Research in Developmental Disabilities 30 (2009) 682–688
Contents lists available at ScienceDirect
Research in Developmental
Disabilities
Aggressive behavior in individuals with moderate to
borderline intellectual disabilities who live in a
residential facility: An evaluation of functional variables
Petri J.C.M. Embregts a,*, Robert Didden b, Nicole Schreuder c,
Cecile Huitink a, M. van Nieuwenhuijzen d
a
Department of Special Education/Behavioral Science Institute, Radboud University Nijmegen, P.O. Box 9104, 6500 HE,
Nijmegen, The Netherlands
b
Radboud University Nijmegen and Trajectum, The Netherlands
c
Trajectum-Hanzeborg, The Netherlands
d
Developmental Psychology, Utrecht University, The Netherlands
A R T I C L E I N F O
A B S T R A C T
Article history:
Received 21 April 2008
Accepted 24 April 2008
We explored functional variables for aggressive behavior in 87
individuals with moderate to borderline intellectual disability who
lived in a residential facility. For this purpose we used the Questions
About Behavioral Function scale (QABF; Matson, J., & Vollmer, T.
(1995). Questions About Behavioral Function (QABF). Baton Rouge,
LA: Scientific Publications). Results show that in most clients
subscales describing social function (i.e., Attention, Escape/Avoidance, Tangible) had significantly higher mean scores than subscales
describing non-social function (i.e., Self-stimulation, Physical
discomfort). Except for gender, there were no significant associations between mean subscale scores and client variables, such as
psychiatric disorder, age, level of intellectual disability. Female
clients had higher mean scores on subscales of Attention, Selfstimulation, and Physical discomfort than male clients. Results of
our study suggest that in most cases, aggressive behavior is
positively/negatively reinforced by social events. Implications for
functional assessment and function-based treatment of aggressive
behavior in these clients are discussed.
ß 2008 Elsevier Ltd. All rights reserved.
Keywords:
Aggressive behavior
Functional assessment
Intellectual disabilities
* Corresponding author. Tel.: +31 24 361 28 22; fax: +31 24 361 6211.
E-mail addresses: [email protected] (Petri J.C.M. Embregts), [email protected] (M. van Nieuwenhuijzen).
0891-4222/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ridd.2008.04.007
P.J.C.M. Embregts et al. / Research in Developmental Disabilities 30 (2009) 682–688
683
1. Introduction
Aggressive behavior is one of the most problematic and common type of challenging behaviors
among persons with severe to mild intellectual disabilities (ID). In the past two decades, functional
analysis approach has provided methods for identifying the cause or function of challenging behaviors
and its behavioral treatment (Didden, 2007). The Question About Behavioral Function (QABF; Matson &
Vollmer, 1995) has been designed for the purpose of functional assessment of challenging behavior in
individuals with ID and as an alternative for experimental functional analysis (see e.g., Thompson &
Iwata, 2001). The QABF is a 25-item questionnaire designed for identifying functional variables
maintaining problem behavior, such as aggression, self-injury and other types of challenging behavior.
It has five subscales corresponding to five possible functions of challenging behavior: (a) Attention, (b)
Tangible, (c) Self-stimulation, (d) Physical discomfort, and (e) Escape/Avoidance. Two subscales (i.e.,
Self-stimulation and Physical discomfort) describe non-social functions whereas the other subscales
describe social functions. The QABF has been shown to be a reliable and valid instrument for assessing
behavioral function of challenging behavior in individuals with ID (see Section 2.2). When applied in
larger groups of clients, outcomes of the QABF may facilitate an understanding of the most common
cause(s) of challenging behaviors, such as aggression and self-injury, in clients with ID.
Applegate, Matson, and Cherry (1999) were the first to explore functional variables of challenging
behavior in a large group of individuals with ID. Functional assessment was assessed for five types of
severe challenging behaviors (i.e., aggression, stereotypies, pica, self-injury, rumination) in 417
individuals with severe or profound ID by using the QABF. They found that the most common
behavioral function for all behaviors except aggression was non-social although social functions were
endorsed as well. Aggressive behavior was primarily maintained by social consequences with Escape/
Avoidance as highest in rank followed by Tangible and Attention. Non-social subscales (i.e., Selfstimulation, Physical discomfort) had low mean scores. Furthermore, the most frequently endorsed
items on the QABF for individuals with aggression were ‘Seems to be saying ‘‘Leave me alone’’’ or ‘Stop
asking me to do this’ and ‘Engages in the behavior to try to get people to leave him/her alone.’
Two years later, Matson and Mayville (2001) examined the function variables of aggressive
behaviors among 135 individuals with severe or profound ID. Results show that in most cases
aggressive behavior served a social function (e.g., Escape/Avoidance, Tangible, Attention).
Furthermore, almost 50% of their sample met criteria for a psychiatric disorder (e.g., Mood Disorder,
Autism) as measured by the Diagnostic Assessment for the Severely Handicapped-II (DASH-II; Matson,
1995). Further analyses revealed that there were no differences in mean QABF subscale score between
individuals with dual diagnosis and those without, except for the Attention subscale whereby
individuals with a psychiatric disorder had significantly higher mean subscale scores than those
without a psychiatric disorder. In general, presence of psychiatric disorder appeared to be unrelated to
any of the functional variables for aggressive behavior in their sample.
Recently, Didden, Korzilius, and Curfs (2007) have investigated functional variables of selfinjurious skin-picking in a relatively large sample of individuals with Prader-Willi syndrome (n = 119).
Outcomes on the QABF suggested that in most cases (i.e., 70%) this type of behavior primarily had nonsocial functions. That is, subscales that indicate non-social behavioral function of skin-picking were
highest in rank, a result which is in agreement with those found by Applegate et al. (1999) in
individuals with more severe levels of ID. In most individuals with this syndrome skin-picking may be
viewed as an operant maintained by contingent sensory reinforcement and/or escape from unpleasant
physical stimulation such as pain, itch and overarousal and anxiety. Such outcomes provide clues for
developing effective behavioral treatments for skin-picking in individuals with Prader-Willi
syndrome. This study was the first to explore relationships between a specific genetic disorder
and functional variables of a behavior that belongs to the behavioral phenotype of this disorder.
Results of the above studies suggest that in most cases aggressive behavior in individuals with
profound or severe ID is related to social functional variables, while self-injurious behaviors are
primarily maintained by non-social consequences. Until present, large n studies on the use of
instruments such as the QABF for functional assessment of aggressive behavior in individuals with
mild ID are lacking. Furthermore, associations between functional variables and other variables (e.g.,
psychiatric disorder, level of ID) have seldomly been explored (see e.g., Matson & Mayville, 2001). The
684
P.J.C.M. Embregts et al. / Research in Developmental Disabilities 30 (2009) 682–688
main purpose of our study then was, to explore functional variables of aggression in 87 clients with
moderate to borderline ID who showed aggressive behavior and who lived in a residential facility. For
this purpose, we used a Dutch version of the QABF. We also explored associations between client
characteristics (e.g., frequency of aggressive incidents, gender, age, IQ level, and psychiatric disorder)
and mean scores on QABF subscales.
2. Method
2.1. Participants and procedure
Managers of four residential facilities were asked to identify clients who met three inclusion
criteria. Individuals were included if they: (a) functioned in the moderate to borderline range of ID (IQ
50–85), (b) were residing at the facility for at least 3 months, and (c) regularly showed severe
aggressive behavior. Aggressive behavior was defined as any verbal/physical behavior directed against
another person and for the purpose of physically and psychologically threatening that person and/or
damaging an object. The aggressive behavior is of such an intensity, frequency, or duration that the
physical safety of the person and/or others is placed in serious jeopardy and/or behavior that is likely
to seriously limit social functioning. Individuals were included if they had shown at least three
incidents of severe aggressive behavior during the last 3 months (Jahoda & Wanless, 2005).
A total of 87 clients met the inclusion criteria. Their mean age was 36 years (range: 13–76 years),
and 62 clients (i.e., 71.3%) were male. Level of ID was borderline in 19 cases (i.e., 27%), mild in 37 cases
(i.e., 53%), and moderate in 14 cases (i.e., 20%). Fifty-three clients (i.e., 61%) had a DSM-IV psychiatric
disorder established by a certified psychiatrist or clinical psychologist. Of these clients, 20 (i.e., 33%)
were diagnosed with Autism spectrum disorder, 13 (i.e., 25%) had Personality disorder, 10 (i.e., 19%)
had Attention Deficit Hyperactivity Disorder, and 10 (i.e., 19%) were diagnosed with Mood disorder.
Clients lived in the facility for a mean of 4.5 years (range: 3 months – 27 years). All clients had given
consent to participate in this study.
Respondents were 87 direct care staff members. Each staff member was asked to complete the
Questions About Behavioral Function (QABF) and a questionnaire addressing demographic information
(see below). Each staff member had worked with the client for at least 3 months. On average, they had
worked with a client for 2.10 years.
2.2. Instruments
2.2.1. Questions About Behavioral Function
The QABF (Matson & Vollmer, 1995) is a 25-item questionnaire designed to identify (functional)
variables maintaining problem behavior in persons with intellectual disability. There are five subscales
corresponding to five possible functions of problem behaviors, such as aggressive behaviors: (a)
Attention, (b) Tangible, (c) Self-stimulation, (d) Physical discomfort, and (e) Escape/Avoidance. Two of
the subscales (i.e., Self-stimulation and Physical discomfort) describe non-social functions whereas the
other subscales describe social functions of problem behavior in individuals with ID. Each subscale has
five items and each item is rated in a 4-point Likert format reflecting how often a target behavior occurs
(0 = never, 1 = rarely, 2 = sometimes, 3 = often). The total subscale score ranges from 0 to 15 (max).
In several studies, the psychometric properties of the QABF have been found to be good to excellent.
For example, Paclawskyj, Matson, Rush, Smalls, and Vollmer (2000) have found good interrater and
test–retest reliability and internal consistency and factor analysis have yielded five subscales.
Furthermore, Matson, Bamburg, Cherry, and Paclawskyj (1999) found the QABF to be valid, that is
treatments designed upon outcomes of the QABF were more effective (in terms of reduction in target
behaviors such as self-injury, aggressive behaviors and stereotypy) than treatments that were not
designed upon identified function(s) with the QABF. Paclawskyj, Matson, Rush, Smalls, and Vollmer
(2001) also found that the QABF has convergent validity when compared to outcomes of analogue
baseline methodology and other checklists such as the Motivation Assessment Scale.
We translated the QABF into Dutch and the translated version was checked by a native English
speaker. We reworded item 5 and item 25 (e.g., ‘toy’ was replaced by ‘object’). For each client, mean
P.J.C.M. Embregts et al. / Research in Developmental Disabilities 30 (2009) 682–688
685
score on each QABF subscale was calculated, as a result of which a rank order of subscales could be
established. The subscale with the highest rank indicates primary functional variable for aggressive
behavior (Matson & Vollmer, 1995).
2.2.2. Questionnaire
Direct care staff members also completed a questionnaire addressing demographic information of
their client (i.e., gender, age, level of intellectual disability, presence of psychiatric disorder, and
frequency of aggressive incidents per month).
2.3. Statistical analyses
First, internal consistency coefficient (i.e., Cronbach’s alpha) for the total QABF as well as for its
subscales were calculated. Then, differences in mean scores between the five QABF subscales were
analyzed using paired t-tests. Finally, associations between mean QABF subscale scores and the
following variables were explored: frequency of aggressive incidents, gender, presence of psychiatric
disorder (i.e., ADHD, Personality disorder, Autism spectrum disorder, and Mood disorder), level of ID,
and age in years.
3. Results
Cronbach’s alpha of the total scale was .90 and coefficients for the subscales ranged from .72 to .83.
Internal consistency of the total QABF and its subscales may be considered good to excellent.
Mean scores on the five subscales ‘Attention’, ‘Escape/Avoidance’, ‘Tangible’, ‘Physical discomfort’,
and ‘Self-stimulation’ were 7.68 (range: 0–4), 8.23 (range: 0–15), 7.95 (range: 0–15), 7.16 (range:
0–15), and 4.94 (range: 0–12), respectively.
Results depicted in Table 1 show that mean scores for subscales that describe social functions (i.e.,
Attention, Escape/Avoidance, Tangible) of aggressive behavior were significantly higher than the
mean score of the subscale describing non-social function (i.e., Self-stimulation). Mean score for
‘Physical discomfort’ was significantly higher than mean score on ‘Self-stimulation’. There were no
significant differences in mean scores for subscales describing social functions (i.e., Attention, Escape/
Avoidance, Tangible) and the subscale ‘Physical discomfort’.
Table 1
Mean QABF-subscale scores, t and p values.
Subscales
Mean
Attention
Escape/Avoidance
Attention
Self-stimulation
Attention
Physical discomfort
Attention
Tangible
Escape/Avoidance
Self-stimulation
Escape/Avoidance
Physical discomfort
Escape/Avoidance
Tangible
Self-stimulation
Physical discomfort
Self-stimulation
Tangible
Physical discomfort
Tangible
7.68
8.23
7.68
4.94
7.68
7.16
7.68
7.95
8.23
4.94
8.23
7.16
8.23
7.95
4.94
7.16
4.94
7.95
7.16
7.95
NB. **p < .01; ***p < .001.
t(86)
p
1.28
.20
7.91
.000***
1.20
.24
0.62
.54
8.69
.000***
2.60
.01**
0.65
.52
5.53
.000***
7.42
.000***
1.57
.12
686
P.J.C.M. Embregts et al. / Research in Developmental Disabilities 30 (2009) 682–688
Finally, we calculated the number of individuals who had the highest rank on each of the five
subscales, which may indicate the functional variable maintaining the aggressive behavior. In 42.5% of
clients, the subscale with the highest score was ‘Tangible’, ‘Escape/Avoidance’ was highest in 34.5% of
the clients, ‘Attention’ and ‘Physical discomfort’ had the highest rank in 28.7% and 19.5% of the clients,
respectively, and ‘Self-stimulation’ had the highest rank in 2.3% of the clients. Subscales that indicate
social behavioral function of aggressive behavior were highest in rank in 78.2% of the clients, while
rank order in the other clients is suggestive of a non-social function of aggression.
Our results suggest that in most clients aggressive behavior may be viewed as an operant
maintained by (reinforcing) socially mediated consequences, such as social attention, escape from
and/or avoidance of task demands or preferred tangibles. On the remaining cases, aggressive behavior
may be maintained by contingent automatic (e.g., sensory, physical) reinforcement and/or escape
from or avoidance of unpleasant physical stimulation, such as overarousal, anxiety or other types of
physical discomfort.
3.1. Associations between other variables and QABF subscales
3.1.1. Frequency of aggressive incidents per month
Except for ‘Physical discomfort’ (r = .18, p = .104), frequency of aggressive incidents per month (as
recorded by staff members in their daily logs) was positively and significantly correlated with mean
score on each QABF subscale: ‘Attention’ (r = .23, p < .05), ‘Escape/Avoidance’ (r = .25, p < .05), ‘Selfstimulation’ (r = .28, p < .05), and ‘Tangible’ (r = .28, p < .05). Four out of five subscales (i.e., functional
variables) were associated with the frequency of aggressive incidents.
3.1.2. Gender
Females (mean = 16.44) had a significantly higher rate of aggressive behavior per month than males
(mean = 8.12), t(80) = 2.98, p < .05. Due to male:female ratio of 3:1, we have also conducted a Mann–
Whitney analysis that shows that female clients had a significantly higher frequency of aggressive
behavior per month (mean rank = 51.14) than male clients (mean rank = 37.27), z = 2,45, p < .05.
Furthermore, females had a significantly higher mean score than males on the subscales
‘Attention’, t(87) = 2.80, p < .01, and ‘Physical discomfort’, t(87) = 2.50, p < .05. There were no
significant differences between males and females with respect to the subscales ‘Self-stimulation’,
t(87) = 1,99, p = .050, ‘Escape/Avoidance’, t(87) = 1.51 and ‘Tangible’, t(87) = 1,76, p = .08. Results
from Mann–Whitney tests revealed that mean rank scores for males were significantly lower than
those of females on three out of five subscales: ‘Attention’, z = 2,83, p < .01, ‘Self-stimulation’,
z = 1,97, p < .05, and ‘Physical discomfort’, z = 2,48, p < .05. For these functional variables,
aggressive behavior was more strongly related to female clients than male clients. There were no
significant differences in mean rank between males and females for the other subscales.
3.1.3. Psychiatric disorder
Results from paired t-tests revealed that there were no significant differences in mean subscale
scores between individuals with a Personality disorder, ADHD, Autism spectrum disorder, Mood
disorder and those without a psychiatric disorder on each of the five QABF subscales.
3.1.4. Level of intelligence and age
Regression analyses revealed that neither level of intellectual functioning (IQ) nor age in years
were significantly related to mean QABF subscale scores.
4. Discussion
In the present study we (a) investigated functional variables for aggressive behavior, and (b)
explored whether client variables were related to behavioral function of aggressive behavior in 87
individuals with moderate to borderline ID who lived in a residential facility. Results show in most
clients (i.e., 78%) aggressive behavior primarily had a social function (i.e., Escape/Avoidance, Attention,
Tangible) whereas in a minority the function of aggressive behavior was a non-social one (i.e., Physical
P.J.C.M. Embregts et al. / Research in Developmental Disabilities 30 (2009) 682–688
687
discomfort, Self-stimulation). These results are in agreement with those of studies by Applegate et al.
(1999) and Matson and Mayville (2001) although their conclusions were based on data collected in
clients with profound or severe ID.
Our findings as well as those from other studies suggest that in most clients with mild ID,
aggressive behavior is maintained and positively reinforced by social events (e.g., access to preferred
objects, activities or attention) and/or negatively reinforced by escape from or avoidance of social
events (e.g., task demands). Results provide important clues for the design of effective behavioral
treatments. Treatment may then consist of changing social environmental events that evoke
aggressive behaviors (i.e., antecedent control, change of establishing operations) and maintain it
(extinction, differential reinforcement, non-contingent reinforcement). Also, training of social and
adaptive skills to clients with mild ID may consist of teaching them to request attention and/or
tangibles (objects, activities) and to escape from or avoid unpleasant social events such as difficult task
demands and/or negative interactions.
Except for gender, there were no associations between mean QABF subscale scores and client
variables. In our sample, aggressive behavior in female clients was more associated with Attention,
Self-stimulation, and Physical discomfort than in male clients. Furthermore, we found no associations
between psychiatric disorders and mean scores on any of the subscales, a result that is in agreement
with those found by Matson and Mayville (2001). We nevertheless agree with Matson and Mayville
who have underscored the significant role that environmental events play in aggressive behavior in
individuals with dual diagnosis. Our knowledge about variables that may affect functional variables of
aggressive behavior remains limited and further studies are clearly needed in this area.
An important shortcoming of our study is its descriptive nature. The QABF identifies behavioral
functions indirectly, as opposed to labor-intensive and expensive naturalistic observations (see e.g.,
Thompson & Iwata, 2001). However, questionnaires such as the QABF have important clinical
advantages, they are more efficient, economical and easy to use than naturalistic observations.
Furthermore, Matson et al. (1999) and Paclawskyj et al. (2001) have shown that the QABF has excellent
convergent and predictive validity and that it is a useful clinical tool for identifying the behavioral
function of challenging behavior in clients with mild ID and design effective behavioral treatments
based on such outcomes. In order to integrate findings from several sources to assess the function(s) of
challenging behavior in individuals with ID, we propose that the QABF be completed together with an
instrument for the identification of more or less specific contextual events that elicit such behavior,
such as for example the Contextual Assessment Inventory (CAI; McAtee, Carr, & Schulte, 2004) (see
Embregts, Didden, Huitink, & Schreuder, submitted, for a study on contextual variables of aggressive
behavior in clients with mild ID and aggressive behavior). Outcomes of both QABF and CAI may
provide a thorough behavioral functional assessment of challenging behavior in individuals with ID
who live in a residential facility.
Despite its descriptive nature, findings of our study suggest that several functional variables may
operate on aggressive behavior in clients with mild ID who live in a residential facility. Clients with
mild ID may display aggression for a variety of wants and needs. Findings from large scale studies with
the QABF, such as the present one, may increase our understanding of the most common functions of
aggressive behavior in clients with (mild) ID. As such, its results may provide a ‘profile’ of the most
common function(s) of problem behaviors (Applegate et al., 1999) as well as clues for designing
effective and function-based behavioral treatments in an attempt to reduce aggressive behavior in
clients with (mild) ID.
Acknowledgement
We would like to thank clients and staff of the residential facilities (i.e., Amarant, Aveleijn,
Dichterbij and Trajectum-Hanzeborg) for their participation.
References
Applegate, H., Matson, J., & Cherry, K. (1999). An evaluation of functional variables affecting severe problem behaviors in adults with
mental retardation by using the Questions About Behavioral Function Scale (QABF). Research in Developmental Disabilities, 20,
229–237.
688
P.J.C.M. Embregts et al. / Research in Developmental Disabilities 30 (2009) 682–688
Didden, R. (2007). Functional analysis methodology in developmental disabilities. In P. Sturmey (Ed.), Functional analysis in clinical
treatment (pp. 65–86). London: Academic Press.
Didden, R., Korzilius, H., & Curfs, L. (2007). Skin-picking in individuals with Prader-Willi syndrome: Prevalence, functional
assessment, and its comorbidity with compulsive and self-injurious behavior. Journal of Applied research in Intellectual
Disabilities, 20, 409–419.
Embregts, P., Didden, R., Huitink, C., & Schreuder, N. Contextual variables affecting aggressive behavior in individuals with mild
intellectual disability who live in a residential facility, submitted for publication.
Jahoda, A., & Wanless, L. (2005). Knowing you: The interpersonal perception of staff towards aggressive individuals with mild to
moderate intellectual disabilities in situations of conflict. Journal of Intellectual Disability Research, 49, 544–551.
Matson, J. (1995). The Diagnostic Assessment for the Severely Handicapped-Revised (DASH-II). Baton Rouge, LA: Scientific Publications.
Matson, J., Bamburg, J., Cherry, K., & Paclawskyj, T. (1999). A validity study on the Questions About Behavioral Function (QABF) scale:
Predicting treatment success for self-injury, aggression, and stereotypes. Research in Developmental Disabilities, 20, 163–176.
Matson, J., & Mayville, E. (2001). The relationship of functional variables and psychopathology to aggressive behavior in persons with
severe to profound mental retardation. Journal of Psychopathology and Behavioral Assessment, 23, 3–9.
Matson, J., & Vollmer, T. (1995). Questions About Behavioral Function (QABF). Baton Rouge, LA: Scientific Publications.
McAttee, M., Carr, E., & Schulte, C. (2004). A contextual assessment inventory for problem behavior: Initial development. Journal of
Positive Behavioral Interventions, 6, 148–165.
Paclawskyj, T., Matson, J., Rush, K., Smalls, Y., & Vollmer, T. (2000). Questions About Behavioral Function (QABF): A behavioral
checklist for functional assessment of aberrant behavior. Research in Developmental Disabilities, 21, 223–229.
Paclawskyj, T., Matson, J., Rush, K., Smalls, Y., & Vollmer, T. (2001). Assessment of the convergent validity of the Questions About
Behavioral Function scale with analogue functional analyses and the Motivation Assessment Scale. Journal of Intellectual
Disability Research, 45, 484–494.
Thompson, R., & Iwata, B. (2001). A descriptive analysis of social consequences following problem behavior. Journal of Applied
Behavior Analysis, 34, 169–178.