Reliability and Validity of The Functional Analysis Screening Tool
Reliability and Validity of The Functional Analysis Screening Tool
Reliability and Validity of The Functional Analysis Screening Tool
net/publication/257647701
CITATIONS READS
86 5,346
3 authors:
Eileen M Roscoe
New England Center for Children
36 PUBLICATIONS 1,270 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
A parametric analysis of the effects of physical activity on stereotypy and on-task behavior. View project
On the relation between treatment efficacy and patient preference for behavioral interventions View project
All content following this page was uploaded by Iser G DeLeon on 16 October 2017.
ISER G. DELEON
KENNEDY KRIEGER INSTITUTE AND JOHNS HOPKINS UNIVERSITY SCHOOL OF MEDICINE
AND
EILEEN M. ROSCOE
NEW ENGLAND CENTER FOR CHILDREN
The Functional Analysis Screening Tool (FAST) is a 16-item questionnaire about antecedent and
consequent events that might be correlated with the occurrence of problem behavior. Items are
organized into 4 functional categories based on contingencies that maintain problem behavior. We
assessed interrater reliability of the FAST with 196 problem behaviors through independent
administration to pairs of raters (Study 1). Mean item-by-item agreement between pairs of raters was
71.5%. Agreement for individual items ranged from 53.3% to 84.5%. Agreement on FAST
outcomes, based on comparison of informants’ highest totals, was 64.8%. We assessed the validity of
the FAST by comparing its outcomes with results of 69 functional analyses (Study 2). The FAST
score predicted the condition of the functional analysis in which the highest rate of problem
behavior occurred in 44 cases (63.8%). Potential uses of the FAST in the context of a clinical
interview, as well as limitations, are discussed.
Key words: assessment, functional analysis, rating scale, problem behavior
Functional analysis (FA) methodology in- have been suggested for conducting interviews
volves observation of behavior under a series of with caretakers (Groden, 1989; Iwata, Wong,
test and control conditions and is regarded as Riordan, Dorsey, & Lau; 1982; O’Neill,
the benchmark standard for assessment of Horner, Albin, Storey, & Sprague, 1990), which
problem behavior in both clinical research and included questions about environmental
practice (Hanley, Iwata, & McCord, 2003). circumstances that might be correlated with
Prior to conducting an FA, therapists often the occurrence of problem behavior. Verbal
gather information about the client and problem reports about behavior, however, often are
behavior by interviewing significant others, unreliable and inaccurate. For example, it has
which may be helpful in designing some aspects been found that caregiver descriptions of
of FA conditions. For example, several formats client preferences often do not correspond
with empirical (direct) assessment of those
We thank SungWoo Kahng, Dorothea Lerman, Jana preferences (Green, Reid, Canipe, & Gardner,
Lindberg, Bridget Shore, Michele Wallace, and April 1991; Green et al., 1988). Given the inability
Worsdell, who helped with initial stages of scale develop- of caregivers to identify stimuli that might
ment; Nancy Sanders, who coordinated pilot testing; and
Kathryn Horton, Kathryn Jann, Griffin Rooker, and Kevin serve as reinforcers for any behavior, identifica-
Schlichenmeyer, who assisted with data analysis. tion of reinforcers that maintain specific
Address correspondence to Brian A. Iwata, 114 Psycho- behaviors would seem to be an even more
logy Building, University of Florida, Gainesville, Florida
32611 (e-mail: [email protected]). formidable challenge beyond the scope of an
doi: 10.1002/jaba.31 interview.
271
272 BRIAN A. IWATA et al.
access to attention and access to tangible items noncompliance). The initial scale consisted of
under a single category (social-positive reinforce- 32 questions, with eight questions for each of the
ment) instead of separating them, as is done in four functions.
some scales such as the Motivation Assessment We subsequently modified the scale based on
Scale (MAS; Durand & Crimmins, 1988) and results of a series of evaluations. First, all members
the Questions about Behavioral Function (N ¼ 9) of a psychology department at a
(QABF, Matson & Vollmer, 1995). In a similar residential center for persons with intellectual
way, we combined escape from task demands disabilities used the initial scale during their
with other types of social escape and avoidance assessments of problem behavior. After a 4-
under the social-negative reinforcement category month trial period, they provided detailed
(neither the MAS nor the QABF makes a written feedback about scale content and format,
distinction between these forms of negative which we used to make format revisions and to
reinforcement). Finally, the automatic-negative modify wording of specific items. Second,
reinforcement category is a tenuous one. Al- reliability analyses were conducted in three state
though alleviation of discomfort has a high degree residential programs for persons with intellectual
of face validity because it seems to be a reinforcer disabilities. The scale was administered to pairs of
for some behavior (e.g., scratching an itch), its direct-care staff who worked closely with
direct influence on problem behavior has been individuals who engaged in varied problem
largely hypothetical, based on either inferences behaviors. Following the first administration,
from nonhuman work or correlational data reliability (percentage agreement) scores were
(Cataldo & Harris, 1982) rather than experi- calculated separately for each of the 32 items and
mental data with clinical populations. For were used as the basis for revision. Items with the
example, although problem behavior may appear lowest reliabilities were reworded or deleted, and
to be exacerbated by illness, it could be a revised scale was administered to another
maintained purely by social consequences that sample of informants. We revised the scale four
become more valuable in the presence of times in this manner based on data collected for a
discomfort. O’Reilly (1997) presented data that total of 182 individuals, yielding a final scale that
showed that an individual’s SIB during attention consists of 16 items.
and demand conditions of an FA was correlated
with the presence of otitis media, which Description of the Scale
suggested that attention and escape from task Figure 1 shows the current version of the
demands were reinforcing when illness was FAST, which consists of three sections. (See
present but not when it was absent. Nevertheless, Supporting Information for a full-page version.)
inclusion of this function complements the other The first section contains brief instructions;
three in spite of a general absence of research on it also is used to record information about
its characteristics. the client, problem behavior, and the client–
Based on an examination of assessment informant relationship. The second section
conditions typically used in FA research on contains 16 questions that focus on antecedent
problem behavior (see Hanley et al., 2003, for a conditions under which problem behavior may or
description of many of these conditions), we may not occur, consequences that typically follow
developed lists of events that have been shown to problem behavior, or correlated behaviors. The
serve as motivating (establishing and abolishing) informant responds to each question by circling
operations or reinforcing consequences, and in yes or no to indicate that the events described do
some cases, behaviors that might be members of a or do not occur or N/A to indicate either a lack of
response class with problem behavior (e.g., information or that the question is not applicable
274 BRIAN A. IWATA et al.
to the client’s problem. Items 1 through 4 focus never versus Anchor 2 seldom, or Anchor 4 usually
on problem behavior maintained by social- versus Anchor 5 almost always on the MAS.
positive reinforcement, which has been shown
to occur when access to attention (Item 1) or
STUDY 1: RELIABILITY ANALYSIS
preferred items (Item 2) is restricted and when
these events are delivered as consequences (Item Method
3), but not usually when free access to positive Subjects, setting, and administration procedure.
reinforcement is available (Item 4). Items 5 Data were collected for 151 individuals drawn
through 8 focus on problem behavior maintained from three client populations (University of
by social-negative reinforcement, which often is Florida, Kennedy Krieger Institute, and New
accompanied by noncompliance (Item 5) and is England Center for Children), all of whom had
likely to occur in the presence of task or social been diagnosed with an intellectual disability or
demands (Item 6) when escape is available (Item autism and had been referred for assessment of
7), but not when demands are absent (Item 8). problem behavior. One hundred individuals were
Items 9 through 12 focus on problem behavior male and 51 were female; their ages ranged from 5
maintained by automatic-positive reinforcement, to 53 years (M ¼ 17.8 years). Some individuals
whose occurrence is unrelated to social interac- exhibited more than one behavior problem (see
tion (Item 9), may or may not be influenced by Table 1 for a complete listing), resulting in a total
the availability of stimulating activities (Items 10 sample of 196 behaviors for which a FAST was
and 12), and usually is maintained by sensory completed.
stimulation (Item 11). Items 13 through 16 focus Informants consisted of parents, relatives,
on problem behavior maintained by automatic- teachers, teacher aides, and direct-care staff
negative reinforcement. Pain-attenuating prob- who were responsible for the care or training of
lem behavior, to the extent that it occurs, seems to clients. Although this sample was heterogeneous
be cyclical (Item 13), to occur more often in and uncontrolled, it was representative of the
individuals who experience recurring medical population from whom information about
problems (Item 14), and when these problems are clients’ problem behavior would be sought.
present (Item 15) rather than absent (Item 16). The educational background of informants
The third section of the FAST consists of a who were training staff ranged from high school
scoring summary. Items from the second section diploma to master’s degree but was unknown for
are grouped according to behavioral function, family members. Two informants independently
and results are summarized by circling numbers
corresponding to questions for which a “yes”
answer was given. Table 1
We selected a yes–no format for answers rather Problem Behaviors for Which FAST and FA Data Were
a numerical scale for two reasons. First, we Collected
conducted an informal analysis of several rating
scales (including a preliminary version of the Problem behavior Study 1 Pairs of FASTs Study 2 FAs
FAST) and found that the factor that accounted Aggression 51 21
for the largest proportion of variability on Elopement 7 0
Inappropriate verbal 21 3
individual item agreement was the number of Noncompliance 8 1
response options: Reliability was inversely corre- Property destruction 24 6
Self-injury 66 31
lated with the number of choices. Second, labels Stereotypy 17 7
that define distinctions among numerical anchors Other (defecation, theft) 2 0
Total 196 69
seemed fairly arbitrary, as in Anchor 1 almost
276 BRIAN A. IWATA et al.
% AGREEMENT SCORES
Reliability analysis. Interrater reliability (agree- 61-70 -
Table 2
Interrater Reliability of Rating Scales Based on Item-by-Item Agreement
Study N Agreement
Motivation Assessment Scale (Durand & Crimmins, 1988)
Conroy et al. (1996) 14 Meana ¼ 56% (range, 25% to 78%)
Duker and Sigafoos (1998) 90 Mediana ¼ 63.3% (range, 54.4% to 71.1%)
Sigafoos et al. (1994) 18 Meana ¼ 41% (range, 18.8% to 62.5%)
Zarcone et al. (1991) 55 Meana ¼ 48% (range, 0% to 88%)
Questions about Behavioral Function (Matson & Vollmer, 1995)
Nicholson et al. (2006) 118 Mediana ¼ 78% (range, 69.5% to 84.8%)
Structured Interview Protocol (Sigafoos et al., 1993)
Sigafoos et al. (1993) 18 Meanb ¼ 43.3% (range, 11.1% to 83.3%)
Data unavailable for:
Behavior Analytic Questionnaire (Hauck, 1985)
Contextual Assessment Inventory (McAtee et al., 2004)
Functional Assessment for Multiple Causality (Matson et al., 2003)
Functional Assessment Checklist for Teachers and Staff (March et al., 2000)
GB Motivating Screening Tool (Barrera & Graver, 2009)
Motivation Analysis Rating Scale (Wieseler et al., 1985)
Problem Behavior Questionnaire (Lewis et al., 1994)
a
Adjacent numeric agreement.
b
Yes–no agreement.
half the values (except when two informants’ attention or when preferred activities are freely
scores are at the extreme opposite values), available?”) yielded the highest agreement
yielding 50% chance agreement that is equivalent (84.5%). Agreement was not noticeably different
to chance agreement for yes–no responses. for questions within a given functional category;
Relative to percentage agreement scores reported however, answers to questions about antecedent
for other scales, the FAST yielded comparable events usually were more reliable (M ¼ 78.9%)
reliability. than were answers to questions about consequent
Table 3 shows percentage agreement scores for events (M ¼ 67.7%). This finding was unusual
each of the 16 items included in the FAST. given that the antecedent event that precedes
Agreement on individual items varied widely: problem behavior often has passed by the time
Item 12 (“Is the problem behavior less likely to behavior occurs, which is one limitation of event-
occur when sensory stimulating activities are based descriptive analyses. Perhaps the higher
available?”) yielded the lowest agreement reliability for antecedent events reflected recol-
(53.3%), whereas Item 4 (“Is the person usually lection of repeated episodes of problem behavior
well behaved when he or she is getting lots of in an ongoing antecedent context.
Table 4 summarizes results obtained for
Table 3 agreement on FAST outcomes, that is, the extent
Percentage Agreement Scores for Individual FAST Items to which two informants’ most frequent yes
answers coincided for the same function.
Item Agreement Item Agreement
Whereas item-by-item agreement reflects the
1 70.8 9 77.4 extent to which informants provide the same
2 78.6 10 71.1
3 66.5 11 69.1 answers to the same questions, outcome agree-
4 84.5 12 53.3 ment reflects the extent to which informants,
5 70.6 13 70.2
6 67.9 14 80.3 regardless of their answers to individual ques-
7 67.5 15 69.9 tions, provide yes answers to more questions in
8 75.9 16 75.0
the same functional category. Using direct-
278 BRIAN A. IWATA et al.
However, because only moderate agreement informants also gave only one fewer yes answer
between pairs of informants was obtained in for a different function (Informant 1: social-
Study 1, arbitrary designation of one FAST as the negative reinforcement; Informant 2: automatic-
primary score for comparison with FA data positive reinforcement). Similarly, both of Kim’s
presented a problem: The function identified by informants gave more yes answers to questions
one FAST might correspond to the function that reflected maintenance by social-negative
identified by the FA, whereas the function reinforcement (corresponding to her FA results).
identified by the other FAST might not. As a Kim’s Informant 1, however, gave only one fewer
result, we compared each pair of FASTs to its yes answer for each of two additional functions
respective FA such that each FAST–FA compari- (social- and automatic-positive reinforcement).
son could yield complete (both respondents), Warren’s case was one of the few examples in
partial (one respondent), or no agreement and which FAST responses not only corresponded
used the corresponding values of 1, .5, and 0 with FA results but also showed a high degree of
when calculating agreement. Thus, correspon- differentiation among FAST responses. Thus, in
dence between one FAST and FA outcome was most cases of correspondence between FAST and
partially canceled out if correspondence was not FA outcomes, informant responses showed little
obtained between the other FAST and FA differentiation across categories; that is, conclu-
outcome. As was the case in calculating reliability sions about function based on FAST data were
for FAST outcomes, if an informant’s FAST determined by a one-question difference.
responses produced a tie for the highest function, Sarah’s Albert’s, and Heather’s results all
an agreement was scored if either matched the showed the absence of correspondence between
outcome of the FA. We conducted a second FASTand FA outcomes. As was typical in cases of
analysis of the data by comparing FAST and FA noncorrespondence, one (Sarah) or both (Albert
outcomes when both informants agreed on the and Heather) of the informants failed to identify
FAST outcome. This analysis answered the the function of problem behavior; in addition,
question, Given that two informants’ FASTs informants again (all three cases) showed little
identified the same function for problem behav- differentiation in responding to questions across
ior, did that function match the FA outcome? different categories. It also is interesting to note
that agreement on function between FAST
Results and Discussion informants did not improve validity. For exam-
Figure 3 shows representative samples of data ple, FAST informants agreed on the function of
illustrating correspondence (left) and noncorres- both Albert’s and Heather’s problem behaviors;
pondence (right) between FAST and FA out- this consensus, however, did not correspond with
comes for each of three functions: social-positive, their FA outcomes.
social-negative, and automatic-positive reinforce- Table 5 summarizes results of all of the FAST–
ment. The automatic-negative reinforcement FA comparisons. There were no cases in which
function does not appear because a typical FA the FA data indicated that problem behavior was
would not yield such an outcome. Each set of FA multiply controlled (maintained by more than
data showed very clear results, whereas the FAST one source of reinforcement); we attributed this
outcomes were not necessarily as clear, even when to the fact that we did not aggregate problem
they showed correspondence with the FA data. behaviors during assessment (Beavers &
For example, both of Sean’s informants gave more Iwata, 2011). Overall correspondence between
yes answers to questions that reflected mainte- FAST and FA outcomes was 63.8%, and the
nance by social-positive reinforcement (corre- highest degree of correspondence was obtained
sponding to his FA results). However, both when results of the FA indicated that problem
280 BRIAN A. IWATA et al.
S+ S- A+ A- S+ S- A+ A-
Sean #1: 3 2 1 0 Sarah #1: 2 2 3 0
40 #2: 3 1 2 0
10 #2: 4 3 3 0
SIB (responses/min)
SIB (responses/min)
8
30 Attention
6
20
Alone/ 4
10 Ignore Demand
Play 2
0 0
0 10 20 30 0 5 10 15
S+ S- A+ A- S+ S- A+ A-
Kim #1: 2 3 2 1 #1: 2 1 0 0
Albert
AGGRESSION (responses/min)
4 4
AGGRESSION (responses/min)
3 #2: 2 0 0 3 #2: 3 1 0
2 2
1 1
0 0
0 5 10 15 0 5 10 15
S+ S- A+ A- S+ S- A+ A-
#1: 0 0 3 0 #1: 4 1 3 0
Warren 3 Heather 4
PROP DESTRUCT (% intervals)
80 80
60 60
40 40
20 20
0 0
0 5 10 15 20 0 5 10 15 20
SESSIONS SESSIONS
Figure 3. Representative data from 69 FAST–FA comparisons (Study 2). Left and right columns show correspondence
and noncorrespondence, respectively, between informants’ FAST ratings and FA results. Numbers in boxes indicate
informants’ yes responses to items in each functional category: social-positive reinforcement (Sþ), social-negative
reinforcement (S), automatic-positive reinforcement (Aþ), and automatic-negative reinforcement (A). Each informant’s
highest score is shown in boldface.
FUNCTIONAL ANALYSIS SCREENING TOOL 281
Table 5
Summary of Agreements Between FAST and FA Outcomes
answers. In addition, a checklist can be complet- a check for inconsistencies and as a basis for
ed in a matter of minutes, creating the illusion clarifying questions or observations. Finally,
that the task is easily accomplished. However, when multiple informants strongly concur that
accurate answers require at least the following: an individual’s problem behavior has a specific
(a) having observed most of the events in function, verification by way of an FA might be
question, (b) organizing disparate events into more efficient through elimination of unneces-
categories contained on the rating scale, (c) sary test conditions. For example, we found
estimating conditional probabilities of events to (Study 2) that when two informants’ FAST scores
determine their relative influence on behavior, agreed that an individual’s problem behavior was
and (d) recalling all of these details while maintained by social-positive reinforcement, the
completing the checklist. Thus, accurate com- results were confirmed in every subsequent FA.
pletion of checklists, such as those currently being Smith, Smith, Dracobly, and Pace (2012) also
used to identify behavioral function, requires reported good correspondence between rating
skills that far exceed those of an individual who scale and FA outcomes when there was consensus
collects direct-observation data; he or she is among four of five informants on the MAS or
required only to detect the occurrence of an QABF. By extension, if several knowledgeable
ongoing event (rather than many that have informants agree that an individual’s problem
occurred in the past), discriminate it from others, behavior is maintained by attention, then perhaps
and mark it while it is being observed. Moreover, a single function test (Iwata & Dozier, 2008) that
it has been shown that correlations that result consists of contingent attention (test) and
even from highly reliable and accurate observa- noncontingent attention (control) conditions
tional (descriptive) data typically are not good might suffice to verify the function of problem
indicators of behavioral function (Thompson & behavior. Whether such a process actually
Borrero, 2011). As a result, it is unclear how decreases the amount of time required to conduct
unsystematic recollection of numerous past an FA is unknown because the benefit of
events is likely to be any better. efficiency (fewer FA conditions) must be weighed
What, then, might be the value of using against the cost of inaccuracy (missing a function
instruments such as the FAST during the course because it was not included in the FA).
of assessment? First, to the extent that structured Future research on function-based rating scales
questionnaires provide a uniform format for also might examine their adaptation for particular
gathering information, reduce the amount of response topographies. For example, scales such
time needed to prepare or conduct an interview, as the MAS, QABF, and FAST encompass
and have known reliability and validity (even multiple functions and thus seem applicable to
though both may be low), they might improve problem behavior that might be maintained by
the consistency and efficiency of clinical inter- varied sources of reinforcement. However, results
views. Second, information obtained from ques- from several studies suggest that stereotypy is
tionnaires might serve as the basis for follow-up most likely to be maintained by automatic
interviews and observations, which might facili- reinforcement (Querim et al., 2013), whereas
tate identification of idiosyncratic antecedent or aggression is most likely to be maintained by
consequent events to be included in an FA. For social reinforcement (Marcus, Vollmer, Swanson,
example, the first section of the FAST contains a Roane, & Ringdahl, 2001). Thus, structured
series of open-ended questions that we did not interviews that focus on these target behaviors
include in any analysis as part of this study. might include additional items related to specific
Perhaps answers to those questions could be functions while items related to unlikely func-
compared with answers to the FAST questions as tions are eliminated.
FUNCTIONAL ANALYSIS SCREENING TOOL 283
A distinct danger in formalizing the role of Autism and Developmental Disorders, 18, 99–117. doi:
verbal report in the assessment of behavior is its 10.1007/BF02211821
Ellingson, S. A., Miltenberger, R. G., & Long, E. S. (1999).
adoption for reasons other than those intended A survey of the use of functional assessment procedures
regardless of what the data show because verbal in agencies serving individuals with developmental
reports are easily obtained. Therefore, we disabilities. Behavioral Interventions, 14, 187–198. doi:
10.1002/(SICI)1099-078X(199910/12)14:4<187:AID.-
reemphasize the fact that the FAST is not an BIN38>3.0.CO;2-A
approximation to an FA of problem behavior; it is Green, C. W., Reid, D. H., Canipe, V. S., & Gardner, S.
simply one way to gather information during an (1991). A comprehensive evaluation of reinforcer
interview. Nevertheless, because the clinical identification processes for persons with profound
multiple handicaps. Journal of Applied Behavior
interview is a common context for posing all Analysis, 24, 537–552. doi: 10.1901/jaba.1991.24-
sorts of questions about behavior, efforts to 537
improve its consistency may be helpful. Green, C. W., Reid, D. H., White, L. K., Halford, R. C.,
Brittain, D. P., & Gardner, S. M. (1988). Identifying
REFERENCES reinforcers for persons with profound handicaps: Staff
opinion versus systematic assessment of preferences.
Barrera, F. J., & Graver, E. E. (2009). A comparison of Journal of Applied Behavior Analysis, 21, 31–43. doi:
behavioral functions in community and facility settings. 10.1901/jaba.1988.21-31
Journal on Developmental Disabilities, 15, 30–34. Groden, G. (1989). A guide for conducting a comprehensive
Barton-Arwood, S. M., Wehby, J. H., Gunter, P. L., & Lane, behavioral analysis of a target behavior. Journal of
K. L. (2003). Functional behavior assessment rating Behavior Therapy and Experimental Psychiatry, 20, 163–
scales: Intrarater reliability with students with emotion- 169. doi: 10.1016/0005-7916(89)90049-9
al or behavioral disorders. Behavioral Disorders, 28, Hanley, G. P., Iwata, B. A., & McCord, B. E. (2003).
386–400. Functional analysis of problem behavior: A review.
Beavers, G. A., & Iwata, B. A. (2011). Prevalence of multiply Journal of Applied Behavior Analysis, 36, 147–185. doi:
controlled problem behavior. Journal of Applied Behav- 10.1901/jaba.2003.36-147
ior Analysis, 44, 593–597. doi: 10.1901/jaba.2011.44- Hauck, F. (1985). Development of a behavior-analytic
593 questionnaire precising four functions of self-injurious
Cataldo, M. F., & Harris, J. (1982). The biological basis for behavior in the mentally retarded. International Journal
self-injury in the mentally retarded. Analysis and of Rehabilitation Research, 8, 350–352. doi: 10.1097/
Intervention in Developmental Disabilities, 2, 21–39. 00004356-198509000-00014
doi: 10.1016/0270-4684(82)90004-0 Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K. E., &
Conroy, M. A., Fox, J. J., Bucklin, A., & Good, W. (1996). Richman, G. S. (1994). Toward a functional analysis of
An analysis of the reliability and stability of the self-injury. Journal of Applied Behavior Analysis, 27,
Motivation Assessment Scale in assessing the challeng- 197–209. doi: 10.1901/jaba.1994.27-197 (Reprinted
ing behaviors of persons with developmental disabil- from Analysis and Intervention in Developmental
ities. Education & Training in Mental Retardation and Disabilities, 2, 3–20, 1982)
Developmental Disabilities, 31, 243–250. Iwata, B. A., & Dozier, C. L. (2008). Clinical application of
Derby, K. M., Wacker, D. P., Sasso, G., Steege, M., functional analysis methodology. Behavior Analysis in
Northup, J., Cigrand, K., & Asmus, J. (1992). Brief Practice, 1, 3–9.
functional assessment techniques to evaluate aberrant Iwata, B. A., Pace, G. M., Dorsey, M. F., Zarcone, J. R.,
behavior in an outpatient setting: A summary of 79 Vollmer, T. R., Smith, R. G., … Willis, K. D. (1994).
cases. Journal of Applied Behavior Analysis, 25, 713–721. The functions of self-injurious behavior: An experi-
doi: 10.1901/jaba.1992.25-713 mental-epidemiological analysis. Journal of Applied
Desrochers, M. N., Hile, M. G., & Williams-Mosely, T. L. Behavior Analysis, 27, 215–240. doi: 10.1901/jaba.
(1997). Survey of functional assessment procedures 1994.27-215
used with individuals who display mental retardation Iwata, B. A., Wong, S. E., Riordan, M. M., Dorsey, M. F., &
and severe problem behaviors. American Journal on Lau, M. M. (1982). Assessment and training of clinical
Mental Retardation, 101, 535–546. interviewing skills: Analogue analysis and field replica-
Duker, P. C., & Sigafoos, J. (1998). The Motivation tion. Journal of Applied Behavior Analysis, 15, 191–203.
Assessment Scale: Reliability and construct validity doi: 10.1901/jaba.1982.15-191
across three topographies of behavior. Research in Kelley, M. E., LaRue, R., Roane, H. S., & Gadaire, D. M.
Developmental Disabilities, 19, 131–141. doi: 10.1016/ (2011). Indirect behavioral assessments: Interviews and
S0891-4222(97)00047-4 rating scales. In W. W. Fisher, C. C. Piazza, & H. S.
Durand, V. M., & Crimmins, D. B. (1988). Identifying the Roane (Eds.), Handbook of applied behavior analysis (pp.
variables maintaining self-injurious behavior. Journal of 182–190). New York, NY: Guilford.
284 BRIAN A. IWATA et al.
Knoster, T. P. (2000). Practical application of functional Applied Behavior Analysis, 30, 165–167. doi: 10.1901/
behavioral assessment in schools. Journal of the jaba.1997.30-165
Association for Persons with Severe Handicaps, 25, Querim, A. C., Iwata, B. A., Roscoe, E. M., Schlichenmeyer,
201–211. doi: 10.2511/rpsd.25.4.201 K. J., Virues Ortega, J., & Hurl, K. E. (2013).
Lewis, T. J., Scott, T. M., & Sugai, G. (2004). The problem Functional analysis screening for problem behavior
behavior questionnaire: A teacher-based instrument to maintained by automatic reinforcement. Journal of
develop functional hypotheses of problem behavior Applied Behavior Analysis, 46, 47–60.
in general education classrooms. Assessment for Effective Sigafoos, J., Kerr, M., & Roberts, D. (1994). Interrater
Intervention, 19, 103–115. doi: 10.1177/ reliability of the Motivation Assessment Scale: Failure to
073724779401900207 replicate with aggressive behavior. Research in Develop-
March, R. E., Horner, R. H., Lewis-Palmer, T., Brown, D., mental Disabilities, 15, 333–342. doi: 10.1016/0891-
Crone, D., Todd, A. W., & Carr, E. (2000). Functional 4222(94)90020-5
Assessment Checklist: Teachers and Staff (FACTS). Sigafoos, J., Kerr, M., Roberts, D., & Couzens, D. (1993).
Eugene, OR: Educational and Community Supports. Reliability of structured interviews for the assessment
Marcus, B. A., Vollmer, T. R., Swanson, V., Roane, H. S., & of challenging behavior. Behaviour Change, 10, 47–50.
Ringdahl, J. E. (2001). An experimental analysis of Smith, C. M., Smith, R. G., Dracobly, J. D., & Pace, A. P.
aggression. Behavior Modification, 25, 189–213. doi: (2012). Multiple-respondent anecdotal assessments: An
10.1177/0145445501252002 analysis of interrater agreement and correspondence
Matson, J. L., Bamberg, J. W., Cherry, K. E., & Paclawskyj, with analogue assessment outcomes. Journal of Applied
T. R. (1999). Validity study on the Questions About Behavior Analysis, 45, 779–795. doi: 10.1901/jaba.
Behavioral Function (QABF): Predicting treatment 2012.45-779
success for self-injury, aggression, and stereotypies. Sturmey, P. (1994). Assessing the functions of aberrant
Research in Developmental Disabilities, 24, 163–176. behaviors: A review of psychometric instruments.
doi: 10.1016/S0891-4222(98)00039-0 Journal of Autism and Developmental Disorders, 24,
Matson, J. L., Kuhn, D. E., Dixon, D. R., Mayville, S. B., 293–304. doi: 10.1007/BF02172228
Laud, R. B., Cooper, C. L., … Matson, M. L. (2003). Thompson, R. H., & Borrero, J. C. (2011). Direct
The development and factor structure of the Functional observation. In W. W. Fisher, C. C. Piazza, & H. S.
Assessment for Multiple Causality (FACT). Research in Roane (Eds.), Handbook of applied behavior analysis
Developmental Disabilities, 24, 485–495. doi: 10.1016/ (pp. 191–205). New York, NY: Guilford.
j.ridd.2003.07.001 Van Acker, R., Boreson, L., Gable, R. A., & Potterton, T.
Matson, J. L., & Vollmer, T. R. (1995). User’s guide: (2005). Are we on the right course? Lessons learned
Questions About Behavioral Function (QABF). Baton about current FBA/BIP practices in schools. Journal of
Rouge, LA: Scientific Publishers. Behavioral Education, 14, 35–56. doi: 10.1007/
McAtee, M., Carr, E. G., & Schulte, C. (2004). A contextual s10864-005-0960-5
assessment inventory for problem behavior: Initial Wieseler, N. A., Hanson, R. H., Chamberlain, T. P., &
development. Journal of Positive Behavior Interventions, Thompson, T. (1985). Functional taxonomy of
6, 148–165. doi: 10.1177/10983007040060030301 stereotypic and self-injurious behavior. Mental Retarda-
Nicholson, J., Konstantinidi, E., & Furniss, F. (2006). On tion, 23, 230–234.
some psychometric properties of the Questions about Zarcone, J. R., Rodgers, T. A., Iwata, B. A., Rourke, D., &
Behavioral Function (QABF) scale. Research in Devel- Dorsey, M. F. (1991). Reliability analysis of the
opmental Disabilities, 27, 337–352. doi: 10.1016/j. Motivation Assessment Scale: A failure to replicate.
ridd.2005.04.001 Research in Developmental Disabilities, 12, 349–360.
O’Neill, R. E., Horner, R. E., Albin, R. W., Storey, K., & doi: 10.1016/0891-4222(91)90031-M
Sprague, J. R. (1990). Functional analysis of problem
behavior: A practical assessment guide. Sycamore, IL:
Sycamore. Received September 24, 2012
O’Reilly, M. F. (1997) Functional analysis of episodic self- Final acceptance November 26, 2012
injury correlated with recurrent otitis media. Journal of Action Editor, Richard Smith
Supporting Information
Appendix S1. Functional Analysis Screening Tool (FAST) Full Page Version