CCerrors

First omissionCC errors 1 7 8 7.3

Second omissionCC errors 10 39 49 44.0

Reverse CC errorsInsert CC errorsOther CC errors

1180

1977

3015

7

27.513.8

6.4

Total 30 79 109 100

Types of consonant cluster errors. Table 2lists the total number of each type of CCerror made by all the children in the con-structed-response tests. The results show thatfor all CC error types (with the exception ofinsert CC errors), more were made for wordsthat had been trained in the multiple-differ-ence condition than in the critical-differencecondition. A total of 109 errors were made,30 in the critical-difference condition and79 in the multiple-difference condition.

Only one first omission error was made

in the critical-difference condition, but sevenwere made in the multiple-difference con-dition. As expected from past research oninitial CC words, the most frequent CC er-ror was the second omission error (a total of49), which accounted for nearly half of allCC errors recorded. Most of these secondomission CC errors (80%) occurred in themultiple-difference condition. Thirty reverseCC errors were recorded in total, the secondmost frequent type of error. Of these, 63%were made in the multiple-difference con-

dition. The combined total of reverse andinsert CC errors, both letters of the CC be-ing present but not in the correct sequence,accounted for nearly half of the CC errors(45). The two training conditions did notdiffer in insert errors.

DISCUSSION

This study was based on the suggestionthat CC errors occur because of selectivestimulus control by one of the consonants.

The present research compared the results oftwo matching-to-sample discriminationtraining conditionscritical-difference andmultiple-difference trainingin reducingthese spelling errors made by children on aconstructed-response transfer test. Overall,the constructed-response test results indicat-ed that the critical-difference training con-dition was more effective than the multiple-difference training condition in reducingCC errors and overall errors. Thus finer dis-crimination training at the level of single let-

ters led to improved spelling performance,as had been suggested by Dube et al. (1991),Lee and Sanderson (1987), and Stromer andMackay (1992b).

The number of training sessions requiredbefore the constructed-response performancereached criterion decreased as training pro-gressed under both conditions. This suggeststhat the critical-difference training conditionfocused the childrens attention on criticalaspects of the CC combination and that this

generalized to performance under the mul-tiple-difference condition. It is very likelythat performance during the multiple-differ-ence condition would have been worse inthe absence of any critical-difference train-ing. However, because training and timewere confounded, it is possible that the over-all improvement in performance was due toextraexperimental experiences. The fact thatsome words were constructed correctly inthe absence of specific discrimination train-ing strengthens the argument for the effect

of such extraexperimental factors.The results from the training conditions

in which whole words were matched to sam-ple showed that a greater number of errorswere made in the critical-difference condi-tion than in the multiple-difference condi-

8/7/2019 Guerin 1997

11/23


tion. These results therefore show that train-ing with the multiple-difference conditionwas easier than with the critical-differencecondition, as would be expected. The datafrom the training condition alone do not es-tablish the controlling factors that governedthe childrens behavior during the matching-to-sample training because selection wasmade from among whole words. However,the fact that more overall errors and CC er-rors occurred in the constructed-responsetests with words trained in the multiple-dif-ference condition indicates that the childrenwere not identifying all four letters of thecompound stimulus during training, whichresulted in more mistakes being made in the

constructed-response tests. These data areconsistent with the results of Allen and Fu-qua (1985), who used other types of stimuli.

With regard to the types and frequency ofinitial CC errors, Treiman (1991) found thatsecond omission errors accounted for 42%of the CC errors. Other researchers havefound that second omission errors accountedfor 33% of the CC errors (Miller & Limber,1985). Although the number of secondomission errors that occurred in the critical-difference condition of the present study was

very low compared to the number that oc-curred in the multiple-difference condition,second omission errors accounted for 44%of the total number of errors, close to thatfound by Treiman (1991). The second omis-sion errors in both conditions therefore sub-stantiate past findings that the most com-mon error made by children trying to spellwords that contain initial CCs is the omis-sion of the second consonant in the cluster(Bruck & Treiman, 1991; Marcel, 1980;Miller & Limber, 1985; Treiman, 1991).

The higher rate of second omission errors inthe multiple-difference condition comparedto the critical-difference condition providesfurther support that the spelling of CCs inwords from the multiple-difference condi-tion came under the control of only one of

the components of the clusterthe initialconsonant.

Another interesting finding was that re-verse and insert CC errors combined ac-counted for nearly half of the CC errors.Both of these errors involve the sequencingrather than the occurrence of the two con-sonants; the consonants were there but werenot in the correct sequence. This meansthat, as suggested by Stromer and Mackay(1992b), further discrimination training forsequencing may be required to overcomesuch CC errors. The training methods usedhere appeared to reduce only the reverse CCerrors. The techniques used by Stromer andMackay (1992b) and Stromer, Mackay, Co-

hen, and Stoddard (1993) seem to be prom-ising for reducing sequencing errors.The computer tasks were extremely useful

in training the spelling of nondisabled chil-dren. However, these children exhibitedsome skills that might not have been presentin the persons with intellectual disabilitieswho participated in previous research (Dubeet al., 1991; Stromer & Mackay, 1992b),and supplementary training might be need-ed to replicate the present results with suchpopulations. Further applications of basic

stimulus control procedures to spellingshould result in improved spelling. One ex-ample is the finer discriminative control ofsequencing mentioned above (Green, Strom-er, & Mackay, 1993).

Of most practical significance was that thecomputer-trained spelling, as in the work ofDube et al. (1991) and Stromer and Mackay(1992a, 1992b), did not require direct in-tervention by a teacher. If normal readingwas to be supplemented by computer tasksof the kind used here and elsewhere (Strom-

er & Mackay, 1992b), some spelling prob-lems, which might occur because naturalcontingencies are not sufficient to shape thefiner discrimination of individual letters,could be circumvented without costly inter-vention by a teacher or a special trainer. The

8/7/2019 Guerin 1997

12/23


present research suggests that the applicationof selective stimulus control research tospelling merits further attention.

REFERENCES

Allen, K. D., & Fuqua, R. W. (1985). Eliminatingselective stimulus control: A comparison of twoprocedures for teaching mentally retarded childrento respond to compound stimuli. Journal of Ex-perimental Child Psychology, 39, 5571.

Bruck, M., & Treiman, R. (1990). Phonologicalawareness and spelling in normal children anddyslexics: The case of initial consonant clusters.Journal of Experimental Child Psychology, 50, 150178.

Cuvo, A. J., Ashley, K. M., Marso, K. J., Zhang, B.L., & Fry, T. A. (1995). Effect of response prac-tice variables on learning spelling and sight vocab-

ulary. Journal of Applied Behavior Analysis, 28,155173.

Dube, W. V., McDonald, S. J., McIlvane, W. J., &Mackay, H. A. (1991). Constructed-responsematching to sample and spelling instruction. Jour-nal of Applied Behavior Analysis, 24, 305317.

Foorman, B. R., Francis, D. M., Novy, D. J., & Lib-erman, D. (1991). How letter-sound instructionmediates progress in first grade reading and spell-ing. Journal of Educational Psychology, 83, 456469.

Gettinger, M. (1985). Effects of teacher-directed ver-sus student-directed instruction and cues versusno cues for improving spelling performance. Jour-nal of Applied Behavior Analysis, 18, 167171.

Gettinger, M. (1993). Effects of invented spelling anddirect instruction on spelling performance of sec-ond-grade boys. Journal of Applied Behavior Anal-ysis, 26, 281291.

Green, G., Stromer, R., & Mackay, H. A. (1993).Relational learning in stimulus equivalence. ThePsychological Record, 43, 599616.

Greenwood, C. R., Dinwiddie, G., Bailey, V., Carta,J. J., Dorsey, D., Kohler, F. W., Nelson, C., Roth-olz, D., & Schulte, D. (1987). Field replicationof classwide peer tutoring. Journal of Applied Be-havior Analysis, 20, 151160.

Griffith, P. L. (1991). Phonemic awareness helps firstgraders invent spellings and third graders remem-ber correct spellings. Journal of Reading Behavior,

23, 215233.Lee, V. L., & Pegler, A. M. (1982). Effects on spellingof training children to read. Journal the Experi-mental Analyses of Behavior, 37, 311322.

Lee, V. L., & Sanderson, G. M. (1987). Some con-tingencies of spelling. The Analysis of Verbal Be-havior, 5, 13.

Lewkowicz, N. K. (1980). Phonemic awareness train-

ing: What to teach and how to teach it. Journalof Educational Psychology, 72, 686700.

Lovaas, O. I., Koegel, R. L., & Schreibman, L.(1979). Stimulus overselectivity in autism: A re-view of research. Psychological Bulletin, 86, 12361254.

Lovaas, O. I., Schreibman, L., Koegel, R. L., & Rehm,R. (1971). Selective responding by autistic chil-dren to multiple sensory input. Journal of Abnor-mal Psychology, 77, 211222.

Marcel, T. (1980). Phonological awareness and pho-nological representation: Investigation of a specificspelling problem. In U. Frith (Ed.), Cognitive pro-cesses in spelling (pp. 373405). London: Academ-ic Press.

Miller, P., & Limber, J. (1985, October). The acqui-sition of consonant clusters: A paradigm problem. Pa-per presented at the annual Boston UniversityConference on Language Development, Boston.

Moxley, R. A., & Warash, B. (19901991). Spellingstrategies of three prekindergarten children on the

microcomputer. Journal of Computing in Child-hood Education, 2, 4761.Neef, N. A., Iwata, B. A., & Page, T. J. (1980). The

effects of interspersal training versus high-densityreinforcement on spelling acquisition and reten-tion. Journal of Applied Behavior Analysis, 13,153156.

Ollendick, T. H., Matson, J. L., Esveldt-Dawson, K.,& Shapiro, E. S. (1980). Increasing spellingachievement: An analysis of treatment proceduresutilizing an alternating treatments design. Journalof Applied Behavior Analysis, 13, 645654.

Read, C. (1975). Childrens categorization of speechsounds in English(NCTE Research Rep. No. 17).Urbana, IL: National Council of Teachers of En-

glish.Reynolds, B. S., Newsom, C. D., & Lovaas, O. I.(1974). Auditory overselectivity in autistic chil-dren. Journal of Abnormal Child Psychology, 2,253263.

Schover, L. R., & Newsom, C. D. (1976). Overse-lectivity, developmental level, and overtraining inautistic children. Journal of Experimental ChildPsychology, 4, 289298.

Schreibman, L., Koegel, R. L., & Craig, M. S. (1977).Reducing stimulus overselectivity in autistic chil-dren. Journal of Abnormal Child Psychology, 5,285289.

Stevens, K. B., Blackhurst, E., & Slaton, D. B.(1991). Teaching memorized spelling with a mi-

crocomputer: Time delay and computer-assistedinstruction. Journal of Applied Behavior Analysis,24, 153160.

Stromer, R., & Mackay, H. A. (1992a). Delayed con-structed-response identity matching improves thespelling performance of students with mental re-tardation. Journal of Behavioral Education, 2, 139156.

8/7/2019 Guerin 1997

13/23


Stromer, R., & Mackay, H. A. (1992b). Spelling andemergent picture-printed word relations estab-lished with delayed identity matching to complexsamples. Journal of Applied Behavior Analysis, 25,893904.

Stromer, R., Mackay, H. A., Cohen, M., & Stoddard,

L. T. (1993). Sequence learning in individualswith behavioural limitations. Journal of IntellectualDisability Research, 37, 243261.

Stromer, R., Mackay, H. A., & Stoddard, L. T.(1992). Classroom applications of stimulus equiv-alence technology. Journal of Behavioral Education,2, 225256.

Treiman, R. (1991). Childrens spelling errors on syl-lable-initial consonant clusters. Journal of Educa-tional Psychology, 83, 346360.

Treiman, R. (1993). Beginning to spell: A study of first-grade children. New York: Oxford University Press.

Treiman, R., & Zukowski, A. (1988). Units in read-ing and spelling. Journal of Memory and Language,27, 466477.

Treiman, R., & Zukowski, A. (1990). Toward an un-

derstanding of English syllabification. Journal of Memory and Language, 29, 6685.

Received October 26, 1995Initial editorial decision December 20, 1995Final acceptance October 25, 1996Action Editor, Joseph E. Spradlin

STUDY QUESTIONS

1. What are words containing consonant clusters (CCs), and how are these words typicallymisspelled?

2. Describe the rules used for generating the experimental words and both the critical-differenceand multiple-difference comparison words. Given shot as an experimental word, provideexamples of its multiple-difference and critical-difference comparisons.

3. Describe the basic procedural difference between the training trials and the constructed-response trials.

4. How did the authors control for the different training approaches used?

5. What types of CC errors did the authors record?

6. Summarize the results in terms of (a) percentage correct during training and constructedresponse trials, (b) number of trials to criterion, and (c) CC errors.

7. What explanation was provided to account for the performance differences that were ob-served during training versus testing?

8. The authors suggested that performance during the multiple-difference condition might havebeen worse than that obtained had the children not been exposed to critical-differencetraining. How might the authors have controlled for the influence of critical-differencetraining?

Questions prepared by SungWoo Kahng and Michele Wallace, University of Florida

8/7/2019 Guerin 1997

14/23


APPENDIX

Individual data for the initial oral test (P) and constructed-response test sessions, shown separately for thecritical-difference and multiple-difference training conditions. At the beginning of the first session, the three words

for training were retested (0). Words that were being trained at the time of any test are shown in capitals. Occasionalprobes for the next words to be trained are shown in lower case letters. Criterion was three correct consecutive

spellings, at which point a new word was introduced to training.

Participant Session

Anna

Critical-difference training

P0123

sowsowSNOWSNOWSNOW

winwenWNWENWHEN

bidbedBEDBEDBED

sda

sea

rat

456

WENWHENWHEN

BREDBREDBRED

STAYSTAYSTAY

ret

78

9

WHINWIEN

WHEN

WRITWRET

WRIT1011

WHENWHEN

WRITWRIT

Multiple-difference training

P0123

sapcipCAPCHAPCAP

fogfogFOGFOGFOG

natnatNATGATGNAT

twn gwin

45678

CAPCHAPCHAPCHAP

FROGFROGRFOGFROFROG

GNATGNAT

twn

TIWNTWNTWIN

gwen

gwinGWEN

91011

FROGFROG

TWINTWIN

GWENGWEN

Bella


P012

cepcipCLIPCLIP

natnatGNATGNAT

towentowinTEWINTIWEN

fog

fog

trwr

34567

CLAPCLIPCLIPCLIP

GNAT TIWNTWINTEWINTIHWNTWIN

FROGFROGFOGFROG

thrig

TIRG8

9101112

TWIN

TWIN

FROG

FROG

TRGE

TRIGTHRGTIRGTRIG

1314

TRIGTRIG

8/7/2019 Guerin 1997

15/23


APPENDIX

(Extended)

bli sca s

show

pla scm

blewu play skimBLWBLEW

BLEW

scabshow

siow

play

play

skim

BLEW scabscab

showshow skim

clp fla tha

they

gro

clip

clip

flay

flay thay

grow

grwclip

flay they they

grow

grow

flak rrop glrad gren thay

they

flye crihp

glad gwen

FLAYFLAYFLAY

CROPCROPCROP

glad

glad

gwen

gwen

they

they

8/7/2019 Guerin 1997

16/23


APPENDIX

(Continued)

Participant Session


P0123

sgemsimSEIMSIGMSKIM

wenewenWENWHENWEN

bolebluBLOWBEWBLWE

rit

wrt

seteay

wsta

4567

SKIMSKIM

WHENWHENWHEN

BIEWBELWBEWBLWE

WITWEIT SITY

89

101112

BLEWBLEWBLEW

WIRTWRITWRITWIRTWRIT

STYESTAYSITYSTIYSTAY

13

14151617

WRIT

WRIT

SATY

SATYSTAYSTAYSTAY

Colin


P0123

ssoSNOWSNOWSONW

lowbwBOLWBLOWBLEW

socsokSMOGSMRGSMOG

sme

skm

reat

45678

SNOWSNOWSNOW

BLAWBLEWBLIWBLOWBLEW

SMOGSMOG

SIKMSIKMSKIM

bied

BREDBDEM

910111213

BLEWBLEW

SKIMSKIM

BREDBREDBRED

14151617


P01

23

wooiroiwGOW

GROWGORW

decfrikTREG

TERGGTR

copicopCROP

GOPCORP

foc

forg

mit

45678

GROWGROWGROW

TREGTROGTREGTROGTRIG

CROPCROPCORPCORPCORP

FORGFORG

nat

8/7/2019 Guerin 1997

17/23


APPENDIX

(Extended - Continued)

sgarb selet brerde

sacb slat bred

SCBESECB

clat bred

CGRB

SCABSCABSCAB

slatslat

slat

bred

sti whit sab sow sat

sta

STAYSTAYSTAY

ritWRITWRITWRIT

caebSCABSCAB

seowslat

saitSCAB SHOW

SHOWSHOW

slat

slatslat

tin ilep ceni dladi fae thai

tein

8/7/2019 Guerin 1997

18/23


APPENDIX

(Continued)

Participant Session

9

10

TRIG

TREG

CORP

CORP

FORG

FORG1112131415

TRIGTRIGTRIG

CROPCROPCROP

FROGFROGFROG

GNATGNAT

16171819

GNAT

David


P0123

croiwgrwGROWGROWGROW

corpcopCROUPCROPCRUP

tewntiwenTEWINTIWNTAWIN

cilp

cihp

tha

toey45678

CROPCROPCROP

TWINTWINTEWINTWINTWIN

CLIBCLPCLIPCLIPCLIP

THAYTHEY

9 TWIN THEY 101112131415

THEY

Multiple-difference trainingP0123

sgiskmSKMSIMSKM

soiwsoiwSOWSHOWSHOW

wenwenWENWENWIEN

pla

pla

beed

bied45678

SKMSKIMSCIMSKIMSKIM

SOWSHOWSHOWSHOW

WHENWHENWHEN

PLAYPLAY BED

9101112

13

SKIM PLAY BEDBEDBREDBRED

BRED14151617

8/7/2019 Guerin 1997

19/23


APPENDIX


TINTIN

cilp

CIPCLIP

gwenGEWN

gaed

giadGLAD

flay

flaythay

theyTWINTWINTWIN

CILPCLIPCLIPCLIP

GENGWENGWENGWEN

GADGLADGLADGLAD

flay

flay

they

they

cap gen nat gilad flaa trik

cehap winnat glad

CAP gladCHAPCHAPCHAP

GWNGWNGWENGWENGWEN

GNATGNATGNAT

gladflay

flay

flay

tig

TRGTRIGTRIGTRIG

rit bliw sta so siat sab

rit bw stae slat

RITRITWRAIT

RIT

BLUEWBLEWBLEW

BLEW

swslat

slat

scab

scab

RITWRITWRITWRIT

STASTAYSTAYSTAY

SOWSNOWSNOWSNOW

scab

8/7/2019 Guerin 1997

20/23


APPENDIX

(Continued)

Participant Session

Elena

Critical-difference trainingP

0123

sasaeSTAESATYSTAY

bluebleBLEWBLEWBLEW

riteritWRITWRITWRIT

scm

sm

bed

bed45678

STASTAYSTAYSTAY

SKMSKIMSKIMSKIM

BREDBREDBRED

91011


P0

sapcab

natnat

tritik

flai tin

12345

CAPCHAPCAPCHAPCHAP

GNATGHATGNATGNATGNAT

TIGTRICTRIGTRIGTRIG

flai twn

6789

1011

CHAP FLAEFLAIFLAFLAYFLAYFLAY

TWINTIWENTWINTWINTWIN

Fiona

Critical-difference trainingP

0123

filfaiFLAIFLAIFILAY

tiktikTIGTRITRIG

gadgadGLADGLADGLAD

lip

cilp

gwn

gwenm4567

FLAYFLAY

TIRGTRIGTRIGTRIG

CLIPCLPCLIPCLIP GWEHN

89

101112

CLIP GWENGWENGWEN

1314

8/7/2019 Guerin 1997

21/23


APPENDIX


plai

plai

sab smo

PLAYPLAY

scrb

SAB

smog

PLAY SCABSCABSCAB

smogsmog

wen cip

gwin

GWEMGWINGWENGWENGWEN

clip

clipclip

twn fog chab nat

twnm frogcap

natTWNTWUNTWINTWENTWIN

frog

frog

CHAPCHAPCHAP GNAT

GNAT

TWINTWIN

GNAT

8/7/2019 Guerin 1997

22/23


APPENDIX

(Continued)

Participant Session


P0123

stsateSTASTAISTAY

blubluBLWBILWBLUW

sgbcabSABSCABSAB

sm

sm

rit

45678

STAESTAESTAYSTAYSTAY

BLIWBLEWBLIWBLEWBLEW

SGABSCABSCABSCAB

SGIM

rit

9101112

BLEW SKIMSKMSKMSKIM

RITRITWRITWRIT

13

141516

SIM

SKIMSKIMSKIM

WRIT

8/7/2019 Guerin 1997

23/23


APPENDIX


sat smoc sow bireed

satsmok

show

bred

SLATSLATSLAT

show bred

SOG

SMOGSMOGSMOG

show bred

Documents

Guerin 1997