The Potential Evocative Effect of Leisure Item Omission in the Attention and Escape Conditions of a Functional Analysis

Abstract

The potential suppressive effects of leisure items in the attention condition of functional analyses have been documented (Ringdahl et al., 2002; Roscoe et al., 2008). However, the potential evocative effects of their absence have not been as thoroughly explored. The current study evaluated the impact of the omission of leisure items in a standard attention or escape condition of a functional analysis. Two participants exhibited leisure item requests and target behavior when the programmed establishing operation for attention or escape was absent during an initial functional analysis of their target behavior. This pattern of responding led to a modification of conditions in which preferred leisure items were included in the attention or escape condition for both participants, resulting in the absence of target behavior. A follow-up component analysis verified that the omission of leisure items and not the programmed establishing operation were responsible for the higher levels of responding in the respective test conditions. Results are discussed in the context of decision-making processes in functional analysis design; a recommendation is made to adopt routine measures of requests and programmed-establishing-operation-absent behavior.

A multielement functional analysis based on the procedures by Iwata et al. (1982/1994) examines the level of challenging behavior in isolated test conditions compared to an omnibus control condition (Jessel et al., 2020). Although this procedure often results in clear outcomes, some functional analyses designed in this manner are inconclusive (Hagopian et al., 2013; Melanson & Fahmie, 2023; Tiger et al., 2009). If differentiation is not initially observed using isolated test conditions, researchers have recommended combining contingencies (e.g., Schlichenmeyer et al., 2013) or adjusting conditions by including or removing leisure items (e.g., Ringdahl et al., 2002) to verify function.

Several studies have examined how the presence or absence of leisure items affects functional analysis outcomes. Call et al. (2005) found that one participant engaged in near-zero rates of challenging behavior during a demand condition that included dissembling a marble game (a highly preferred [HP] leisure item). However, in a separate demand condition absent of the HP leisure item, challenging behavior occurred at high, differentiated rates. Similarly, Call and Lomas Mevers (2014) compared three functional analysis demand condition variations. In the first two conditions (HP leisure item absent; HP leisure item present, but out of reach), only a break from demands was delivered for challenging behavior. In the third condition, an HP leisure item was delivered with a break from demands. The participant engaged in differentiated rates of challenging behavior in all three conditions, but challenging behavior persisted during breaks from demands (e.g., when the programmed establishing operation was absent) for the two conditions in which the HP leisure item was absent or out of reach and did not persist in the absence of the programmed establishing operation in the condition in which the HP leisure item was present during breaks.

Likewise, Ringdahl et al. (2002) and Roscoe et al. (2008) explored the impact of leisure items on responding in the attention condition of a functional analysis. Ringdahl et al. included two separate attention conditions (i.e., attention-only; attention-enriched) for one participant and a communication response with a microswitch for another. In both conditions, target behavior produced access to therapist attention; however, in the attention-enriched condition, the participant received noncontingent access to preferred leisure items throughout the session as well. These same items were absent in the attention-only condition. Responding occurred at differentially higher rates for both participants under the attention-only condition. Roscoe et al. extended these findings by manipulating the presence and preference for leisure items included in the attention condition. Through a series of component analyses, Roscoe et al. showed that challenging behavior was suppressed when leisure items were present but not absent, and HP leisure items produced this suppression to a greater extent than low-preferred leisure items. Collectively, these researchers warned that including HP leisure items in the attention condition may have masked an attention function through a process of reinforcer competition.

Thus, both the presence and absence of leisure items have been shown to influence responding in various functional analysis test conditions by either suppressing or evoking behavior (respectively) in a manner that may obscure clear outcomes. Past research has provided a helpful framework to analyze the impact of leisure items on functional analysis outcomes. In the current study, we sought to demonstrate a hypothesis-testing approach for two individuals with challenging behavior presumed to be evoked by the omission of leisure items (an unprogrammed establishing operation) in standard attention and escape conditions. Our hypothesis was guided by our analysis of two measures (programmed-establishing-operation-absent responding, appropriate requests) that are not routinely included in functional analyses. First, we demonstrated a clinically responsive approach to our three dependent measures. Next, we extended the component analyses used by Roscoe et al. (2008) to confirm the hypothesized effect of the absence of leisure items on responding.

Method

Participants, Materials, and Setting

We included two participants who were actively receiving early intervention applied behavior analytic (ABA) services at a clinic in the Midwestern United States. Both participants also actively participated in another study assessing the function of emerging (i.e., low severity) challenging behavior.1 Brooke was an 8-year-old white female diagnosed with autism spectrum disorder who communicated vocally in short, disfluent sentences. Harper was a 5-year-old Black female diagnosed with autism spectrum disorder who communicated by exchanging pictures of preferred items or activities.

Functional analysis sessions primarily took place in an individualized therapy room containing a cabinet, table, and chairs for the experimenter and participant. Developmentally appropriate work materials and leisure items identified from a preference assessment (see below) were also available depending on the condition. The experimenters conducted Brooke’s sessions in a novel clinical room separate from her daily treatment space. The room was relatively empty, apart from the specific materials (work, toys, etc.) relevant to each functional analysis condition. As requested by her Board-Certified Behavior Analyst (BCBA), the door remained open throughout the assessment. On a few occasions, Brooke walked out of the individualized therapy room and directly to a large table in a common area. When this occurred, the sessions were continued at the large table. No additional people or materials were present at the large table. The experimenters conducted Harper’s sessions in the space where she regularly received therapy. Additional items relevant to her daily treatment goals were in the room but kept out of reach or in closed cabinets, and Harper did not attempt to engage with those items throughout the assessment.

Response Measurement and Interobserver Agreement

For both participants, target behaviors included a combination of nondangerous2 and dangerous topographies. Nondangerous behaviors for both participants included whining/groaning, defined as making cry-noises (e.g., “aww,” “uhh”) without tears, and facial expressions of disgust, defined as frowning, grimacing, or snarling. Additional nondangerous behaviors for Brooke included shaking limbs, defined as rapid, small back and forth or side-to-side movements of limbs, and pushing a person away, defined as pushing a person or person’s body part away from her body at least 2.5 cm. Dangerous behaviors for Brooke included biting self, defined as making forceful contact between the top and bottom teeth and herself, and hitting self, defined as making forceful (greater than 7.5 cm) contact between open palm and herself. Additional nondangerous behaviors for Harper included tapping a person, defined as making repetitive limb/body contact with another person from a distance of 7.5 cm or less, stomping feet, defined as bringing a foot off the ground and then placing it back on the ground with force, and loud vocalizations, defined as making a vocalization above conversational level. Dangerous behavior for Harper included throwing objects, defined as clasping an object with her hands and then projecting away from her body (greater than 7.5 cm).

The experimenters calculated the rate of target behavior by dividing their frequency by the total session duration in minutes. Data were further analyzed during periods when the programmed establishing operation was present (e.g., attention withdrawn in the attention condition) and during periods when the programmed establishing operation was absent (e.g., attention available in the attention condition). The rate of programmed-establishing-operation-present behavior was calculated by dividing the frequency of target behavior during programmed-establishing-operation-present periods by the total duration in minutes from those same periods (Roane et al., 1999). The rate of programmed-establishing-operation-absent behavior was calculated by dividing the frequency of target behavior during programmed-establishing-operation-absent periods by the total duration in minutes from those same periods. The experimenters also collected data on requests for HP and moderately preferred (MP) leisure items for Brooke and Harper, respectively. HP requests were defined as any request for the coloring sheets and utensils for Brooke. MP requests were defined as any request for the bubble toy for Harper. The experimenters reported on the rate of requests, calculated by dividing the frequency of HP (Brooke) or MP (Harper) requests by the total session duration in minutes.

An independent observer collected interobserver agreement (IOA) data from 33.3% and 21.1% of randomly identified sessions during the functional analyses for Brooke and Harper, respectively. An independent observer collected IOA data from 33.3% of randomly identified sessions of the follow-up component analyses for both participants. IOA was calculated by comparing the scored frequency of target behavior in each 10-s interval of a session using the proportional agreement method. Intervals received a score of 1 if the frequency was identical; intervals with partial or no agreement were scored by dividing the smaller frequency by the larger frequency of responses within the interval. The experimenters divided the sum of the intervals by the total number of intervals and multiplied by 100 to obtain a percentage. IOA data were calculated at 94.3% (range, 80–100%) for Brooke’s functional analysis, 94.6% (range, 91.7–100%) for Brooke’s follow-up component analysis, 98.9% (range, 95.6–100%) for Harper’s functional analysis, and 98.8% (range, 97.8–100%) for Harper’s follow-up component analysis.

Procedure

Preference Assessment

Both participants completed a single-session multiple-stimulus without replacement (MSWO) preference assessment to identify MP and HP items to be used throughout the functional analysis (Conine et al., 2021). Upon entering the clinical room, the participants sat at the table across from the experimenter. The experimenter presented five leisure items, identified by the participant’s BCBA, one at a time, allowing 30 s of presession access to each item. Subsequently, the experimenter placed the items in a random array on the table and instructed the participant to “pick one.” After each selection, the experimenter allowed 30 s access to the item while simultaneously shifting the remaining items in the array one position to the left, with the item placed furthest left in the previous trial now positioned at the far-right end of the array.

After 30 s had passed, the experimenter regained control of the selected item, removed it from the array, and instructed the participant to “pick one” again. The assessment continued until each item was selected (Brooke) or 10 s had passed without a selection response (Harper; trial 4), whichever came first. The experimenters considered the item selected first as HP and the item selected third as MP. For Brooke, preference assessment results identified coloring sheets and utensils as the HP item and the Minnie Mouse figurines as the MP item. For Harper, preference assessment results identified yo-yos as the HP item and a bubble toy operated by the experimenter as the MP item.

Functional Analysis

Initial Procedures

 Brooke and Harper’s target behavior was initially analyzed in a multielement functional analysis (Iwata et al., 1982/1994). Conditions included tests for attention, escape, and tangible items, as well as an omnibus (play) control. Each session lasted five min (Griffith et al., 2021), was rotated in a fixed order (Hammond et al., 2013), and continued until differentiated responding occurred. Before initiating a session, the experimenter ensured the participant was calm (i.e., positive or neutral affect; no target behavior) for at least one min. In addition to target nondangerous and dangerous behavior, the experimenter also provided the condition-specific reinforcer for any appropriate requests relevant to the condition (requests for leisure items in the tangible condition; requests for attention in the attention condition; requests for a break or help in the escape condition) based on the programmed establishing operation (LaRue et al., 2011). Although this represented a departure from the standard procedures for a functional analysis of challenging behavior, requests were reinforced in this study based on the protocol used by Melanson et al. (2025), which sought to measure the correspondence between the outcomes of sensitivity tests (wherein reinforcement of appropriate requests is standard; see Fahmie et al., 2020) and functional analyses for both target behavior and appropriate requests. Table 1 summarizes the reinforcers scheduled for target behavior and leisure item requests during the functional analysis.

Table 1.

Conditions when target behavior and HP/MP requests were reinforced

		Brooke		Harper
		Target behavior	HP requests	Target behavior	MP requests
Initial FA	Attention	X		X
	Play		X		X
	Escape	X		X
	Tangible	X	X	X	X
Component Analysis	Attention No HP	X		–	–
	Attention HP	X	X	–	–
	Play No HP			–	–
	Play		X		X
	Escape MP	–	–	X	X
	Escape No MP	–	–	X
	Play No MP	–	–

HP highly preferred; MP moderately preferred; FA functional analysis; X indicates a reinforcement contingency; blank cells indicate no reinforcement contingency; dashes indicate that the condition was not included in the analysis

Attention

Prior to the start of the attention condition, the experimenter provided 30 to 60 s of noncontingent attention and ensured the participant was engaged with the experimenter’s attention before saying, “I have some work to do,” while redirecting their attention elsewhere. The participant received noncontingent access to their MP leisure item throughout the session. Because Harper’s MP leisure item required assistance from the experimenter, the experimenter operated the bubble toy continuously and continued to redirect their attention elsewhere. Contingent on target behavior or an appropriate request for attention, the experimenter turned toward the participant and provided attention for 20 s before redirecting their attention again.

Play (Control) 

During the play condition, the experimenter provided continuous access to attention and HP and MP leisure items. The experimenter cooperated with all reasonable requests, did not present instructions, and did not differentially respond to target behavior.

Escape 

During the escape condition, the experimenter implemented two-step prompting (verbal, model) to present continuous task instructions. Contingent on target behavior or an appropriate request for escape, the experimenter removed the task materials and allowed the participant a break for 20 s before presenting a new instruction. General praise was provided contingent on cooperation with task instructions.

Tangible 

Prior to the start of the tangible condition, the experimenter provided access to HP and MP leisure items for 30 to 60 s and ensured the participant was engaged with the items before withdrawing the items and turning away. Contingent on target behavior or an appropriate request for leisure items, the experimenter provided access to the withdrawn items for 20 s before withdrawing them again.

Modified Procedures

 For Brooke, requests for HP items (not present) occurred throughout the first attention session, and target behavior persisted throughout programmed-establishing-operation-absent periods when attention was continuously provided. Similarly, for Harper, requests for MP items (not present) occurred throughout the first and second escape sessions, and target behavior persisted throughout programmed-establishing-operation-absent periods when escape from task instructions was provided. Because of these response patterns, the experimenters hypothesized that the target behavior might have been evoked by the absence of HP items for Brooke and MP items for Harper. In an attempt to better isolate the programmed-establishing-operation in each test condition, HP items were included in the attention and escape conditions for Brooke starting in session 5, and MP items were included in the escape condition for Harper starting in session 11 (Note: MP items were already present in the attention condition). Brooke received noncontingent access to the coloring pages and utensils while attention was withdrawn in the attention condition or tasks were presented in the escape condition. Harper received noncontingent access to the bubble toy, and requests for the experimenter to operate the bubble toy (which was her preferred way to engage with the item) were honored in the escape condition.

Follow-up Component Analysis

After completing the modified functional analyses, the experimenters further tested the impact of the presence or absence of leisure items on responding by comparing four unique conditions (Roscoe et al., 2008). Table 1 summarizes the reinforcers scheduled for target behavior and leisure item requests in the follow-up analysis.

Attention with HP Items (Brooke) 

The attention with HP items condition was identical to sessions 5 and 9 of Brooke’s functional analysis. Specifically, Brooke had access to her coloring sheets and utensils throughout the duration of the session, including when attention was withdrawn. Contingent on target behavior, the experimenter provided 20 s of access to attention before redirecting attention to reset the programmed establishing operation.

Attention with No HP Items (Brooke) 

The attention with no HP items condition was identical to session 1 of Brooke’s functional analysis. Specifically, Brooke received noncontingent access to Minnie Mouse figurines (an MP item) throughout the duration of the session, including when attention was withdrawn. Contingent on target behavior only, the experimenter provided 20 s of access to attention before withdrawing attention to reset the programmed establishing operation. HP items were not present in this condition, and requests for HP items were not reinforced.

Escape with MP Item (Harper) 

The escape with MP item condition was identical to sessions 11 and 15 of Harper’s functional analysis. Specifically, Harper received noncontingent access to the bubble toy throughout the duration of the session, and requests for the experimenter to operate the bubble toy were honored. Task instructions were provided continuously. Contingent on target behavior, the experimenter provided 20 s of escape from task instructions.

Escape with No MP item (Harper) 

The escape with no MP item condition was identical to sessions 3 and 7 of Harper’s functional analysis. Specifically, the experimenter provided continuous task instructions, no additional leisure items were present in the room, and the experimenter did not attend to requests for the bubble toy. Contingent on any instance of the target behavior only, Harper received a 20-s break from instructions. The MP item was not present in this condition.

Play (Control; Brooke and Harper)

The play condition was identical to that of the initial functional analysis. Specifically, Harper and Brooke received noncontingent access to attention, HP and MP items, and no task instructions were presented. All reasonable requests were honored, and the experimenter did not attend to target behavior.

Play with No HP Items (Brooke)

The play with no HP items condition was identical to the play (control) condition, except only the MP items, not the HP items, were present, and no reinforcement was programmed for HP item requests.

Play with No MP item (Harper)

 The play with no MP item condition was identical to the play (control) condition, except only the HP items, not the MP item, were present, and no reinforcement was programmed for MP item requests.

Results

Figure 1 (left panel) shows the results of Brooke’s functional analysis. Target nondangerous and dangerous behavior did not occur in any play session. During the first attention session, Brooke engaged in an elevated rate of target behavior, both during programmed-establishing-operation-present (1.4 responses per minute [RPM]) and programmed-establishing-operation-absent periods (2.4 RPM). Brooke also engaged in a high rate of requests for the HP items (1.6 RPM), which were not present in this particular session. After observing this response pattern, the experimenters added HP items to the attention and escape conditions starting with session 5. For the final two attention sessions, target behavior did not occur. Lastly, target behavior occurred at a high level, and there was an increasing trend during the escape and tangible conditions. As a result, the experimenters concluded that Brooke’s target behavior was maintained by escape from instructions and access to tangibles.

Fig. 1.

Functional analysis and follow-up component analysis results for Brooke. Note. The top panel depicts the rate of target behavior, the middle panel depicts the rate of requests for HP items, and the bottom panel depicts the rate of programmed-establishing-operation-absent target behavior during Brooke’s functional analysis (left panel) and follow-up component analysis (right panel). HP highly preferred, EO establishing operation

Given the limited opportunities to replicate effects in the initial functional analysis, the extent to which the absence of leisure items in the first attention session impacted Brooke’s responding was still unclear. Figure 1 (right panel) shows the results of Brooke’s follow-up component analysis. Brooke did not engage in target behavior when the HP items were present, regardless of the presence of the experimenter’s attention. Conversely, target behavior occurred at moderate to high rates when the HP items were absent in both the attention with no HP items (M = 1.2 RPM) and play with no HP items (M = 0.5 RPM) conditions. Moreover, target behavior occurred (M = 0.5 RPM in each condition) during programmed-establishing-operation-absent periods during these same conditions. As a result, the experimenters concluded that Brooke’s target behavior was impacted by the absence of HP items and not by access to attention.

Figure 2 (left panel) shows the results of Harper’s functional analysis. Target behavior did not occur in the play condition but did occur at a high and stable rate during the attention condition. During the tangible condition, target behavior occurred at low rates for the first two sessions, elevated rates in the third session, and zero rates in the final three sessions. Recall that requests for the MP item were reinforced during the tangible condition, per the research protocol (Melanson et al., 2025). When considered together, requests and target behavior occurred at a steady rate in the tangible condition for Harper and were inversely correlated, suggesting sensitivity to tangibles as a reinforcer for both response topographies. Finally, during the escape condition, Harper engaged in low to moderate rates of target behavior in the first two sessions, both during programmed-establishing-operation-present (M = 0.4 RPM) and programmed-establishing-operation-absent (M = 0.5 RPM) periods. Harper also engaged in a high rate of requests for the absent MP item (1.6 RPM) in the second escape session. After observing this response pattern, the experimenters added the MP item to the escape condition starting in session 11. For the final two escape sessions, target behavior did not occur. As a result, the experimenters concluded that Harper’s target behavior was maintained by access to attention, and her target behavior and requests (in combination) were maintained by access to leisure items.

Fig. 2.

Functional analysis and follow-up component analysis results for Harper. Note. The top panel depicts the rate of target behavior, the middle panel depicts the rate of requests for MP items, and the bottom panel depicts the rate of programmed-establishing-operation-absent target behavior during Harper’s functional analysis (left panel) and follow-up component analysis (right panel). MP moderately preferred, EO establishing operation

The impact of the absence of the MP leisure item in the first two escape sessions was further evaluated in Harper’s follow-up component analysis. Figure 2 (right panel) shows that Harper did not engage in target behavior when the MP item was present, regardless of the presence of task instructions. Conversely, target behavior occurred at moderate to high rates when the MP item was removed in both the escape with no MP item (M = 0.9 RPM) and play with no MP item (M = 1.3 RPM) conditions. Moreover, target behavior occurred during programmed-establishing-operation-absent periods (M = 0.3 RPM during the escape with no MP item; M = 1.3 RPM during the play with no MP item). As a result, the experimenters concluded that Harper’s target behavior was impacted by the absence of the MP item and not by escape from task instructions during the escape condition.

Discussion

The current study replicated and extended research on the impact of leisure items on functional analysis outcomes. We demonstrated that response patterns in functional analyses may have been impacted by the presence or absence of preferred leisure items for two individuals with nondangerous and dangerous behavior partially maintained by tangible reinforcement. These results underscore the importance of a hypothesis-driven approach to functional analysis. That is, we modified our procedures when initial outcomes suggested the presence of an unprogrammed establishing operation and potential false positive outcome. We then confirmed our hypothesis in a component analysis that isolated the impact of leisure items on responding in the relevant test condition.

Critical to our initial hypothesis was the inclusion of two measures that are not routinely included in functional analyses. First, we collected data on the rate of requests for leisure items and showed high rates in sessions in which those leisure items were absent. Second, we analyzed data on target behavior when the programmed establishing operation was present versus absent. When rates of target behavior were higher during programmed-establishing-operation-absent periods than during programmed-establishing-operation-present periods, it suggested the presence of unprogrammed establishing operations. We encourage the regular adoption of these supplementary measures. These data are rarely reported in published functional analyses; interestingly, though, researchers are beginning to evaluate new functional analysis formats that routinely consider within-session responding and measures other than challenging behavior (e.g., Iovino et al., 2022; Jessel et al., 2023; Reilly et al., 2025).

Several methodological considerations should be acknowledged when interpreting our findings. Given our data on Harper’s consistent requests for MP items (and not HP) items, it is possible that our expedited single-session MSWO produced a hierarchy that was only partially accurate. Additional sessions of the MSWO may have revealed more reliable outcomes (Richman et al., 2016) and are encouraged in future research on this topic. More importantly, our functional analysis design did not permit a conclusion about whether our initial functional analyses would have produced false positive outcomes if they had not been modified. That is, it is not uncommon for target behavior occurring in initial sessions of a functional analysis to eventually decrease with more contact with the programmed contingencies. Because we modified conditions quickly during the functional analysis, we could not test whether responding assumed to be impacted by omitted leisure items would have eventually decreased in attention or escape sessions via a process of extinction. Our decision to intervene before stability was achieved was based on our supplementary measures (presence of requests for leisure items and target behavior during programmed-establishing-operation-absent periods) suggestive of an unprogrammed establishing operation. For example, target behavior occurred at nearly double the rate (2.4 RPM in programmed-establishing-operation-absent periods vs. 1.4 RPM in programmed-establishing-operation-present periods) during the first attention session of Brooke’s functional analysis. Thus, our quick modification of the initial functional analysis reduced potentially aversive features (unreinforced requests for items) of our protocol.

Despite this important limitation, data from the present study suggested that the initial sessions of the functional analysis were impacted by the presence of unprogrammed establishing operations, which would have likely necessitated a longer analysis had we not modified the conditions quickly. Many practitioners and researchers have shown interest in maximizing the efficiency of functional analyses in practice (Jessel et al., 2022a, 2022b; Saini et al., 2020). Data from our follow-up analysis showed continued responding (and not extinction) under conditions when leisure items were absent, suggesting that this response pattern may have persisted in the initial functional analysis if we had not made modifications. Both the total number of sessions (12) and the design (multielement) of the follow-up analysis were similar to that of a typical functional analysis, suggesting that these same response patterns could have persisted across the initial functional analysis unless a different design (e.g., reversal) was adopted.

The results of the present study should be considered alongside a few additional limitations. First, requests for the HP (Brooke) and MP (Harper) leisure items were reinforced in some sessions but not in others (see Table 1). The procedural choice to reinforce condition-specific requests may have had unintentional effects on the target behavior of participants when requests were no longer reinforced in sessions in which they were not deemed “condition specific” (e.g., not reinforcing leisure item requests in attention sessions). For example, reinforcement of leisure item requests in the tangible sessions may have strengthened discriminative stimuli and increased an establishing operation for tangibles during the first attention session for Brooke and the second escape session for Harper. Published functional analyses that have reinforced requests only and not challenging behavior have shown relatively low correspondence with functional analyses reinforcing challenging behavior only (Rahaman et al., 2024). However, the effects of reinforcing both requests and target behavior on the outcomes of functional analyses are yet unclear. Future research is warranted to identify the extent to which our outcomes generalize to traditional functional analysis formats in which requests are not reinforced. Additional research is necessary on measures (e.g., indicating responses) that may supplement appropriate requests when a child does not have the communication skills displayed by our participants.

Second, our functional analyses did not include a test for automatic reinforcement, so it remains unclear whether any features of automatic reinforcement influenced our results. Typically, MP leisure items are programmed in the attention condition of a functional analysis to reduce the likelihood of evoking automatically maintained challenging behavior (Roscoe et al., 2008). Given the consistent suppression of target behavior in conditions that included leisure items and the maintenance of responding in at least one social test condition, it seems unlikely that our participants had an undetected automatic reinforcement function. Nevertheless, it is possible that our participants engaged in target behavior either partially or fully maintained by automatic reinforcement and that including leisure items across multiple conditions masked the automatic reinforcement function. This is particularly likely if the leisure items we included competed with the sensory consequences of the target behavior (Hagopian et al., 2017). One barrier that researchers will encounter in responding to this limitation will be how to interpret steady rates of target behavior in a no-interaction condition that does not include leisure items, particularly if requests for leisure items persist in that condition as well. That is, the absence of leisure items in a typical no-interaction condition may also produce false positive outcomes if an unprogrammed establishing operation for tangibles is present. Perhaps arranging tests for automatic reinforcement immediately, prior to introducing leisure items in preference assessments or other rapport-building activities, may decrease the likelihood of challenging behavior evoked by the absence of leisure items.

Third, although our analysis incorporated valuable additional measures that informed our decision-making, we did not record item engagement during each condition nor present data on follow-up treatment procedures.3 Researchers interested in extending this work should consider measuring leisure item engagement in each condition to help identify if requests for leisure items correspond with their engagement. Longer durations of engagement with leisure items could help support the hypothesis that the presence or absence of leisure items impacts outcomes. In addition, data showing a successful follow-up treatment for participants who engage in similar response patterns could also help support the validity of the modified functional analysis.

The present study documented the evocative effect of the absence of leisure items in standard attention and escape conditions of a functional analysis, supporting a hypothesis-driven approach to functional analysis modifications. Our data support the use of supplementary measures of within-session responding and requests to drive clinically responsive practice. We encourage future researchers and clinicians to consider adopting these two measures, as well as our follow-up component analyses, when relevant, to inform their decision-making process.

Author Contributions

Isaac J. Melanson – Conceptualization, methodology, validation, formal analysis, investigation, resources, writing – original draft, writing – review and editing, visualization, project administration.

Tara A. Fahmie – Conceptualization, methodology, validation, formal analysis, writing – original draft, writing – reviewing and editing, supervision, project administration, funding acquisition.

Kendall M. Kastner – Formal analysis, resources, writing – original draft, writing – reviewing and editing.

Funding

Autism Speaks, Grant/Award Number: 13411.

Data Availability

Additional data to support this study are available upon request.

Declarations

Ethics Approval

The authors conducted this study in accordance with the standards approved by the University of Nebraska Medical Center’s Institutional Review Board.

Informed Consent

Before the study, the authors obtained written informed consent from each participant’s caregiver.

Conflict of Interest

The authors have no conflicts of interest to declare.

Clinical Trial Number

Not applicable.