ERIC EJ767621: Evaluating Preschool Children's Preferences for Motivational Systems during Instruction

2007·0.26 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2007, 40, 249–261 NUMBER2 (SUMMER2007)
EVALUATING PRESCHOOL CHILDREN’S PREFERENCES FOR
MOTIVATIONAL SYSTEMS DURING INSTRUCTION
NICOLE A. HEAL AND GREGORY P. HANLEY
UNIVERSITYOFKANSAS
Preschool teachers rely on several strategies for motivating children to participate in learning
activities.Inthecurrentstudy,weevaluatedtheeffectivenessofandpreferenceforthreeteaching
contexts in which embedded, sequential, or no programmed reinforcement was arranged. The
embeddedcontextincludedhighlypreferredteachingmaterials,thesequentialcontextincluded
highlypreferrededibleitemsforcorrectresponding,andacontrolcontextincludedneither.In
addition,anexclusivelyplay-orientedactivitywasincludedasafourthoptiontodetermineifone
of the direct teaching contexts could compete with a relatively unstructured and exclusively
child-led activity. All participants preferred the sequential context (use of high-quality
consequences) over the embedded context (use of high-quality teaching materials), 2 of the 4
participants preferred some motivational system to none at all, and the play area was selected
overallvariantsoftheinstructionalcontextsduringthemajorityoftrials.Wefoundeithernoor
smalldifferencesincorrectrespondinginthedifferentinstructionalcontexts;however,ratesof
undesirable behavior were highest in the least preferred interaction area for 3 of the 4
participants. Implications for the design of effective and preferred teaching environments for
young children arediscussed.
DESCRIPTORS: choice, concurrent-chains arrangement, embedded reinforcement, in-
struction,preschoolers, sequentialreinforcement
_______________________________________________________________________________
Determining children’s preferences for dif- selection of practices, such as relative efficacy or
ferent classroom contexts is important because, implementation effort, are similar across prac-
inadditiontoefficacymeasures,preferencedata tices.
may influence the adoption of classroom Anumberofmethodshavebeendevelopedfor
practices. The acceptability of behavioral inter- determiningyoungchildren’spreferencesamong
ventions has often been determined by inviting food, toys, activities, and curriculum materials
teachers or caregivers to complete rating scales (DeLeon & Iwata, 1996; Fisher et al., 1992;
or questionnaires (Miltenberger, 1990). How- Foster-Johnson,Ferro,&Dunlap,1994;Roane,
ever, there are distinct advantages to determin- Vollmer,Ringdahl,&Marcus,1998).Inspiteof
ing the acceptability of different teaching these advancements, procedures for accurately
practices with the learners themselves. For determining preschoolers’ preferences for class-
example, Dunlap et al. (1994) showed that room practices or teaching strategies have not
incorporating a child’s preferences into in- been described in the research literature. A
structional contexts resulted in less off-task general strategy for directly determining prefer-
and disruptive behavior. Information about ences for protracted events or behavioral inter-
children’s preferences may be most important ventions has, however, been described in several
when factors that would typically influence the studies (Hanley, Iwata, & Lindberg, 1999;
Hanley, Piazza, Fisher, Contrucci, & Maglieri,
Theresearchpresentedinthisarticlewascompletedin 1997;Hanley,Piazza,Fisher,&Maglieri,2005),
partialfulfillmentofthesisrequirementsbythefirstauthor
and these procedures may be applicable for
atthe University ofKansas.
RequestsforreprintsshouldbeaddressedtoGregoryP. determining young children’s preferences for
Hanley, Applied Behavioral Science Department, 4023 classroom practices. Common features in these
Dole Center, University of Kansas, Lawrence, Kansas
studies were the use of (a) modified concurrent
66045(e-mail: [email protected]).
doi:10.1901/jaba.2007.59-05 chains arrangements with (b) individuals who
249
250 NICOLE A. HEAL and GREGORY P. HANLEY
could not readily express their biases due to teaching materials that areattractive (e.g.,color-
(c) severe language impairments or the complex ful,iconic,multidimensional)andthoughttobe
nature of the assessed events. For example, generally preferred by children are used. Thus,
Hanleyetal.(1997,2005)determinedchildren’s childrenmayautomaticallyderivereinforcement
preferences among different function-based simply by manipulating instructional materials.
treatments for problem behavior by recording By contrast, preferred stimuli are systematically
each child’s selections in initial links of concur- identified and then used as consequences for
rent chains in which different treatments were correct responding with the sequential motiva-
presented in terminal links. In initial links, tional system. Therefore, in addition to the
children pressed colored switches located on manner in which putative reinforcers are de-
a table outside the treatment room. Pressing livered (embedded in materials or provided by
a switch resulted in exposure to a treatment a teacher after a correct response), the type and,
procedure inside the room, and each switch was presumably, quality of reinforcers differ across
associated with a different treatment. Following motivational systems. Although both systems
repeatedexposuretothisarrangement,children’s have been evaluated in preschool classrooms
responding in the initial link differentiated, (Bryant&Budd,1982;Driscoll&Nagel,1999;
suggesting preferences among the treatments. Gordon & Williams-Browne, 2000; Warren,
In the current study, we applied a similar Rogers-Warren, & Baer, 1976; Zanolli &
concurrent-chains arrangement to determine Daggett, 1998), their relative effectiveness or
preschoolers’ preferences for different teacher- valuetothelearnershasnotyetbeendetermined.
led instructional practices. In the current study, either embedded or
Two general strategies for motivating chil- sequential reinforcement was implemented in
dren to participate in teacher-led instruction two contexts. Thus, the teaching contexts varied
havebeendescribed.First,reinforcershavebeen along two primary dimensions: the type and
timing of reinforcement. We also included an
embedded within the task itself (Essa, 2003;
exclusivelyplay-orientedactivitydevoidofteach-
Herr & Libby, 1990; Mayesky, 1998). Second,
er instructions and an instructional context
reinforcers have been delivered following com-
devoid of either motivational system in our
pletion of the task (Daly, Martens, Hamler,
evaluation. Thus, four contexts were included as
Dool,&Eckert,1999;Freeland&Noell,1999;
terminal links in our concurrent-chains arrange-
Rincover & Newsom, 1985). The relative
ment: instructions with embedded reinforce-
efficacy of embedded and sequential reinforce-
ment,instructionswithsequentialreinforcement,
ment has been evaluated during skills training
instructions only (control), and no instructions
with adults with severe disabilities (Thompson
(play). We compared measures of children’s
& Iwata, 2000) and interventions aimed at
correct responding and undesirable behavior
increasing children’s consumption of nonpre-
while they experienced the different teaching
ferred foods (Kern & Marder, 1996). By
conditions to determine their relative effective-
keeping the type of reinforcement constant
ness. We determined children’s preferences
across conditions and manipulating only the
among the four classroom contexts by allowing
manner in which reinforcement was delivered,
and measuring their selections of each context.
these studies provided strong evidence for the
superiority of the embedded procedure.
METHOD
However, sequential and embedded motiva-
tional systems differ in other ways when in- PARTICIPANTSAND SETTING
tegrated within teacher-led instructional activi- The participants attended a full-day inclusive
ties. With embedded motivational systems, preschool program that served typically and
PREFERENCES FOR MOTIVATIONAL SYSTEMS 251
atypically developing children ranging from 3.5 50%ofallsessionsacrossallparticipants(range,
to 5 years of age. Charlie was 5 years old with 50% to 90%). Agreement was scored when
a diagnosis of nonspecified developmental de- both observers recorded the same card selection
lay. James, Amy, and Anne were typically within a trial; coefficients were calculated as
developing4-year-olds.Inadditiontoobtaining described previously. Agreement was 100%
informed consent from parents, assent was across all participants.
obtained from the participants each day prior Performance data also were collected during
to conducting sessions. Children were selected the interaction area preference assessment. A
for participation based on their and the first mean task-completion score was generated by
author’s availability. assigning points for correct responding to vocal
Observations took place in a large unoccu- (3 points), model (2 points), or physical (1
pied classroom (12 m by 7 m) that contained point) prompts within each instructional trial.
child-sized tables and chairs, toys, and teaching Undesirable behavior was defined as vocal
materials. The four terminal-link interaction attempts to leave the area (e.g., ‘‘Did the timer
areas were located along one wall, three of gooff yet?’’or‘‘Iwantto gobackoverthere.’’),
which contained a table, chair, and teaching and physical or vocal avoidance of instructions
materials. There was a shelf with nine toys that (e.g., pushing materials away or putting his
were often available in the children’s classroom or her head down on the table). All terminal-
(e.g.,woodenblocks,dinosaurs,andconnecting link interactions were videotaped and later
blocks) in the fourth interaction area. A scored in 10-s intervals using handheld com-
different-colored laminated card (60 cm by puters for Charlie and James. Paper-and-pencil
75 cm) was placed on the wall in each of the measures were used to record performance data
interaction areas. On the opposite wall, there in 20-s intervals during sessions for Amy and
was a table containing four smaller colored Anne. A second observer recorded data simul-
cards (15 cm by 10 cm). taneously but independently during at least
28% of all sessions across participants (range,
RESPONSE MEASUREMENTAND 28% to 70%). For Charlie and James, agree-
INTEROBSERVER AGREEMENT ment percentages were calculated by dividing
A second observer simultaneously but in- the smaller number of responses recorded
dependently recorded data on children’s selec- in each interval by the larger number of
tions during 100% of task material and responses and multiplying the result by 100%.
consumable preference assessments. Interob- Interobserver agreement was 95% (session
server agreement was calculated by dividing range, 68% to 100%) for correct responding
the number of agreements by the number of and 99% (session range, 73% to 100%) for
agreements plus disagreements and multiplying undesirable behaviors, respectively. Agreement
by 100%. An agreement was scored when both was scored when both observers recorded the
observers scored the same child selection within same responses in each 20-s interval for Amy
a trial. Agreement was 100% across all task and Anne. Agreement was calculated by di-
materialandconsumablepreferenceassessments viding the number of agreements by the
for all children. number of agreements plus disagreements and
Duringthe interaction area preference assess- multiplying by 100%. Interobserver agreement
ment, observers recorded card selections, which was 99% (session range, 87% to 100%) for
were defined as the first contact of the correct responding and 97% (session range,
participant’s hand on a card. Interobserver 67% to 100%) for undesirable behaviors,
agreement data were collected during at least respectively.
252 NICOLE A. HEAL and GREGORY P. HANLEY
Table 1
Description ofTask Materials inSets 1and2
Skill Set1teachingmaterials Set2teachingmaterials
A Counting:1–20 Multicoloredcountingvehicles(cars,trucks,airplanes, 2cmby2cmnaturalwoodenblocks
andbuses)
B Patterning:Color Multicoloredpegbearswithyellowfoampegboard 5cmby5cmmulticoloredconstruction
paper
C Patterning:Shape Orangeshapeswithfacesonthefrontgluedtoclothes Whiteconstructionpapershapes
hangers
D Patterning:Size RedandgreenMartians(bigandlittle)gluedtofoam 8cmby8cmand15cmby15cmwhite
pieces constructionpaper
E Writing:Lettersa 20cmby28cmlaminatedconstructionpaperwith 20cmby28cmwhiteconstructionpaper
F Writing:Shapes decorativeedgesandweterasemarkers andpencils
G Matching:Numbers Greenfishwithglitteronthefinsgluedtoa2cm 5cmby5cmwhiteconstructionpaper,
H Pointing:Numbers by2cmblueblock,numberslocatedonthebody numberswritteninblackink
I Naming:Numbers ofthefishoryellowstarfishwithglitteronthebow
gluedtoa2cmby2cmyellowblock,numbers
locatedonthebodyofthestarfish
J Pointing:Lettersb 8cmby15cmyellowlaminatedcardswithbluedogs 5cmby5cmwhiteconstructionpaper,
K Ordering:Lettersb onthefrontgluedtofoampieces,letterslocated letterswritteninblackink
onthebodyofthedog
L Matching:Lettersa 8cmby15cmpurplelaminatedcardswithorange 5cmby5cmwhiteconstructionpaper,
M Pointing:Lettersa catsonthefrontgluedtofoampieces,letters letterswritteninblackink
N Naming:Lettersa locatedonthebodyofthecator8cmby10cm
O Ordering:Lettersa orangelaminatedcardswithgreenfrogsonthefront
gluedtowoodenblocks,letterslocatedonthebody
ofthefrog
P Ordering:Name 8cmby15cmbluelaminatedcardswithgreen 5cmby5cmwhiteconstructionpaper,
andredMartiansonthefrontgluedtogreen letterswritteninblackink
Popsiclesticks,letterslocatedonthebodyofthe
Martian
Q Pointing: Bluebutterflieswithglitteronbody,greenwormswith 2cmby10cmwhiteconstructionpaper,
Sightwords glitterontail,yellowsunflowerswithglitterontops sightwordswritteninblackink
gluedtoPopsiclesticks,sightwordslocatedonthe
frontofthepicture
aUppercaseletters.
bLowercaseletters.
PROCEDURE academic task, the two sets of materials were
Task Materials Preference Assessment placed on the table, and the child was asked,
Academic tasks were chosen according to ‘‘Do you want to work on [task name] with
each child’s skill level, as determined from their these materials or these materials?’’ The teacher
individualized classroom curricula. For exam- paused for approximately 2 s and then slid the
ple, if the child had mastered matching child’s chair closer to the table. The child chose
numbers 1 through 20, then the next skill, a set by touching the materials. Following
pointing to numbers 1 through 20, would have a child selection, one academic instruction
been used in the assessment. Two sets of relevant to the teaching materials was delivered.
teaching materials were created for each aca- The order of task presentation was random and
demic task. One set consisted of colorful, counterbalanced, and this sequence of events
iconic, multidimensional materials; the other continued until both sets of teaching materials
consisted of simple black-and-white materials for each academic task were presented once.
(Table 1). A paired-item preference assessment Two sessions were conducted if selections
(Fisheretal.,1992)wasconductedwiththetwo appeared stable; five sessions were conducted
sets of teaching materials to identify the more if different materials were selected across the
and less preferred sets of materials. For each initial two assessments. The most preferred set
PREFERENCES FOR MOTIVATIONAL SYSTEMS 253
of teaching materials was used in the embedded control for James. The color assignments were
context, and the least preferred set of teaching held constant throughout the interaction area
materialswasusedinthesequentialandcontrol preference assessment for each child.
contexts. Interaction area exposure. Exposure sessions
were conducted to allow the child to experience
Consumable Items Preference Assessment
the interaction areas in the terminal links and
Ten consumable items deemed acceptable by learn the associations between colored cards in
parentsandteacherswereselectedforeachchild the initial links and the corresponding terminal
and included in a paired-item preference links. Standing behind the child in the initial-
assessment (Fisher et al., 1992). Each item link area, the teacher vocally and physically
was paired with every other item once and were guided the child to touch one of the four cards
presented to the child on separate plates; and then provided access to the correlated area
approach responses (touching one of the two for 5 min. The selection of each session’s
plates) were scored. After a plate had been academic task was counterbalanced and ran-
touched,thechildwasgivenwiththeitem.The
domly determined for each child, and the same
order of stimulus presentations was random. A
academic task was used across instructional
hierarchy of preference among the items was
contexts within each session. In addition, the
obtained for each child, and the three items the
order of link selections was randomly deter-
child preferred most were used as consequences
mined and counterbalanced.
forcorrectrespondinginthesequentialcontext.
Instructions were delivered approximately
once every 20 to 30 s in each instructional
Interaction Area Preference Assessment
context. A three-step prompting procedure was
Description of interaction areas. Based on the
used to deliver instructions, and consequences
results of the preference assessments, three
(e.g., praise or praise plus consumable item)
instructional contexts were designed. Highly
wereprovidedfollowingcorrectresponsestothe
preferred teaching materials were present and
vocal or model prompt. There were no
praise was delivered following correct respond-
differential consequences for undesirablebehav-
ing in the embedded context, less preferred
ior. To provide similar amounts of teacher
teaching materials were present and highly
attention in all four interaction areas, attention
preferred consumable items plus praise were
was provided every 20 to 30 s in the play area
delivered following correct responding in the
for Amy and Anne. Teacher attention was
sequential context, and less preferred teaching
nearly continuous in the play area for Charlie
materials were present and praise was delivered
following correct responding in the control and James. After the child experienced each
context. Finally, various toys that were typically interaction area once, the exposure session was
available in the preschool classroom were terminated. Four exposure sessions were initial-
present in the play area. The play area was ly conducted with each participant; one addi-
included in the preference assessment to de- tional exposure session was conducted immedi-
termine the conditions under which a child ately prior to the second interaction area
might select a teacher-led instructional activity preference assessment.
over an exclusively child-led play activity. Choice sessions. Choice sessions were identical
Laminated colored cards (60 cm by 75 cm) to the exposure sessions, except that the child
correspondingtoeachconditionwerepostedon made independent selections in the initial link,
the wall of each interaction area. For example, and 3-min access to one of the four interaction
a red card was correlated with play, yellow with areaswasarrangedintheterminallinks.During
sequential,bluewithembedded,andgreenwith thechoicesessions,theteacherstoodbehindthe
254 NICOLE A. HEAL and GREGORY P. HANLEY
child and provided the instruction, ‘‘Touch delivered all instructions in a neutral tone of
a card.’’ If the child did not respond, the same voice, and the quality of the social praise
instruction was repeated twice until 1 min had delivered by the teacher was similar across all
passed. Sessions were conducted once daily and interaction areas. Finally, for each child, the
were terminated after a child made eight same academic task was used across the in-
selections or refused to make a selection within structional contexts within each session.
1 min of a choice opportunity (this happened
once for Charlie and once for James). EXPERIMENTAL DESIGN
Restrictionrules.Tworestrictionrulesallowed Aconcurrent-chainsarrangementwasusedto
the identification of a preference hierarchy evaluate preference for the instructional con-
among the four interaction areas (Hanley, texts and play area. Multielement and reversal
Iwata, Lindberg, & Conners, 2003). First, if designs were used to evaluate relations between
aninteractionareawasselectedon60%ormore the interaction areas and terminal-link behav-
trials for two consecutive sessions or for two of iors (i.e., correct responding and undesirable
three sessions, the card correlated with that area behaviors).
was removed from the choice array. Second, if
two interaction areas were alternately selected RESULTS
for60%ormoretrialsacrossfoursessions,then
Materials Preference Assessment
cards correlated with both of the areas were
The materials preference assessments identi-
removed. After an initial preference hierarchy
fied preferred materials for each participant.
was obtained, a second interaction area prefer-
Charlie showed a relative preference for Set 1
enceassessmentwasconductedtodeterminethe
(colorful, iconic, multidimensional) over Set 2
reliability of the hierarchy with 3 of the 4
(black-and-white) teaching materials for 15 of
children. All four interaction areas were again
the 16 academic tasks. James showed an
available to be selected, an exposure session was
exclusive preference for Set 1 materials. Skills
conducted, and the children were given the
were selected for inclusion in the interaction
opportunity to make independent selections.
area preference assessment based on three
Prevention of bias. The following variables
criteria: (a) There were highly discrepant
were held constant throughout the study to
selection percentages between Sets 1 and 2,
ensure that interaction area selections were
(b) the initial skill in a developmental sequence
a function of the independent variables. The
was selected over a later skill (e.g., pointing to
teacher provided no differential consequences
numbers was selected over naming numbers
beyond access to the different interaction areas
because we generally taught the former prior to
for selections in the initial link. The colored
the latter), and (c) the skill had not been
cards in the initial link were sequenced
mastered in the classroom at the onset of the
randomly at the start of each session. After
interaction area assessment (e.g., patterning was
each selection, the card located at the far left
not used with James because he had previously
end was rotated to the far right end. The acquired this skill in the classroom). Therefore,
position of the interaction areas also was counting, patterning by color, writing shapes,
sequenced randomly for each session. One and pointing to numbers were used in Charlie’s
teacher conducted all sessions throughout the interaction area preference assessment; count-
evaluation. The same teacher also was available ing, writing letters, naming numbers, and
in each interaction area to prevent the partic- naming upper case letters were the four tasks
ipant from choosing an interaction area to gain included in James’ interaction area preference
access to a particular teacher. The teacher assessment.
PREFERENCES FOR MOTIVATIONAL SYSTEMS 255
Only two academic tasks were included in selections to the play area. When the play area
Amy’s and Anne’s materials preference assess- was removed, a preference for the embedded
ments because each had already mastered many context emerged relative to the control context.
of the academic tasks included in Charlie’s and The preference assessment was then repeated
James’ assessments. Amy showed an exclusive with Charlie to determine whether the prefer-
preference for the Set 1 materials, and Anne ence hierarchy would be replicated. Although
selected the Set 1 teaching materials on 80% of results of the second assessment showed that
trials. Naming letters and pointing to sight Charlie initially chose play in lieu of all
words were alternately used in Amy’s and instructional contexts, the hierarchy among
Anne’s interaction area preference assessments. the instructional contexts remained intact in
In sum, the selected Set 1 materials were more that the sequential context was preferred to the
highly preferred than the selected Set 2 embedded context, and both were selected in
materialsforalltasksforallchildren.Therefore, lieu of the control context. Therefore, Charlie
Set 1 materials were included in the embedded preferred the sequential motivational system
instructionalcontexts;Set2materialswereused and, at leastinitially,demonstrated apreference
in the sequential and control contexts. for an instructional context with the sequential
motivational system over an exclusively play-
Consumable Items Preference Assessment
oriented activity.
Theconsumableitemspreferenceassessments
James allocated the majority of his choices to
identified three highly preferred items for each
the play area when all four options were
participant(M&MsH,KitKatBitesH,andGold
available (Figure 1). After the play area was
FishH for Charlie; SkittlesH, Gold FishH, and
restricted, he allocated his selections relatively
jelly beans for James; stickers, Teddy Gra-
evenly between the sequential and embedded
hamsH,andFrootLoopsHforAmy;andKitKat
contexts. In the second assessment, the prefer-
BitesH, Froot LoopsH, and Fruit SmilesH for
ence patterns were replicated (note the similar-
Anne). These items were presented as conse-
ity in the data paths from the first and second
quencesforcorrectrespondinginthesequential
assessments).Insum,Jamespreferredplaytoall
context.
variants of instruction, and he showed a prefer-
ence for the use of motivational systems relative
Interaction Area Preference Assessment
to their absence (i.e., control), but he did not
Initial links (preference data). The results of
show a strong preference for either of the
the interaction area preference assessments are
motivational systems.
shown in Figures 1 and 2. These figures show
Initially, Anne’s selections were variable
cumulative trial-by-trial data. Rows of data
(Figure 2); however, a preference for the play
represent selections towards a single interaction
area eventually emerged. When only the three
area; columns of panels represent data from
instructional contexts were available, Anne
single assessments. Restriction of an interaction
allocated almost all of her selections to the
area is indicated when the data path stops on
sequential context. After the sequential context
each panel. Charlie initially allocated the
majorityofhischoicestothesequentialcontext, was removed from the array, she alternated her
as indicated by the immediate upward trend in selections between the two remaining instruc-
selections for the sequential context and the tional contexts. The patterns in the second
initial flat data paths for the other contexts. assessment were somewhat similar to the first;
Whenthesequentialcontext wasremovedfrom however, Anne often selected the sequential
the array and three interaction areas were context instead of the play area. When the
available, Charlie allocated almost all of his sequentialcontextwasrestricted,Anneallocated
256 NICOLE A. HEAL and GREGORY P. HANLEY
Figure 1. Cumulative number of initial-link selections during the first and second interaction area preference
assessments for Charlie (first andsecondcolumns) andJames (third andfourthcolumns).
all of her selections to the play area. When the Amy because she declined further participation.
play area was restricted, Anne selected the In sum, Amy showed a preference for the play
embedded context. area over all of the instructional contexts.
Amy initially showed variability in her Among the instructional contexts, Amy pre-
selections (Figure 2), but a preference for the ferred the sequential context.
play area emerged. When the play context was ThetoppanelofFigure 3displaysasummary
restricted, she almost exclusively selected the of all children’s interaction area preference
sequentialcontext.Whenthesequentialcontext assessments and shows the percentage of
was restricted, she selected the embedded over selections for each interaction area given the
the control context. The interaction area opportunity to select the interaction area.
preference assessment was not repeated with Overall selection percentages indicate that all
PREFERENCES FOR MOTIVATIONAL SYSTEMS 257
Figure 2. Cumulative number of initial-link selections during the first and second interaction area preference
assessments for Anne (first andsecondcolumns) andAmy (thirdcolumn).
children preferred the play area to instructional lowest in the control context for all children.
contexts, all children preferred the sequential Anne and Amy required less prompting to
contexttotheembeddedcontext,and2ofthe4 complete the academic tasks relative to Charlie
children preferred the use of some motivational and James, which was not surprising consider-
system to none at all. ing these participants’ more extensive histories
Terminal link (performance data). Only very with the preschool classroom curriculum.
small differences were observed in mean task- The percentage of intervals with undesirable
completion scores (Figure 3) across instruction- behaviors is also shown in Figure 3. Charlie
al contexts for all children. Nevertheless, scores engaged in undesirable behaviors during 0%,
were the highest in the sequential context and 6%, 20%, and 23% of the intervals in the play,
258 NICOLE A. HEAL and GREGORY P. HANLEY
at similarly low levels across all interaction
areas; therefore, there was no strong correlation
between undesirable behavior and the value of
the interaction areas.
DISCUSSION
Weadaptedaconcurrent-chainsarrangement
that has been used to identify children’s
preferences for behavioral interventions to de-
termine preschool children’s preferences for
common classroom practices. The 4 children
preferred unstructured play to teacher-led in-
structional contexts that involved different
motivational systems. When the play context
was unavailable, all children preferred the
sequential to the embedded motivational sys-
tem, and 2 of the 4 children preferred the
embedded motivational system to none at all.
This study was primarily concerned with
comparing teaching conditions that are com-
monly found in preschool classrooms. Early
childhood textbooks (Eliason & Jenkins, 1986;
Graves, Gargiulo, Sluder, & Holmes, 1996;
Roopnarine & Johnson, 1993) indicate that
potentially reinforcing teaching materials are
frequently integrated into early childhood
classroom activities, as are enhanced conse-
Figure 3. Summaries of initial-link responding (pref- quences(edibleitems,stickers,trinkets,etc.)for
erence indexes; top), mean task-completion scores (mid-
correct responding. Thus, we selected contexts
dle), and the percentage of intervals with undesirable
with high ecological validity; however, we
behaviors (bottom) for all participants during the in-
teraction areapreference assessments. sacrificed some experimental integrity by vary-
ing two elements across the sequential and
sequential, embedded, and control interaction embedded contexts (i.e., both reinforcer type
areas, respectively. For Charlie, there was and the manner in which the reinforcers were
a negative correlation between the value of the delivered changed across contexts). Thus, by
interaction area (evident in the top panel of evaluating children’s preferences for practices
Figure 3) and undesirable behavior (i.e., the common to preschool classrooms, we may have
likelihood of undesirable behavior decreased as demonstrated the rather obvious outcome that
thepreference valueincreased).Thiscorrelation children prefer candy and stickers to material
also was apparent with Anne and James, who reinforcers(i.e.,thenatureofthedeliveryofthe
engaged in undesirable behaviors in 0%, 0.3%, reinforcement may not have influenced prefer-
7%, and 18% and 0%, 6%, 10%, and 12% of ences). Nevertheless, because all children pre-
the intervals in the play, sequential, embedded, ferred the sequential context, we recommend
and control interaction areas, respectively. By that teachers systematically identify preferred
contrast, Amy engaged in undesirable behaviors items that both teachers and parents find