ERIC EJ910173: The Effects of Behavioral Skills Training on Mand Training by Staff and Unprompted Vocal Mands by Children

2010·0.23 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2010, 43, 757–761 NUMBER4 (WINTER2010)
THE EFFECTS OF BEHAVIORAL SKILLS TRAINING ON MAND
TRAINING BY STAFF AND UNPROMPTED VOCAL MANDS
BY CHILDREN
DARLENE NIGRO-BRUZZI AND PETER STURMEY
QUEENSCOLLEGEANDTHEGRADUATECENTER
CITYUNIVERSITYOFNEWYORK
Weevaluatedtheeffectsofatrainingpackage,includinginstructions,modeling,rehearsal,and
feedback, for training staff members to conduct mand training with children. Experimenters
collected data on staff performance on each step of a task analysis of mand training and on
unprompted child vocal mands. Training resulted in increases in staff performance in mand
trainingandinunpromptedmandsbychildren.Weobservedreplicationoftheseeffectsacross
settings for allstaff andfor 3 ofthe children.
Keywords: mandtraining, staff training
_______________________________________________________________________________
Communication deficits among children are One strategy experimenters have used to
often associated with problem behavior (Mat- promotegeneralityofmandtrainingeffectsisto
son, Boisjoli, & Mahan, 2009; Stevenson & train staff or parents to conduct mand training
Richman,1978).Inaddition,problembehavior (Laski,Charlop,&Schreibman,1988).Laskiet
may serve a similar social function to that al. demonstrated that training was effective in
produced by appropriate forms of communica- promoting mand training among staff and
tion (Carr & Durand, 1985; Wacker et al., parents; however, lengthy instructions were
2005). In these cases, an effective form of required. Behavioral skills training (BST) is a
treatmentinvolveswithholdingthemaintaining training package that is effective for promoting
reinforcer for problem behavior and delivering acquisition of discrete-trial teaching responses
itcontingentonanappropriatecommunication (Sarokoff & Sturmey, 2004) and preference
response, often referred to as a mand. Skinner assessment implementation (Roscoe & Fisher,
(1957) defined a mand as a response that one 2008), which includes instructions, modeling,
reinforces with a specific consequence and is rehearsal, and feedback. It is unclear whether
under the control of deprivation or aversive BST would be effective and efficient when used
stimulation (pp. 35–36). Given the clinical to train staff to implement mand training.
importanceofestablishinganappropriatemand Therefore, the purpose of the current study was
repertoire among children with communication to evaluate the effectiveness of BST for training
deficits, experimenters have developed a num- staff to implement mand training.
ber of mand training procedures.
METHOD
A grant from the Graduate Center, CUNY supported
Participants, Setting, and Stimuli
part of this research, which was part of the first author’s
doctoral dissertation. We thank Volunteers of America, Six children with autism spectrum disorders
the Children’s Home Intervention Program, Inc., Haven participated. The children in Dyads 1, 2, and 3
Bernstein,TinaRovitoGomez,andJacquelineMarksfor
were 4 years old and attended school. The
theirassistance withthe study.
Pleaseaddresscorrespondenceconcerningthisarticleto childreninDyads5and6were2yearsold,and
DarleneNigro-Bruzzi,DepartmentofPsychology,Queens the child in Dyad 4 was 6 years old. These
College, CUNY, Flushing, New York 11367 (e-mail:
children received in-home services. All children
[email protected]).
doi:10.1901/jaba.2010.43-757 imitated vocal models. Six staff (three special
757
758 DARLENE NIGRO-BRUZZI and PETER STURMEY
Table 1
Task Analysis ofMand Training
1. Placepreferredleisureitemsonthefloor,with3cmbetweeneachtoy.
2. Bringthechildafewcentimetersawayfromtheleisureitemsandinstructhimorherto‘‘[get],[find],[showme],[pointto]theone
youwant.’’
3. Level1:Whenthechildtouchesaleisureitem,immediatelyremoveitfromhisorherhandsandholditupafewcentimetersinfront
ofthechild’sfaceuntilheorshemandsfortheitemor3selapses.Ifthechildmands,presenthimorherwiththeleisureitemfor
1minandrecordthenameoftheitemandthecorrespondingpromptlevelonthedatasheet.
4. Level2:IfthechilddidnotmandduringLevel1,askhimorher‘‘Whatdoyouwant?’’Ifthechildmandswithin3s,delivertheitem
andrecorddataasdescribedforStep3.
5. Level3:IfthechilddidnotmandduringLevel2,presenttheitemwithavocalapproximation(e.g.,‘‘mm’’formusic).Ifthechild
subsequentlymandswithin3s,delivertheitemandrecorddataasdescribedforStep3.
6. Level4:IfthechilddidnotmandduringLevel3,presenttheitemwiththecompletevocalmodel(e.g.,‘‘music’’).Repeatthevocal
modelthreetimeswith3sbetweenprompts.Ifthechildsubsequentlymandswithin3s,delivertheitemandrecorddataasdescribed
forStep3.
7. IfthechilddoesnotmandduringLevel4,placetheitembackinthearray;recordthenameoftheitemonthedatasheet.
8. Afterthechildhashadaccesstothemandeditemfor1minordidnotmandduringLevel4,reinitiatemandtrainingstartingatStep
1.Repeatuntiltheexperimenterinstructsyoutostop.
education teachers and three speech therapists), initiatedbyachild(e.g.,whenachildtoucheda
who worked with the children on a regular toy). During each mand opportunity, only
basis, participated. None had received training certain steps of the task analysis were applicable
in teaching manding. We paired the staff and for data collection, depending on whether a
the children into dyads, and each staff member staff member or a child initiated the mand
worked with the same child throughout the opportunityanddependingonthepromptlevel
study. Staff training and mand training sessions in effect when the child emitted a mand. For
wereconductedinthechild’sclassroom,aroom example, if a staff member initiated a mand
in their home, or a private office (staff training opportunity and the child emitted a mand at
only). Staff conducted generality probes in a Level 1 (Step 3), then only Steps 1, 2, 3, and 8
room outside the child’s classroom. Experi- were applicable for data collection, whereas if
menters identified five to eight items for use the child initiated an opportunity and per-
during mand training with each participant formed a mand at Level 1 (Step 3), then only
through staff and parent interviews and prefer- Steps 3 and 8 were applicable for data
ence assessments (DeLeon & Iwata, 1996). collection. For each applicable training step,
Experimenters taped all sessions. experimenters scored a correct response if the
staff completed all components associated with
Response Measurement and that step in the sequence listed in the task
Interobserver Agreement analysis. If a staff member skipped a prompt
Aneight-stepmand-trainingtaskanalysiswas level, experimenters scored that prompt level as
developed (see Table 1). Each training step incorrect. The experimenter recorded staff
consisted of multiple components (e.g., behavior during staff training role play in the
prompting, delivery of the item, and data same manner. We calculated the percentage of
collection) to replicate the complexity of mand correct staff responses by adding the number of
training in the natural environment where staff applicable steps scored as correct, dividing that
must perform multiple responses in a short numberbythetotalnumberofapplicablesteps,
period of time. Observers collected data for and converting the ratio to a percentage.
each mand opportunity, including those initi- Observers scored the number of mand
atedbystaff(e.g.,whenthestaffaskedachildto opportunities and the corresponding prompt
‘‘find the one he or she wants’’) and those level (Levels 1 to 4) in effect when the first
TRAINING STAFF TO TEACH MANDING 759
mand was emitted for a given opportunity; member a video depicting all steps listed in
subsequent mands emitted within the same Table 1. Next, the experimenter role played the
opportunity were ignored and were excluded childwhilethestaffmemberpracticedthesteps.
from data collection. Experimenters summa- The experimenter provided feedback (positive
rized data on the percentage of Level 1 mands and corrective statements) on performance
by adding mands emitted during the Level 1 during role plays. Training sessions continued
step, dividing that number by the number of until staff performed 90% correct across three
mandopportunities,andconvertingtheratioto consecutive sessions. All staff members met
a percentage. Experimenters summarized data criterion in three sessions.
on percentage of steps performed correctly by Posttrainingsessions.Sessionswereidenticalto
dividing the number of applicable steps per- baseline.
formed correctly by the total number of Generality sessions. We assessed generality of
applicable steps and converting the ratio to a children’s manding and staff’s correct perfor-
percentage. Experimenters collected interob- mance across settings. Generality sessions were
server agreement data during a minimum of
similar to baseline and lasted 5 min.
20% of sessions across participants. Experi-
menters summarized agreement scores by
RESULTS AND DISCUSSION
dividing the number of agreements by the
number of agreements plus disagreements and
Figure 1 depicts the percentage of staff
converting the ratio to a percentage. Mean
members’ correct performance of applicable
agreementforstepsacrossstaffwas90%(range,
task analysis steps and Level 1 child mands.
83% to 97%), and mean agreement for mands
Duringbaseline,Dyads1,2,4,5,and6showed
acrosschildrenwas92%(range,77%to100%).
low percentages of staff members’ correct
performance and Level 1 child mands. Because
Procedure
Dyad3hadanincreasingtrendduringbaseline,
Experimenters used a multiple baseline
experimenters did not implement training with
design across participants. Baseline and post-
this dyad. During posttraining, percentages of
training sessions lasted 20 min.
staff members’ correct performance and Level 1
Baseline. During Session 1, the experimenter
child mands increased for Dyads 1, 2, 4, 5, and
presented staff with written instructions and a
6. The increases in mands for Dyad 4 were
copy of Table 1 and read its contents aloud.
lower than those observed for the other dyads.
The instructions described the definition of a
For generality probes conducted in the context
mand,howtoconductmandtraining,andhow
of baseline, staff and children in all dyads
to collect data on child performance. In
subsequent sessions, staff received an abbreviat- showed outcomes similar to those observed
ed set of instructions that included the purpose during their baseline sessions. For generality
and length of the session. The experimenter probes conducted in the context of posttraining
answered staff members’ questions before every sessions, staff andchildreninDyads 1,2,and 4
session. Staff rarely asked questions, and they showed outcomes similar to those observed
typically involved procedural clarification relat- during their posttraining sessions. By contrast,
ed to data collection and the task analysis. only staff in Dyads 5 and 6 showed outcomes
Stafftraining.During30-to60-minsessions, similartothoseobservedduringtheirposttrain-
the experimenter provided instructions, a video ing sessions; the children in these dyads showed
model, role-play rehearsal, and performance discrepant responding (i.e., low percentages of
feedback. After reading the abbreviated instruc- Level1mands)fromtheirposttrainingsessions.
tions, the experimenter showed the staff Table 2 shows the range of Level 1 mands and
760 DARLENE NIGRO-BRUZZI and PETER STURMEY
Figure 1. Percentage of Level 1 child mands (circles) and percentage of staff members’ correct performance of
applicable task analysis steps (squares). Open data points represent responses during sessions, and filled data points
representresponses during generality probes.
mand opportunities and demonstrates that a may be due to her having previously acquired a
greater number of mand opportunities in manding repertoire for a particular item
baseline compared to posttraining alone did (movie) that was present in the generality
not result in more Level 1 mands for Dyads 2 setting.
and 5. This study extended previous research on
Poor integrity of mand training during BST by showing that experimenters can use it
baseline was often due to staff presenting only to train staff to teach children with autism. In
some prompt levels (e.g., Level 4, a model) and addition, the mand training procedure was
omitting others (e.g., Level 1, initiating a effective in increasing independent mands in
delay). During baseline, the children’s mands childrenwithautism,andthiseffect generalized
occurredmostoftenwhenstaffinitiatedamand across settings for three of the five children. An
opportunity(atLevel2or4);thechildinDyad important feature of the staff training program
2 was the only exception, in that she frequently was that staff completed BST in fewer than
initiatedmandopportunities.Duringposttrain- three 60-min sessions. The results suggest that
ing, mands occurred most often at Level 1. It is experimenters can implement the mand train-
unclear why mand training generalized across ing procedure across settings with minimal
settings for children in Dyads 1, 2, and 3 but effort for staff and concomitant clinical benefit
notfor children in Dyads 5 and 6. The child in to children with autism.
Dyad 2 showed a high percentage of Level 1 Experimenters note some limitations. First,
mands during baseline generality probes; this we could not evaluate the effects of the
TRAINING STAFF TO TEACH MANDING 761
Table 2
Rangeof Level1 ChildMands andMand Opportunities
Baseline Posttraining
Dyad Sessions Generalityprobes Posttraining Generalityprobes
1 0–2,13–20 1,4 10–19,15–19 4–7,4–7
2 0–24,28–40 5,6–7 18–19,19–20 5,7
3 3–23,21–36 0–6,4–6
4 0,16–20 0,8 2–10,18–24 0–1,6
5 2–9,13–15 0,2 14–18,15–25 0,2–5
6 0–14,20–42 2–10,11–12 16–21,16–21 0,5
Note.ThetableliststherangeofLevel1mandsandthenumberofopportunitiesforeachcondition.Experimenters
didnotobtain posttraining data for Dyad3.
independent variable with the staff member in Laski, K. E., Charlop, M. H., & Schreibman, L. (1988).
Trainingparentstousethenaturallanguageparadigm
Dyad 3 because she exhibited increases in
to increase their autistic children’s speech. Journal of
correct performance during baseline. Second, Applied BehaviorAnalysis,21, 391–400.
each step of the task analysis consisted of Matson, J. L., Boisjoli, J., & Mahan, S. (2009). The
relation of communication and challenging behavior
multiple staff responses resulting in a less
ininfantsandtoddlerswithautismspectrumdisorder.
sensitive dependent variable (i.e., it was unclear
JournalofDevelopmentalandPhysicalDisabilities,21,
whether an error was due to one or more 253–261.
incorrect responses). Third, the number of Roscoe,E.M.,&Fisher,W.W.(2008).Evaluationofan
efficient method for training staff to implement
mand opportunities differed across sessions,
stimulus preference assessments. Journal of Applied
whichmayhaveaffectedthepercentageofLevel Behavior Analysis,41,249–254.
1 mands. Potential areas for future research Sarokoff, R. A., & Sturmey, P. (2004). The effects of
behavioral skills training on staff implementation of
include teaching staff to implement mand
discrete-trial teaching. Journal of Applied Behavior
training using a pyramidal approach and Analysis,37,535–538.
modifying the task analysis for group instruc- Skinner, B. F. (1957). Verbal behavior. New York:
tion to further increase staff training efficiency. Appleton-Century-Crofts.
Stevenson,J.,&Richman,N.(1978).Behavior,language
and development in 3-year-old children. Journal of
REFERENCES Autism andChildhoodSchizophrenia,8,299–313.
Wacker,D.P.,Berg,W.K.,Harding,J.W.,Barretto,A.,
Carr,E.G.,&Durand,V.M.(1985).Reducingbehavior Rankin, B., & Ganzer, J. (2005). Treatment
problems through functional communication train- effectiveness,stimulusgeneralizationandacceptability
ing. Journal of Applied Behavior Analysis, 18, to parents of functional communication training.
111–126. Educational Psychology, 52,233–256.
DeLeon, I. G., & Iwata, B. A. (1996). Evaluation of a
multiple-stimulus presentation format for assessing Received March 12,2009
reinforcer preferences. Journal of Applied Behavior Final acceptanceMarch 2,2010
Analysis, 29,519–533. Action Editor,Eileen Roscoe