ERIC EJ853178: The Effects of Varied versus Constant High-, Medium-, and Low-Preference Stimuli on Performance

2009·0.13 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2009, 42, 321–326 NUMBER2 (SUMMER2009)
THE EFFECTS OF VARIED VERSUS CONSTANT HIGH-, MEDIUM-,
AND LOW-PREFERENCE STIMULI ON PERFORMANCE
BYRON WINE AND DAVID A. WILDER
FLORIDAINSTITUTEOFTECHNOLOGY
The purpose of the current study was to compare the delivery of varied versus constant high-,
medium-,andlow-preferencestimulionperformanceof2adultsonacomputer-basedtaskinan
analogue employment setting. For both participants, constant delivery of the high-preference
stimulus produced the greatest increases in performance over baseline; the varied presentation
produced performance comparable to constant delivery of medium-preference stimuli. Results
are discussed in terms of their implications for the selection and delivery of stimuli as part of
employee performance-improvement programs in the field of organizational behavior
management.
DESCRIPTORS: organizational behavior management, reinforcer assessment, stimulus
variation
_______________________________________________________________________________
The selection and delivery of effective versus higher quality constant stimuli among 7
reinforcers are important topics for the field of childrenwithdevelopmentaldisabilities.Results
applied behavior analysis in general and orga- indicated that 5 participants preferred varied
nizational behavior management (OBM) in presentation, and 2 participants preferred
particular.Onemethodofenhancingtheeffects constant presentation.
of reinforcers on employee performance might Research on the effects of stimulus variation
be to vary the stimuli delivered during on performance by employees in organizational
performance-improvement programs. Although settings is needed. If varied presentation does
no research on varied versus constant reinforcer enhance employee performance, its use may
presentation exists in the OBM literature, lead to the development of more cost-effective
previous research on this topic has been and efficient performance-improvement pro-
conducted with children in educational and grams in organizations. The purpose of this
clinical settings. study was to compare the delivery of varied
Egel (1980) demonstrated that varied, as versus constant high-, medium-, and low-
opposed to constant, presentation of stimuli preference stimuli on the performance of 2
produced more and faster bar pressing by adults on a computer-based task in an analogue
children with autism. This finding was repli- employment setting.
cated by Egel (1981) with 3 children with
autism using curriculum-based responses in a
METHOD
naturalistic context. Similarly, Bowman, Piazza,
Participants and Setting
Fisher, Hagopian, and Kogan (1997) examined
Participants were recruited via informational
preference for lower quality varied stimuli
flyers distributed on a university campus and
This report is based on a thesis submitted by the first weretoldthatthepurposeoftheresearchwasto
author in partial fulfillment of the requirements for the
evaluate a computer software program that may
masterofsciencedegreeinappliedbehavioranalysisatthe
assist with the completion of paperwork. Two
FloridaInstituteof Technology.
Address correspondence to David A. Wilder, Florida individuals participated. Sara was a 22-year-old
Institute of Technology, School of Psychology, 150 W. female undergraduate student at the university.
University Blvd., Melbourne, Florida 32901 (e-mail:
Jane was a 34-year-old woman who was
[email protected]).
doi:10.1901/jaba.2009.42-321 employed as an instructor at the same univer-
321
322 BYRON WINE and DAVID A. WILDER
sity. Sessions took place in a laboratory room pack.Jane’shigh-preferenceitemwasthemovie
designed to look like an office (i.e., a desk, theater GC, her medium-preference item was
office chair, computer, and office supplies were the candy, and her low-preference item was the
present). pencil pack.
Comparison of varied versus constant presenta-
Data Collection
tion. A combination reversal (ABA, where A is
Each participant performed a task using a baselineandBiscomparison)andmultielement
check-writing software program run on a design was used to evaluate the effects of varied
laptop computer. Specifically, experimenters versus constant stimuli on performance. Each
instructed participants to type in the correct participant was first exposed to a baseline
monetary amount indicated by the program on
condition during which no programmed con-
computer-generated blank checks. The depen-
sequences forcompletingcheckswereprovided.
dent variable was number of checks completed
Participants were told, ‘‘You can do as many as
per session.
you want, as few as you want, or none at all.
Whenyouarefinished,pleasetellme.’’Notime
Procedure
limitwasplacedonsessions.Afterdeliveringthe
Stimulus preference assessment. Experimenters
instructions,theexperimenterlefttheroomand
used a reinforcer survey (Wilder, Therrien, &
waited in an adjoining room. Baseline contin-
Wine, 2005) to determine participants’ prefer-
ued until the participant declined the opportu-
ences for six stimuli. All six stimuli had about
nity to complete checks for three consecutive
the same monetary value (i.e., $5). The stimuli
sessions. Each participant’s mean number of
evaluated were a $5 gift certificate (GC) to a
checks completed during baseline sessions,
movietheater,$5GCtoageneralstore,$5GC
excluding the last three sessions, was then
to a health food shop, a pack of specialty
calculated, multiplied by two, and set as the
pencils, a box of randomly selected specialty
response requirement to earn one stimulus in
candy, and some stationery. Participants rated
the comparison phase.
eachstimulusona4-pointLikertscale(15low
Following baseline, each participant was
preference, 2 and 3 5 medium preference, and 4
exposed to one of five conditions (high
5 high preference).
preference, medium preference, low preference,
BothparticipantsratedthemovietheaterGC
varied, and control) on a quasirandom basis.
as high preference. Sara rated three items
The specified stimulus (high, medium, or low
(general store GC, health food shop GC, and
preference) was delivered in its analogous
stationery)asmediumpreferenceandtwoitems
condition(highpreference,mediumpreference,
(pencils and candy) as low preference. Jane
or low preference) on the fixed-ratio (FR)
rated three items (general store GC, candy, and
schedule determined following baseline. Each
stationery)asmediumpreferenceandtwoitems
participant received access to a randomly
(pencils and health food shop GC) as low
preference. Medium- and low-preference stim- determined stimulus (either high, medium, or
uli used in the comparison of varied versus low preference) on their individually deter-
constant stimuli were chosen randomly from minedFRscheduleduringthevariedcondition.
the items each participant rated as medium and Eachtypeofstimulushadanequallikelihoodof
low preference during the preference assess- being selected, and each participant was
ment. In this phase, Sara’s high-preference item unaware of which item she would receive. For
was the movie theater GC, her medium- the varied preference condition, Sara received
preference item was the health food shop GC, access to her high-preference, low-preference,
and her low-preference item was the pencil and medium-preference items in Sessions 19,
VARIED VERSUS CONSTANT STIMULI 323
21, and 24, respectively. Jane received access to were delivered at the beginning of the next
her high-preference, medium-preference, and scheduled blockof sessions, which was within 3
low-preference items in Sessions 10, 18, and days.
20, respectively. Participants received access to Interobserver agreement and integrity of the
one white card (8 cm by 14 cm) on their independent variable. Interobserver agreement
individual FR schedules during the control data on the number of checks completed were
condition. collected on at least 50% of sessions for each
Each of the conditions was presented three participant. An agreement was defined as both
times, for a total of 15 sessions. Experimenters observers scoring the same number of checks
gave participants the same instructions as in completed; a disagreement occurred when
baseline and to stop and say ‘‘I’m done’’ when observers scored a different numbers of checks
they finished. No time limit was placed on completed. Interobserver agreement was calcu-
sessions, although all sessions were less than lated by dividing the number of agreements by
20 min in duration, and no limits were placed the number of agreements plus disagreements
on the number of stimuli earned per session. and converting this ratio to a percentage. For
both participants, mean agreement was 100%.
For the varied condition, the earned stimulus
Integrity of the independent variable was
was held constant during a specific session (i.e.,
assessed by having a second observer verify that
if the stimulus earned was the low-preference
the items participants earned during sessions
stimulus, only the low-preference stimulus was
were subsequently delivered. Mean integrity,
earned each time the criterion was met during
which was conducted on at least 50% of
that session). At the beginning of each session,
sessions for both participants, was 100%.
the experimenter told and showed the partici-
pant which item she would receive (except in
the varied condition, during which the exper- RESULTS AND DISCUSSION
imenter told the participant that she would get
Figure 1 displays Sara’s and Jane’s results.
either her high-, medium-, or low-preference
Saracompletedameanof16checks(range,1to
item) and the number of checks needed to
26)duringthefirsteightsessionsofbaseline(no
receive one item. Participants received no
responding in the final three sessions). There-
formal feedback on the number of checks
fore, her response requirement was set at 32
completed until after they finished a session.
checks for the comparison phase. Sara complet-
However, they were free to monitor their own
ed no checks during the low-preference and the
performance.Oncefinishedwithasession,they
controlconditions.Shecompletedameanof82
received a 2-min break. After the break, the checks (range, 68 to 95) during the medium-
experimenter randomly selected another one of preference condition, a mean of 143 checks
the five conditions, and the procedures were (range, 77 to 198) during the high-preference
repeated. No more than two consecutive condition, and a mean of 70 checks (range, 58
presentations of the same condition were to 89) during the varied condition. She did not
permitted. Participants completed two to five complete any checks during the return to
sessions per visit. The experimenter continued baseline.
to conduct one of the five conditions randomly Jane completed a mean of 61 checks (range,
until each condition had been conducted three 35 to 77) during the first three sessions of
times. At the end of each block of sessions, the baseline (no responding in the final three
experimenter wrote the names and quantity of sessions). Therefore, her response requirement
the items earned on a sheet of paper for the was set at 122 checks for thecomparison phase.
participant to take with her. The items earned She completed no checks during the control
324 BYRON WINE and DAVID A. WILDER
Figure1. Numberofcheckscompletedacrossbaseline,constanthigh-preference(HP),constantmedium-preference
(MP),constantlow-preference(LP),variedpreference,andcontrolconditionsforSara(top)andJane(bottom).Forthe
variedpreferencecondition,Sarareceivedaccesstoherhigh-,low-,andmedium-preferenceitemsinSessions19,21,and
24, respectively. Jane received access to her high-, medium-, and low-preference items in Sessions 10, 18, and
20,respectively.
VARIED VERSUS CONSTANT STIMULI 325
conditionandfivechecksduringtheinitiallow- stimuli. Similarly, more recent research (Fran-
preference condition. No subsequent respond- cisco, Borrero, & Sy, 2008) has suggested that
ing was observed during the low-preference lower preference stimuli can function as
condition. She completed a mean of 99 checks reinforcers. The results of the current study
(range, 59 to 147) during the medium- add to those of Bowman et al. by suggesting
preference condition. During the high-prefer- that the delivery of varied stimuli produces
ence condition, she completed a mean of 213 improvements in performance comparable to
checks(range,184to231),andshecompleteda those produced by constant delivery of medi-
mean of 135 checks (range, 106 to 151) during um-preference stimuli.
the varied condition. She did not complete any These results have implications for the
checks during the return to baseline. selection and delivery of items as part of
It is interesting to note that on a few employee performance-improvement programs
occasions, responding was just short of a in the field of OBM. Specifically, this study
criterion (e.g., during Session 19, Sara com- suggests that the delivery of varied, randomly
pleted 58 checks, 8 shy of criterion). Such selected stimuli as part of performance-im-
findings were likely due to a miscalculation on provement programs may result in improved
the part of the participant, because both performance.Inaddition,thispracticemightbe
participants reported that they attempted to advantageous for other reasons. For example, if
tally their own responses. a number of employees in an organization are
Unlike previous research that has compared on performance-improvement programs, the
varied and constant reinforcer presentation delivery of each individual’s most preferred
(e.g., Egel, 1980, 1981), the results of the item contingent on improved performance
currentstudysuggestthatvariedpresentationof might be difficult to arrange on a consistent
items produced performance above baseline basis. On the other hand, delivery of stimuli
levels but lower than conditions in which selectedatrandomfromamongasmallnumber
constant access to a high-preference item was of items most preferred by some, but not all,
providedcontingent on performance.However, employees would likely be easier to coordinate.
inthetwostudiesbyEgel,itisnotknownifthe The results of this study suggest a number of
item delivered in the constant condition was topics for future research. First, price arrange-
highly preferred; a stimulus preference assess- mentofthevariousitemscouldbeexaminedby
ment was not conducted. Other differences increasing or decreasing the price (i.e., criteria
between the current study and the Egel studies for check completion) of items. Second, future
that could account for the discrepant findings researchcouldexaminetheeffectsofreceiptofa
were the items delivered and timing of delivery. high- or low-preference item immediately
Egel used edible items delivered immediately before receipt of an item of a different value.
after responding. In the current study, some of Participants may respond less for medium- or
the items were secondary reinforcers delivered low-preference items after they have just earned
after the performance. access to a high-preference item.
The current study adds to the literature by Becausethedependentvariableinthecurrent
demonstrating that the effects of stimulus study was number of checks completed as
variation relative to constant presentation may opposed to rate of check completion, these
depend on preference for the items being results should be viewed as preliminary. A
delivered. Bowman et al. (1997) demonstrated measure of response rate would provide infor-
that some individuals preferred lower quality mation about response efficiency, which might
varied stimuli over higher quality constant be particularly important in organizations. A
326 BYRON WINE and DAVID A. WILDER
second limitation is the restricted set of stimuli Egel, A. L. (1980). The effects of constant versus varied
reinforcer presentation on responding by autistic
used in the study. Future research on this topic
children. Journal of Experimental Child Psychology,
should employ a wider variety of stimuli, and
30,455–463.
perhaps some stimuli that are exchangeable for Egel, A. L. (1981). Reinforcer variation: Implications for
goodsandservicesatmultiplesites(e.g.,aVISA motivatingdevelopmentallydisabledchildren.Journal
ofAppliedBehavior Analysis, 14, 345–350.
check card). Finally, the accuracy of check
Francisco, M. T., Borrero, J. C., & Sy, J. R. (2008).
completion was not examined. Future research Evaluation of absolute and relative reinforcer value
using a task such as the one employed in this using progressive-ratio schedules. Journal of Applied
study should examine the integrity of task BehaviorAnalysis,41,189–202.
Wilder, D. A., Therrien, K., & Wine, B. (2005). A
completion.
comparison between survey and verbal choice meth-
ods of identifying potential reinforcers among
employees. Journal of Organizational Behavior Man-
REFERENCES
agement,25(4),1–13.
Bowman,L.G.,Piazza,C.C.,Fisher,W.W.,Hagopian,
L.P.,&Kogan,J.S.(1997).Assessmentofpreference Received May 31,2007
for varied versus constant reinforcers. Journal of Final acceptance October8,2007
AppliedBehavior Analysis, 30,451–458. Action Editor,Joel Ringdahl