Managing Challenging Behaviours in Disabilities

Introduction

Challenging behaviours (CBs) are culturally unusual behaviours, which threatens either the selves and/or others (Emerson & Bromley, 1995). The individuals with intellectual or developmental disabilities are more likely to engage with the challenging behaviours (Devine, 2014). These behaviours make it difficult for the individuals to integrate with their community and society in general, thereby reducing ones’ overall quality of life and their family (Edrisinha, 2011; Argumedes, Lanovaz, & Larivee, 2018). According to Repp and Horner (1999), challenging behaviours were often maintained by its reinforcing consequences, even sometimes, it initially emitted by biological/physical factors. Repp and Horner (1999) noted that, common reinforcing consequences include tangibles, social attention, demand avoidance, and sensory satisfaction (automatic reinforcement). Repp and Horner (1999) also argued that, although indirect assessment (e.g., interview) is advantageous for high efficiency and low cost, its reliability can be inevitably influenced by the confounding variables. For example, if an individual engages in more CB, when wanting tangibles (e.g., toy), the tangible cannot be determined as the maintaining variable, because the CB could also be maintained by the attention of someone providing the tangibles. Experimental analysis, however, by only manipulating one variable and controlling all others to stay the same, can determine whether/how much the manipulated variable (independent) changes/maintains the target behaviour (dependent) (Repp & Horner, 1999).

Behaviour of Interest.

Recalling from baseline data collection, the CBs will be recorded in different methods. Hand biting and aggression will be recorded, based on the frequency (how many times), because they terminate quickly. Noise making and Hand flapping will be recorded by duration (how long), because different occurrence of them varies a lot in time.

Hand Biting. This is a self-injurious behaviour, defined as when Sam were observed using his teeth to touch any part of his hand, no matter on his fingers or palms. To note that behaviours would be recorded as biting only when the teeth are clearly observed. A clear non-example would be hand mouthing, when Sam only puts his hand(s) into/next to his mouth, while he does not actually bite. This behaviour is important because it may hurt or pose risk to significant harm to the self (Soke et al., 2018).

Aggression. Aggressive behaviours toward people and objects will both be recorded. Aggression towards the people includes (1) pinching the experimenter with two or more fingers, (2) hitting on the experimenter with his open/closed fists, and (3) shouting towards the experimenter accompanied with staring on him or the toys held by him. Aggression towards objects, in this case, was defined as throwing toys away with a force for any distance. Aggressive behaviours are paid attention for the harm or potential harm towards others and property.

Noise Making. This is a form of vocal stereotyped behaviour, consisting of nonsense vocalisations and huff/puff, no matter what specific type is displayed. Nonsense vocalisation is the unintelligible shouting or muttering, that can be clearly heard but cannot be understood by the observer and/or not responded by the experimenter. Non-examples include phrases that (1) gained response, (2) repeated the instructions before conducting tasks, (3) can be clearly detected for sharing/requiring (e.g., “comes”, “help”, “back”). Huff and puff were recorded, when his breathe was observed louder than his normal (which cannot be clearly heard without purpose) and no jumping/shouting were engaged before its occurrence (i.e., feeling tired).

Hand flapping. This is a motor stereotyped behaviour defined as sudden/rapid hand movement, which does not gain any response (i.e., not for communication). Flapping a single hand or two hands together are both included. Hand gestures for communication were not considered as problematic (e.g. showing this fist and saying “respect”). Vocal/motor stereotyped behaviour was considered because of its potential to impair one's ability to learn (Ming, Brimacombe, & Wagner, 2007), communicate (Shriberg et al., 2001), and socialise (Jones, Wint, & Ellis, 1990).

Literature Review

Related literature of some common functional analysis design (from Iwata et al., 1982 to nowadays) will be reviewed in this session in terms of procedure, advantages, and disadvantages. These designs include multielement design, brief functional analysis, latency functional analysis, and trial-based functional analysis.

Multielement Design. Iwata et al. (1982) conducted multielement design to investigate maintaining variables of the participants’ self-injurious behaviours. Each participant was assigned to some different conditions (e.g., social disproval, academic demand) in a randomly decided order. Eight sessions for each condition was conducted to avoid the influence of confounding variables. The data was presented by the lines with the same number of conditions. Each line consisted data points (CBs’ frequencies/durations) of difference sessions within a same condition. If one (or more) line was higher, the manipulated variable(s) was suggested to be reinforcing/maintaining for the CBs.

Iwata et al., (1982) argued that, multielement design was advantageous of avoiding the confounding variables and therefore increasing result reliability, by conducting each condition for more than one time.

In terms of disadvantage, first, repeating a behaviour for more than one time restricted its application of severe and low-intensity behaviours. Second, it is complex and time-consuming – plenty of trials to rule out and record, which reduced its operability for practitioners (Northup et al., 1991). For instance, Iwata et al.’s study (1982) measured the CBs of participants in four conditions and each condition was repeated for eight sessions. Therefore, 32 sessions were conducted for one participant. Third, repeatedly providing reinforcing consequences on CBs is likely to build up a continuous reinforcement schedule, which could increase their future frequency.

Brief Functional Analysis (BFA). According to Northup et al. (1999), similar with multielement design, each participant was assigned into different conditions. The difference was that instead of manipulating each condition for many times (e.g., Iwata et al., 1982), each condition was conducted only once. After all conditions finished, the condition(s) with a higher rate of CB was conducted for another time to avoid the influence of confounding variables.

BFA is more ethical to measure severe aggressive or self-injurious behaviours, because if a behaviour brings severe hurt/damage to self/others/properties, emitting for many times is not safe. Additionally, this design saves resources (e. g., time) spent on each individual and thus, more persons will have the opportunity to be assessed (Northup et al., 1999).

However, without manipulating each condition/variable repeatedly, it is hard to measure their maintaining effect, because the influence of other stimuli under the same condition cannot be excluded (Iwata & Dozier, 2008), such as participants’ pre-session condition (sleep deprivation, drug intake), deprivation/satiation of reinforcers.

Latency Functional Analysis. Davis et al. (2013) conducted this for studying elopement, where different conditions were manipulated for more than one times in a randomly decided order. The data, being recorded, is the time length between the beginning of a session and the initiation of CBs. When CBs occurred, the participants were provided by reinforcing consequences immediately. This experimental design is based on the assumption that, the latency should be shorter, when the possibly provided items are maintaining/reinforcing to the CBs (Kazdin, 1982).

In experimental conditions (attention, tangible, and demand condition), once the target behaviour occurred and the consequence was provided, the session was ended. In the unstructured play condition, the target behaviour was ignored. Therefore, before the experiment, the researcher needed to set a maximum time length. If the target behaviour did not appear in a limited time, the session would be terminated. Latency functional analysis had a few advantages (Davis et al., 2013). Firstly, it reduced the confounding effect of extra attention. This was observed especially in demand sessions (reinforcing escaping behaviour), where the participants were easy to run away. In order to emit a second occurrence of CB, experimenters needed to retrieve them back and asked them to do the task. Such retrieval inevitably provided them with extra attention, so that maintaining effect of escaping was easily to be confounded by the attention. Recording the latency, however, eliminated this effect since it did not require more than one occurrence of the target behaviour. Moreover, demonstrating the consequences for only one time reduced the chance that CB will be reinforced.

However, Davis et al. (2013) also mentioned the limitations of this method. First, without the repeated manipulation, the influence of confounding variables cannot be excluded. Second, the participants might find new reinforcers, when CBs occurred. Davis et al. (2013) provided an example of demand condition, where the participant successfully escaped from task demand (by CBs) and found some toy interesting/attractive. The next time he/she engaged in CBs, it was hard to determine, whether his motivation was the escaping or toy.

Trial-Based Functional Analysis. This design was always applied into school settings, whose most essential component of was the immediate transfer from the control condition to one of experiment conditions (or in the opposite direction). In Bloom et al.’s study (2011), after providing free access of reinforcers for a period (this is the control condition, usually 1-2 minutes), experimenters immediately switched into a non-access condition (experiment condition). Reinforcers will be contingent on the occurrence of CBs in non-access conditions. The opportunities of conducing the experiments depended on the nature of the measured condition and ongoing activity. To be specific, tangible and attention conditions were conducted during leisure time, while demand condition appeared at teaching time.

The data collected by Bloom et al. (2011), consists of frequencies of the CBs in both experiment and control conditions. A bar graph was applied to represent the data. The bar graph consisted of the same number of clusters and each cluster contained two bars – one for experiment condition and another for control condition. If the frequency was higher in the experiment condition, the manipulated variable was likely to be maintained for the CBs. Trial-based functional analysis had some advantages. First, manipulating in a naturalist setting made it adaptive, when the resources to arrange a standard FA was not available (Bloom et al., 2011). Second, instead of creating a context similar with participants’ living circumstance, this design allowed FA to be conducted in naturalist settings (e.g., classroom). Thus, behaviours observed, tended to be more representative and the analysis result would be more reliable (Lambert & Bloom, 2012). Third, the immediacy transfer between experiment and control trials eliminated the confounding effect from the differences (even slight) between the settings of different sessions (Bloom et al., 2011) – confounding variables only need to be controlled for a brief period.

Bloom et al. (2011) also argued that, trial-based functional analysis should be used with an extra concern. First, in standard FBA, participants’ motivation of gaining reinforcer can be modified by establishing operation (EO), while trial-based design relied heavily on the occasional antecedents. Such background variability made it impossible to repeat a session for checking the stability of the results. Second, for practical constrains, it was hard to find opportunities for long-lasting sessions, during the daily life of individuals. Thus, most trial-based design had sessions varying from 1 to 2 minutes (e.g., Bloom et al., 2011; Lambert & Bloom, 2012; LeJeune et al., 2018). As a result, the participants would have limited access to EO and reinforcing consequences so that pre-session events were more likely to influence the results.

Plan

Preparation Needed before FBA. First, indirect assessment such as interviews, checklists, and questionnaires will be conducted as a starting point to learn the possible reinforcers to Sam’s different challenging behaviours. Second, Sam’s likability towards different tangibles will be measured by preference assessment and placed in order. Third, skill assessment will be used to determine the difficulty level of different tasks for Sam. Fourth, experimenter should exclude the possibility that the CB is a form of skill deficit – if the behaviour frequency does not increase after reinforcing, under the absence of teaching, it would be considered as a skill deficit rather than a CB (Rue, 2011).

Experimental design. The analysis of this study will be based on multielement design to avoid the influence of confounding variables by conducting more than one session of each condition. Based on previous numerous researchers, (e.g., Iwata et al., 1982; Iwata & Dozier, Diller, Barry, & Gelino, 2016), the number of sessions being conducted for each condition varied from three to eight. As suggested by these examples, participants’ behaviours were shown to be distinguished across three/more sessions. Thus, four sessions are determined to be conducted for each condition and therefore the total number of sessions will be 16. Four strings of difference sessions will be conducted, and the order of each strings will be randomly determined.

Settings. According to baseline data collection, there were two settings, where Sam was observed: bedroom and teaching room. In order to investigate whether the setting difference is a discriminative stimulus (SD) of his CBs, all the following four conditions will be repeated in both settings. Based on the baseline data collection (TZ870 Assignment 1), Sam’s CBs can be clearly observed in five-minute sessions. Therefore, each session will last for five minutes. Moreover, before experiment sessions (tangible, attention, and demand), Sam will be deprived from related reinforcers for enough time to reach a deprivation situation.

Tangible Condition. Based on the baseline observation, Sam will be seated .5 meter away from and face to the experimenter. Tangible items (e.g., toy train) will be held by the experimenter, but can be seen by Sam to keep him motivated to engage in CBs. Reinforcers/stimuli will be available right before each session and removed, when the session starts but still visible (Vollmer, Marcus, Ringdahl, & Roane, 1995). The experimenter will not have communication with him during the whole session (e.g., eye contact, verbal communication) and only the target CBs will be responded. For example, verbal use (e.g., saying “back”) will be ignored, and if Sam grabs by himself, he will be physically blocked. Based on the baseline observation, Sam will be provided with a five-second access to the tangible item once the target CB occurs. Tangible then will be grabbed back by the experimenter.

Attention Condition. Sam will be provided with free access to his liking items during this condition. The experimenter will sit 2-2.5 meters away from him (based on baseline data) and pretend doing something (e.g., reading a book) to suggest his attention is not available. It is worthy, to note that, the items provided to him should not be extremely attractive to him (e.g., ipad) in which situation attention is no more motivation of engaging in challenging behaviours. When Sam engages in target behaviour, the experimenter will verbally reprimand him, pairing with short and non-punishing physical contact (e.g., hand on shoulder) for three seconds. Other ways of requiring attention will be ignored or physically blocked.

Demand Condition. Sam will be asked to do a task (e.g., counting the number of some item) for more than one trials in this session. The experimenter will first verbally ask him to do the task. If Sam failed to do it, the experimenter then will model and ask him to do it again. If the modelling is not enough, physical prompt will be used (e.g., holding his hand to point at each item and count). All the successful complement of the task will be reinforced by verbal approval (e.g., “well done!”) no matter it is finished by Sam himself or by prompt. Experimenters should pay extra attention when selecting the task. It should be a task comparatively difficult to Sam to keep his motivation of escaping. At the same time, it should not be excessive difficult which he cannot gain success instances and refuse to engage in. The occurrence of challenging behaviour will be immediately followed termination of the demand for 10 seconds. Other ways of escaping from the task demand will be ignored or even physically blocked.

Unstructured Play Condition. During this condition, Sam will have free access to all his tangible items and contingent attention. The experimenter will play his liking items with him in the way he likes/requires, respond to his sharing language, and provide help when Sam has any difficulty. All challenging behaviour will be ignored. This is a control condition, in which challenging behaviour is expected to occur in a low rate, since all possible motivations of engagement are satisfied. If challenging behaviour still occur at a high rate in this condition, experimenters may concern it as automatically reinforced.

Inter-Observer Agreement (IOA).

In order to check the quality of behaviour definitions, one extra observer will be introduced. The extra observer will record the CBs independently and no communication should occur before the data collection procedure finished. He/she will observe all the experiment and control sessions and record CBs based on the provided definition. Total IOA will be calculated and if the definitions are clear enough, the IOAs are expected to be greater than 80%.

References

Bloom, S. E., Iwata, B. A., Fritz, J. N., Roscoe, E. M., & Carreau, A. B. (2011). Classroom application of a trial-based functional analysis. Journal of Applied Behavior Analysis, 44(1), 19–31. doi: 10.1901/jaba.2011.44-19.

Davis, T. N., Durand, S., Bankhead, J., Strickland, E., Blenden, K., Dacus, S., & Machalicek, W. (2013). Brief report: Latency functional analysis of elopement. Behavioral Interventions, 28(3), 251–259. doi: 10.1002/bin.1363.

Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K. E., & Richman, G. S. (1982). Toward a functional analysis of self-injury. Analysis and Intervention in Developmental Disabilities, 2(1), 3-20. doi: 10.1016/0270-4684(82)90003.

Iwata, B. A., & Dozier, C. L. (2008). Clinical application of functional analysis methodology. Behavior Analysis in Practice, 1(1), 3–9. doi: 10.1007/BF03391714.

Kazdin, A. (1982). Single-case Designs: Methods for Clinical and Applied Settings. New York: NY: Oxford University Press.

Lambert, J. M., Bloom, S. E., & Irvin, J. (2012). Trial-based functional analysis and functional communication training in an early childhood setting. Journal of Applied Behavior Analysis, 45(3), 579–584. doi: 10.1901/jaba.2012.45-579.

LeJeune, L. M., Lambert, J. M., Lemons, C. J., Mottern, R. E., & Wisniewski, B. T. (2018). Teacher-conducted trial-based functional analysis and treatment of multiply controlled challenging behavior. Behavior Analysis: Research and Practice. Advance online publication, doi: 10.1037/bar0000128.

Northup, J., Wacker, D., Sasso, G., Steege, M., Cigrand, K., Cook, J., & Deraad, A. (1991). A brief functional analysis of aggressive and alternative behavior in an outclinic setting. Journal of Applied Behavior Analysis, 24(3), 509–522. doi: 10.1901/jaba.1991.24-509.

Thomason-Sassi, J. L., Iwata, B. A., Neidert, P. L., & Roscoe, E. M. (2011). Response latency as an index of response strength during functional analyses of problem behavior. Journal of Applied Behavior Analysis, 44(1), 51–67.

Vollmer, T. R., Marcus, B. A., Ringdahl, J. E., & Roane, H. S. (1995). Progressing from brief assessments to extended experimental analyses in the evaluation of aberrant behavior. Journal of Applied Behavior Analysis, 28(4), 561–576. doi: 10.1901/jaba.1995.28-561.