Embed Size (px)
Citation preview
Behavioral Development Bulletin
School, Meds, and Moms: Using the Standard CelerationChart for a Contextual Analysis of BehaviorTiffany Aninao, Timothy Fuller, Kendra Newsome, and Donny NewsomeOnline First Publication, September 21, 2015. http://dx.doi.org/10.1037/h0101311
CITATIONAninao, T., Fuller, T., Newsome, K., & Newsome, D. (2015, September 21). School, Meds, andMoms: Using the Standard Celeration Chart for a Contextual Analysis of Behavior. BehavioralDevelopment Bulletin. Advance online publication. http://dx.doi.org/10.1037/h0101311
School, Meds, and Moms: Using the Standard Celeration Chart for aContextual Analysis of Behavior
Tiffany Aninao, Timothy Fuller, Kendra Newsome, and Donny NewsomeFit Learning, Reno, Nevada
Contextual factors can have a significant impact on treatment outcomes. However,systematic analysis is difficult in the absence of appropriate measurement tools. TheStandard Celeration Chart provides a way for evaluating the effects of these variablesthrough its standardization and availability of immediate data analysis and decisionmaking. Standard Celeration Charts are presented demonstrating how the date syn-chronization feature of the Standard Celeration Chart is used to identify and assess theinfluence of contextual variables. Specifically, using the Standard Celeration Chart canmake contextual analysis possible by illustrating how school enrollment status, med-ication changes, and inconsistent session attendance can be observed as factors criticalto academic task performance.
Keywords: date synchronization, contextual analysis, confounding variables, Standard CelerationChart
It is commonly accepted that there are vari-ables influencing behavior outside the behaviorscientist’s purview (Cooper, Heron, & Heward,2007; Fryling & Hayes, 2009; Sidman, 1960).These contextual factors can have significantimpact on treatment outcomes in a clinical set-ting. Examples of factors that can affect treat-ment outcomes include, but are not limited to,variability in session attendance, tardiness, hol-idays that reduce session time, academic curric-ulum, behavior management strategies imple-mented outside the treatment environment, andmedication changes. It is impossible to controlfor all possible contextual features of this sort ina clinical setting, and systemic analysis of suchfactors is difficult in the absence of sensitivemeasurement tools. The Standard CelerationChart (SCC) provides a lens for evaluating theeffects of these factors through its standardiza-tion (Kubina & Yurich, 2012). The sensitivityof the SCC’s date-synchronization feature pro-vides the benefit of analyzing data in calendar
time (see Kubina & Yurich, 2012, pp. 159–160). In this article, we discuss how the chartcan be used as an analytic tool for a contextualanalysis. Data plotted on three Standard Celera-tion Charts and three corresponding conven-tional Excel graphs are presented to demon-strate the superior clinical utility of the SCC fordetecting sources of variance in performance inreal calendar time. In these demonstrations, theSCC was used to assess the influence of vari-ables, including changes in schedule due tobreaks in school, medication changes, and at-tendance. A discussion of how to use the SCCas a communication tool for interested partiesand how it relates to the breadth of interventionstrategies is provided.
Method
In the first demonstration, changes in behav-ior were detected due to a learner not attendingschool. Frank was a 7-year-old boy diagnosedwith attention-deficit hyperactivity disorder(ADHD). Frank was referred for services at FitLearning for reading remediation. Due to a his-tory of noncompliance, data were recorded andcharted for duration of work refusal for eachsession from the beginning of his enrollment. Inan effort to account for variability in these data,phase change lines were added to his chart
Tiffany Aninao, Timothy Fuller, Kendra Newsome, andDonny Newsome, Fit Learning, Reno, Nevada.
Correspondence concerning this article should be ad-dressed to Tiffany Aninao, Fit Learning, 9408 Double RBoulevard,SuiteB,Reno,NV89521.E-mail: [email protected]
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
Behavioral Development Bulletin © 2015 American Psychological Association2015, Vol. 20, No. 2, 000 1942-0722/15/$12.00 http://dx.doi.org/10.1037/h0101311
1
retroactively to indicate when school sessionsbegan and ended.
In our second demonstration, the impact ofmedication on mathematics acquisition pro-gramming and a global progress-monitoringprobe in the form of a curriculum-based mea-surement (CBM) was assessed. Cathy was an11-year-old girl diagnosed with ADHD. Cathywas attending Fit Learning for mathematics re-mediation. Cathy’s mother reported that Cathyhad started a new medication on a particulardate. This medication change was indicated onall of Cathy’s skill acquisition, behavior reduc-tion, and skill probe charts to analyze the po-tential effects of this medication across pro-gramming.
Jack was a 7-year-old boy diagnosed withdyslexia and the subject of our third demonstra-tion. Jack was referred for services at Fit Learn-ing for reading remediation. He had both on-site(i.e., conducted at the center) and off-site (i.e.,conducted at his school) sessions. Weekly read-ing CBM measures were collected only duringon-site sessions. Jack often did not attend on-site sessions. Patterns of attendance and nonat-tendance of on-site sessions were evaluated onJack’s weekly reading CBM measure.
Results
Contextual factors were recorded and ana-lyzed using the SCC. Frank’s data are displayedin Figure 1. The record floors indicate the totalduration of the session (100 min). The dotsrepresent total duration of work refusal minutes.Dots below the record floor indicate days whereno duration of work refusal was recorded.While reviewing the data, the clinicians ob-served a 2-week period of decreased work re-fusal followed by an increase in work refusalacross several weeks. Upon observing this pat-tern, periods of no school attendance (e.g., hol-iday break) and periods of school attendancewere indicated on the work refusal chart. Aperiod where Frank was not attending school isindicated from the dates December 29, 2013,through January 12, 2014, by dashed phasechange lines on this chart.
Using the SCC, clinicians were able to gleanthat work refusal was more likely while Frankwas attending school compared to when he wasnot attending school. These data are summa-rized in Table 1, where we can see that Frank
engaged in an average of 7.75 min of workrefusal per instructional hour while school wasin session. During periods where Frank was notattending school, an average of 1.25 min ofwork refusal per instructional hour was ob-served. These observations indicate a correla-tion between school attendance and work re-fusal.
Cathy’s data are displayed in Figure 2. Ef-fects of medication changes were analyzed inthe context of her performance on a fifth-grademath computation CBM. The dots representcorrect digits written. The “x” represents incor-rect digits written. Upon starting a medicationregimen, a noticeable reduction in the variabil-ity of her math computation CBM performancewas afforded by plotting these data on the SCC.Bounce envelopes were calculated using anelectronic template for the SCC (Harder, White,& Born, 2004). Before medication changes,variability in performance was at �3.25. AfterCathy started ADHD medication, the variabilitywas reduced to �1.59. A statistical analysisshows that before medication, Cathy was per-forming at an average of 13.8 digits written perminute with a standard deviation of 5.58. Aftermedication, Cathy’s performance increased toan average of 24.45 digits written per minutewith standard deviation of 3.81.
Jack’s data are displayed in Figure 3. Effectsof on-site session attendance were analyzed incontext of his performance on a first-grade-levelreading CBM. The dots represent words readcorrectly. The “x” represents words read incor-rectly. Dots and the “x” symbol that are on thesame vertical line are due to a technique re-ferred to as “stacked dots,” where multiple1-min timings were conducted on the readingpassage, and all of these timings are charted onthe corresponding weekly line. Because data onthe SCC are plotted in real calendar time, pat-terns of attendance were evidenced by thelength of horizontal spaces between data pointson the weekly per minute chart. Celerationsacross months of consistent on-site session at-tendance versus months of inconsistent atten-dance or nonattendance of on-site sessions wereconsiderably different. Celeration lines onJack’s chart were calculated using the QuarterIntersect Method (Pennypacker, Gutierrez, &Lindsley, 2003). Periods of consistent on-sitesession attendance can be detected in themonths of July and October. Jack’s data show
2 ANINAO, FULLER, NEWSOME, AND NEWSOME
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
Figure 1. Frank’s duration of work refusal during a 100-min session.
3USING THE SCC FOR A CONTEXTUAL ANALYSIS OF BEHAVIOR
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
a �1.33 celeration in July and a �1.31 celera-tion in October. Periods of inconsistent on-sitesession attendance include August through Sep-tember and November through March. Jack’sdata show a �1.25 celeration in August throughSeptember and a �1.06 celeration in Novemberthrough March. Table 2 presents an analysis ofJack’s percentage of on-site sessions attended.The steepest celerations in CBM performancecorrespond to months of high percentage ofon-site sessions attended (e.g., July and Octo-ber).
Discussion
The SCC provides behavior scientists with avaluable tool for evaluating the effects of bothprogrammed and unprogrammed contextualfactors. Specifically, the SCC’s sensitivity anddate synchronization feature have proved to beof particular value to assess the influence ofcontextual variables. By indicating variablessuch as the nonattendance of school, medicationchanges, or variability in attendance on theSCC, we are able to evaluate the effects of theseunprogrammed events on behavior.
The data presented here are evidence of ben-efits of the SCC over other data managementconventions such as graphing by session ratherthan calendar date. Use of another form of datadisplay may have masked these sources of vari-ability from the clinician. The use of nonstan-dardized data displays may make it difficult forthe clinician to accurately evaluate rate ofchange, potentially exaggerating or muting theappearance of effects, and may negatively affectdecision making (Kubina, 2014). To illustratethis point, we have included comparisons of thesame data graphed using a more traditional Ex-cel-based data display for each leaner (Figures4, 5, and 6).
Frank’s data are displayed on an Excel linegraph in Figure 4 with sessions across the x-axis
and percentage of time in work refusal acrossthe y-axis. A period of decreased work refusalcan be detected (between Sessions 14 and 28).However, the clinician would not have beenable to put in phase change lines to indicatewhen Frank was in or out of school because dataare plotted by session and do not correspond tospecific dates. Additionally, it is important tonote that Sessions 14–28 are not the sessionsthat correspond to when Frank was out ofschool. If the clinician were to only use the linegraph in Figure 4, the functional relation be-tween school attendance and work refusalwould have gone undetected.
Cathy’s data are shown using an alternativedata display in Figure 5 with CBM weeks acrossthe x-axis and percent correct across the y-axis.A phase change line could be indicated on thisdisplay because the parents provided the medi-cation information the day that it went intoeffect. Detection of the same decrease in vari-ability is evident in this display, as seen on theSCC it had been originally plotted on. At firstglance, this may appear to not support ourargument that the SCC provides a unique ana-lytical vantage for clinicians. However, the pur-pose of this demonstration is not that Excel-created line graphs are inappropriate in allinstances. Rather, it is to show the SCC iseffective in providing clinicians with preciseanalytical vantage in all instances. Furthermore,Cathy’s percent correct performance depicted inFigure 5 is not without limitation. For instance,although decreases in variability are detectedand this detection can be linked to the introduc-tion of a medication regimen, these changes arein different units of measure from those plottedon the SCC. More specifically, in Figure 5, they-axis is percent correct, whereas the y-axis ona SCC is correct responses per minute. In thecase of Cathy’s performance, percent correcthas improved, but what is not available foranalysis on this conventional display is what
Table 1Display of Frank’s Instructional Hours and Work Refusal Minutes SeparatedInto Periods of School Attendance and Nonschool Attendance
Analysis period (November 10,2013–March 23, 2014)
Total instructionalhours
Total refusalminutes
Refusal minutes perhour of instruction
In school 57 441.75 7.75Out of school 8 10 1.25
4 ANINAO, FULLER, NEWSOME, AND NEWSOME
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
Figure 2. Frequency of correct and incorrect digits written on Cathy’s math computationcurriculum based measurement. ADHD � attention-deficit hyperactivity disorder.
5USING THE SCC FOR A CONTEXTUAL ANALYSIS OF BEHAVIOR
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
Figure 3. Frequency of correct and incorrect words read on Jack’s reading curriculum-basedmeasurement.
6 ANINAO, FULLER, NEWSOME, AND NEWSOME
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
constitutes this statistical representation of be-havior. For example, it might have been the casethat her percent correct had improved, but thenumber of correct digits written decreased afterthe introduction of medication. This informa-tion could be determined using non-SCC con-ventions but would require additional displays.The SCC provides clinicians the analytical pre-cision to evaluate multiple aspects of a learner’sperformance on one display without the need tocreate multiple graphs.
Jack’s alternative data display (see Figure 6)is constructed with CBM weeks across the x-axis and percent correct across the y-axis. Byexcluding the date synchronization feature, thisdisplay method masks the variability in atten-dance of on-site sessions. Similarly to Cathy’sgraph analysis, by excluding the frequency
count feature of the SCC, this alternative dis-play also masks the comparison between prog-ress during periods of on-site session attendanceand progress during periods of inconsistent at-tendance that was clearly visible when datawere displayed on the SCC. Again, these twocomparisons display different units of measure(i.e., the y-axis on the line graph is percentcorrect, whereas the y-axis on a SCC is count).
The impact of using insensitive data displayspotentially leads clinicians and researchers aliketo false conclusions about the contributing fac-tors. For example, given the same data withoutany reference to calendar time, one might haveattributed the observed variability to a variety ofputative explanatory factors (e.g., inadequatemotivational operations, poorly selected or pin-pointed curriculum, or poor program fidelity byinstructors). Interventions based on these pre-sumed factors would be unlikely to yield de-sired results because the underlying contribut-ing factors would not be adequately addressed.
Discovery of contextual influences on perfor-mance from outside the treatment environment,such as those described above, is valuable notonly to the clinician but also to other stakehold-ers. The standardization of the SCC makes it aparticularly useful communication tool to beused with interested parties (e.g., parents andphysicians). Even if effects of changes could bedetected using other data display tools, the abil-
Table 2Display of Jack’s On-Site Session Attendance
Month Percentage of on-site sessions attended
July 88August 25September 50October 100November 50December 33January 0February 50March 50
Figure 4. Frank’s percentage of time in work refusal during each session.
7USING THE SCC FOR A CONTEXTUAL ANALYSIS OF BEHAVIOR
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
ity to analyze these changes is limited by thereorientation required with each new display.Laypersons should not be expected to have thesame data consumption skills as scientists. TheSCC reduces the number of component skillsrequired of the consumer by having a standard-ized display. A benefit of using the SCC as acommunication tool is that charting conventionshave be learned only one time, eliminating the
need to relearn the dimensions of each chart(Kubina & Yurich, 2012). The use of the SCCas a data display tool makes communicationmore efficient due to its standardization. WithFrank’s SCC, behavioral differences noted dur-ing periods of no school attendance can becommunicated to his parents and teachers. Withdata displayed on Cathy’s SCC, the beneficialimpact of the medication modification on her
Figure 5. Cathy’s percentage of correct digits written on a math computation curriculum-based measurement (CBM) before and after introduction of attention-deficit hyperactivitydisorder medication.
Figure 6. Jack’s percentage of correct words read on a reading curriculum-based measure-ment (CBM).
8 ANINAO, FULLER, NEWSOME, AND NEWSOME
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
grade-level math skills can be communicated toher caregivers and physician. Last, the impor-tance of consistent attendance and the impactthis variable has on grade-level progress can bemade clear to Jack’s parents.
References
Cooper, J. O., Heron, T. E., & Heward, W. L. (2007).Applied behavior analysis. Upper Saddle River,NJ: Pearson Education.
Fryling, M. J., & Hayes, L. J. (2009). Psychologicalevents and constructs: An alliance with Smith. ThePsychological Record, 59, 133–142.
Harder, S., White, O., & Born, S. (2004). Exceltemplate for the Standard Celeration Chart. Re-trieved from http://harderchartingtemplates.pbworks.com/
Kubina, R. (2014, December). A critical review oftime-series graphics in behavior analytic journals.PowerPoint presentation at the meeting of the In-ternational Precision Teaching Conference, Chi-cago, IL.
Kubina, R., & Yurich, K. K. L. (2012). The precisionteaching book. Lemont, PA: Greatness AchievedPublishing Company.
Pennypacker, H. S., Gutierrez, A., & Lindsley, O. R.(2003). Handbook of the Standard CelerationChart. Cambridge, MA: Cambridge Center for Be-havioral Studies.
Sidman, M. (1960). Tactics of scientific research:Evaluating experimental data in psychology. Bos-ton, MA: Authors Cooperative, Inc.
Received December 19, 2014Revision received May 27, 2015
Accepted May 28, 2015 �
9USING THE SCC FOR A CONTEXTUAL ANALYSIS OF BEHAVIOR
Thi
sdo
cum
ent
isco
pyri
ghte
dby
the
Am
eric
anPs
ycho
logi
cal
Ass
ocia
tion
oron
eof
itsal
lied
publ
ishe
rs.
Thi
sar
ticle
isin
tend
edso
lely
for
the
pers
onal
use
ofth
ein
divi
dual
user
and
isno
tto
bedi
ssem
inat
edbr
oadl
y.
