Marissa E. Kamlowsky | Claudia L. Dozier | Stacha C. Leslie | Ky C. Kanaman | Sara C. Diaz de Villegas
Department of Applied Behavioral Science, University of Kansas, Lawrence, KS, USA

Correspondence
Claudia L. Dozier, Department of Applied Behavioral Science, University of Kansas, Lawrence, KS, 66045, USA.
Email: cdozier@ku.edu

Editor-in-Chief: John Borrero
Handling Editor: Craig Strohmeier

© 2024 Society for the Experimental Analysis of Behavior (SEAB).

Abstract

We replicated and extended Kanaman et al. (2022) by comparing outcomes of solitary (leisure items only), social (leisure items with social interaction), and combined (leisure items alone and leisure items with social interaction) stimulus preference assessments to determine the extent to which the inclusion of social interaction influenced the outcomes of preference assessments for five children with autism. We then conducted reinforcer assessments to determine the reinforcing efficacy of high- and low-preferred leisure items when presented with and without social interaction. The results showed that both high- and low-preferred items functioned as reinforcers to varying degrees for all participants and the inclusion of social interaction increased the reinforcing efficacy of some items for all participants. Additionally, the results showed that combined preference assessments predicted reinforcer assessment outcomes for two of five participants but produced false-negative outcomes for three participants. Clinical implications and directions for future research are discussed.

Keywords: preference assessment, reinforcer assessment, reinforcer efficacy, social interaction

 

Stimulus preference assessments (SPAs) are commonly used to determine preferred items and activities that may be used as reinforcers to increase target behavior during clinical programming for individuals with autism and related disabilities (Hagopian et al., 2004). Numerous methods for conducting SPAs have been developed over the past several decades (see Saini et al., 2021, for a review), and most research has demonstrated that preference is a relatively reliable predictor of reinforcer efficacy (DeLeon et al., 2009; Piazza et al., 1996). The extent to which a stimulus functions as a reinforcer may be determined using a reinforcer assessment (Pace et al., 1985; Piazza et al., 1996), and researchers have found that high-preferred (HP) stimuli typically function as more effective reinforcers relative to low-preferred (LP) stimuli. That is, HP stimuli result in higher rates of responding, more persistent responding, or faster skill acquisition relative to LP stimuli (Penrod et al., 2008; Taravella et al., 2000). However, research has also shown that LP stimuli can function as reinforcers for some individuals or behaviors (N. M. Goldberg et al., 2023; Graff et al., 2006), particularly under singleoperant reinforcer arrangements (Roscoe et al., 1999). Given the clinical utility of both HP and LP stimuli as reinforcers (e.g., increased flexibility in the use of reinforcers of different qualities, decreased likelihood of satiation; Paden & Kodak, 2015), empirical investigations surrounding variables that influence preference and reinforcer efficacy are warranted.

Several variables may influence the outcomes of SPAs and reinforcer assessments. For example, motivating operations (i.e., conditions that momentarily alter the value of stimuli) such as the degree of deprivation or satiation associated with stimuli included, may influence SPA outcomes (Hanley et al., 2006). Hanley et al. (2006) measured shifts in preference with 10 adults with developmental delays following motivating-operation manipulations by providing extended daily access to HP items (a preference-weakening manipulation) and pairing access to LP items with social and edible reinforcers (a preference-strengthening manipulation). The researchers found that shifts in preference could be imposed by directly manipulating the motivating operations; however, the overall outcomes aligned with research on preference stability that suggests that hierarchies of preference remain generally stable over time (Carr et al., 2000; Hanley et al., 2006; MacNaul et al., 2021).

Another variable that may influence outcomes is the magnitude or duration of stimulus access. For example, Paden and Kodak (2015) evaluated the effects of varying reinforcement magnitude on skill acquisition outcomes and preference for reinforcers with four children with autism. All participants preferred the large-magnitude reinforcers; however, both large- and small-magnitude reinforcers produced rapid skill mastery. In another study, Jones et al. (2014) compared the preference stability of 11 typically developing children when the duration of access to HP items varied from 30 s to 5 min. Preferences remained generally stable across durations; however, for some participants preference for certain leisure items (e.g., videos) was influenced by the duration of access.

In addition to magnitude and duration, the types of items (i.e., stimulus categories) that are used may also influence outcomes. For example, Conine and Vollmer (2018) compared relative preferences for edible and leisure items in combined SPAs (i.e., SPAs involving stimuli from different categories) for 26 children with autism; they found that edible items were ranked higher than leisure items for most participants. However, the researchers observed less displacement of items (i.e., changes in rankings as a result of inclusion or exclusion of a stimulus category) than was observed in previous research (e.g., DeLeon et al., 1997; Fahmie et al., 2015). Specifically, Conine and Vollmer found that multiple leisure items outranked edible items during combined SPAs. The researchers speculated that leisure items may have outranked edible items because screen-based media were included (e.g., tablets, computers). This finding suggests that the type of stimulus category (e.g., edible, leisure) influences outcomes and that the inclusion of screen-based media within leisure-item SPAs may also influence hierarchies.

Along with edible, leisure, and screen-based stimulus categories, researchers have begun to evaluate the effects of social stimuli (e.g., social interaction, leisure items or activities presented with social interaction) on SPA outcomes. For example, N. M. Goldberg et al. (2023) demonstrated that when social interaction was evaluated as a separate stimulus category in a combined SPA, displacement of preferences by stimulus category occurred for three of five participants. M. C. Goldberg et al. (2017) evaluated preference for and reinforcing efficacy of activities that were presented with and without social interaction provided by participants’ mothers for 21 boys with autism. Results showed that participants preferred certain activities presented with social interaction at levels similar to those of typically developing peers. However, interpretation of these results is limited in that participants’ mothers provided all social interaction and the extent to which individuals with autism would prefer social activities when provided by other adults (e.g., staff or therapists) is unknown.

More recently, Kanaman et al. (2022) evaluated the effects of social interaction provided by an experimenter (i.e., classroom teacher) during leisure-item presentation. Specifically, the researchers compared preference for and reinforcer efficacy of leisure items that were presented with and without social interaction for 33 typically developing children during solitary (leisure items alone), social (leisure items with social interaction), and combined (leisure items alone and leisure items with social interaction) paired-stimulus preference assessments (PSPAs; Fisher et al., 1992). The researchers then conducted a concurrent-operants reinforcer assessment to determine the extent to which HP-leisure items identified from combined PSPAs predicted the results of reinforcer evaluations when presented with and without social interaction. The results of the solitary and social PSPAs indicated relatively stable preferences across participants; however, when leisure items were presented with and without social interaction in the same combined PSPA, the ranked order of the participants’ preferences shifted. Specifically, for most participants, preference and reinforcing efficacy of leisure items increased when items were presented with social interaction indicating that solitary and social stimuli are qualitatively different. However, there are some limitations worth noting.

First, Kanaman et al. (2022) conducted a concurrent-operants reinforcer assessment in which only the relative reinforcing efficacy of stimuli was assessed. In a concurrent-operants reinforcer assessment, participants may respond exclusively to a single task and the reinforcing effects of other stimuli (e.g., LP stimuli) may not be determined (Francisco et al., 2008). Alternatively, the use of a single-operant reinforcer assessment would allow for evaluation of the absolute reinforcing efficacy of each stimulus, which may provide information about the potency of LP stimuli (Francisco et al., 2008; Roscoe et al., 1999). Determining the reinforcing efficacy of LP items is important given that combined SPAs may be particularly susceptible to displacement effects, which may produce false-negative outcomes (DeLeon et al., 1997; N. M. Goldberg et al., 2023). That is, combined SPAs may suggest that specific stimuli or categories of stimuli are LP and less reinforcing. However, those LP stimuli may function as high-quality reinforcers, which would allow for increased flexibility and variety of stimuli used in programming (N. M. Goldberg et al., 2023). For example, after conducting combined PSPAs including edible, leisure, and social-interaction stimuli, N. M. Goldberg et al. (2023) conducted single-category reinforcer assessments and found that stimuli that were identified as LP produced levels of responding that were equal to those observed in the HP condition for two of five participants. Further, N. M. Goldberg et al. found that social interaction alone functioned as a reinforcer for four of five participants despite combined SPA outcomes indicating that social interaction was LP. These findings suggest the importance of evaluating LP stimuli that are identified from combined SPAs in subsequent reinforcer assessments. Second, Kanaman et al. used a relatively dense schedule of reinforcement (fixed-ratio [FR] 1 or 6), which did not reveal the extent to which HP stimuli would maintain reinforcing efficacy under progressively thinner schedules. Alternatively, using a progressive-ratio (PR) schedule of reinforcement in which the response requirement progressively increases within a single session may better highlight reinforcer efficacy (Francisco et al., 2008; Hodos, 1961; Roane, 2008; Roane et al., 2001).

Although results of M. C. Goldberg et al. (2017), N. M. Goldberg et al. (2023), and Kanaman et al. (2022) provide information about the influence of social interaction on displacement of preferences and reinforcing efficacy of leisure items and activities, the generality of these findings is somewhat limited. Specifically, all participants included in Kanaman et al. were typically developing children and all social interaction provided by M. C. Goldberg et al. (2017) was delivered by the participants’ mothers. Thus, the extent to which similar findings would be obtained with individuals who commonly participate in SPAs (e.g., children with autism) and those who program reinforcer arrangements (e.g., therapists or teachers) remains unknown.

Furthermore, it is especially important to extend these findings to children with autism given that they may exhibit difficulties in social–emotional reciprocity (e.g., failure to initiate or respond to social interaction) and may show less preference for social stimuli relative to neurotypical peers (American Psychiatric Association, 2013). However, recent research evaluating the value of social and nonsocial activities for individuals with autism suggests that social interactions can be both preferred and reinforcing for this population (e.g., Call et al., 2013; N. M. Goldberg et al., 2023; Gutierrez et al., 2013; Kelly et al., 2014; Morris & Vollmer, 2019; Nuernberger et al., 2012). Despite these findings, it remains possible that clinical providers for children with autism will assume that social interaction is not preferred and may avoid providing interactions during reinforcement periods. Furthermore, understanding the extent to which including social interaction enhances a reinforcer’s value may lead to more efficient programming for skill acquisition and reduction of challenging behavior. For example, if social interaction enhances the reinforcing efficacy of specific leisure items, clinicians may strategically modify differential reinforcement procedures to include social interaction.

Taken together, these findings suggest that social interaction may be preferred and reinforcing for some individuals with autism. However, continued research on preference for and reinforcing efficacy of solitary and social stimuli for children with autism is warranted to more thoroughly program learning opportunities and arrange environments that are both individualized and preferred. Therefore, the purpose of the current study was to replicate previous research by comparing the influence of social interaction on preference for leisure items within solitary (leisure items presented without social interaction), social (leisure items presented with social interaction), and combined (duplicate leisure items presented with and without social interaction in the same assessment) SPAs with children with autism. Additionally, we evaluated the reinforcing efficacy of both HP and LP stimuli presented with and without social interaction in a subsequent reinforcer assessment using a single operant arrangement with a PR schedule for most participants.

Method

Participants, setting, and materials

Five children who were diagnosed with autism were included in this study. All received their diagnoses from professionals who were not affiliated with the university based child development center that they attended, and all participants were enrolled in an early intervention program within the university-based child development center. Josie was a 5-year-old female who communicated vocally using full sentences and whose cultural identities included White and Afghan. Ophelia was a 3-year-old female who communicated vocally using one-to-two-word utterances and whose cultural identities included White and Nigerian. Andrew was a 5-year-old male who communicated vocally using one-to-three-word utterances and whose cultural identities included White and Hispanic. Masie was a 5-year-old female who communicated vocally using one-to-three-word utterances and whose cultural identities included White and Sri Lankan. Spencer was a 4-year-old White male who communicated vocally using one-to-three-word utterances.

Only children who met 80% accuracy in the preassessment (see below), which involved a brief assessment to determine whether they could discriminate the social interaction stimulus from other stimuli, were included; all recruited children met this criterion. Preference assessments for four children were conducted using a PSPA format; however, we conducted a free-operant preference assessment with response restriction (Hanley et al., 2003) for one participant (Spencer) due to the occurrence of challenging behavior when items were removed during previous attempts at the PSPA (see below for modifications to the free-operant assessment). A graduate student who was familiar to each participant (i.e., had known the participant for 6 months or more within the context of educational services, such as the supervisor in the participant’s classroom or adjacent classroom) served as the experimenter for all SPA sessions for a participant. The experimenters conducted all SPAs at a desk or on the floor of a session room at the child development center that contained a table, two chairs, and relevant session materials. Preference-assessment sessions were conducted once per day for a maximum of 30 min per day so as not to disrupt the participants’ early intervention programming until all SPA trials were complete. This period ranged from 1 to 3 weeks for all participants.

Materials for the solitary SPAs included four (Spencer only) or six different leisure items that were reported by caregivers and staff to be preferred for a participant (e.g., iPad, magnet letters, books, play food, dolls). Materials for the social SPAs included the same leisure items that were used in the solitary SPAs and two or four (Spencer only) pictures of the experimenter to be presented during the social SPA trials (i.e., one picture was presented with each leisure item that was presented with social interaction). The pictures were 30  21 cm in size and depicted the experimenter with a pleasant expression (e.g., smiling). Materials for the combined SPAs included duplicate sets of the leisure items that were used in the previous assessments and two pictures of the experimenter such that solitary and social leisure items could be included in the same assessment.

During the reinforcer assessment, session materials for Masie and Andrew included a binder with 20 laminated sheets of paper (22  28 cm). Each laminated paper depicted five squares (5  5 cm) with a single shape inside, which were used for a shape-matching task. A bin of laminated shape cutouts (5  5 cm) was used for the shape-matching task. Each cutout had a small piece of Velcro attached for ease of matching. A large number of laminated cutouts were included to ensure that participants had enough cutouts to match for the duration of the session without requiring the experimenter to reset the materials. For Josie and Ophelia, materials for the reinforcer assessment included the same bin of laminated shapes with corresponding bins for a sorting task. For most reinforcer assessments, we chose tasks that involved discrete responses and were similar to mastered educational tasks that were used in participants’ day-to-day programming. For Spencer, the reinforcer assessment involved separating the session room into three equidistant and concurrently available squares that were outlined on the floor with tape. Each square was approximately 1  1.5 m.

Additional session materials for all participants’ reinforcer assessments included the highest and lowest ranked leisure items from the combined SPA. That is, the leisure items (whether presented with or without social interaction) with the overall highest and lowest selection percentages from the combined SPAs were evaluated with and without social interaction in an HP and LP reinforcer assessment, respectively. If the HP leisure item from the combined SPA was the iPad, the experimenters evaluated both the iPad and the next highest ranked leisure item in separate reinforcer assessments. We chose to evaluate the reinforcing efficacy of both the iPad and the next-highest ranked leisure item in these circumstances based on previous research suggesting that screen-based media can influence preference hierarchies and potentially displace preference for other HP leisure items (Conine & Vollmer, 2018). Additionally, the reinforcer assessments included an alternative item or activity that was provided across all reinforcement sessions to ensure that the participant could engage in something other than the target response. The participant’s staff suggested the alternative item or activity as something that was continuously available in the participant’s classroom but not something hypothesized to be highly preferred (e.g., a puzzle or blocks always available in the free play area of the classroom). Finally, iPods were used to record sessions and for data collection.

Data collection and analysis

For all SPAs and reinforcer assessments, trained graduate and undergraduate observers collected data through a one-way observation booth or retroactively via videotaped sessions. Data were collected using paper data sheets and pencils or using iPods. For the four participants with whom we conducted PSPAs (all but Spencer), observers recorded participants’ leisure-item selection for each SPA trial. Leisure-item selection was defined as any instance in which the participant placed their hand on or pointed to a presented item within 5 s of its presentation. If the participant vocally selected (i.e., said the name of) an item, the experimenters prompted the participant to physically select the item by saying, “Point to the one you want.” Experimenters determined participants’ selection percentages for each item by summing the number of times an item was selected, dividing the result by the number of times that item was presented, and multiplying by 100. The experimenters then determined the rankings for the leisure items, which ranged from 1, indicating the item with the highest selection percentage, to 6 (solitary or social SPAs) or 12 (combined SPA), indicating the item with the lowest selection percentage. If two items produced equal selection percentages, the experimenters reviewed the raw data to determine which item was selected more frequently when paired with the other item and assigned the higher rank to the more frequently selected item.

For Spencer, with whom we conducted a free-operant preference assessment with response restriction, the observers recorded interactions with leisure items by using iPods for each SPA session. Leisure-item interaction was defined as any instance in which Spencer’s hand contacted any part of the item for at least 1 s with an immediate onset and offset. Leisure-item interaction was recorded as duration in seconds, and the experimenters determined Spencer’s interaction percentage for each leisure item by dividing the total number of seconds that Spencer interacted with each item by the total number of seconds in the session and multiplying the quotient by 100. The experimenters then determined the rankings for leisure items, which ranged from 1, indicating the item with the highest interaction percentage, to 4 (solitary or social SPAs) or 8 (combined SPA), indicating the item with the lowest interaction percentage.

During reinforcer assessments, the data collectors scored correct and incorrect responding toward the target task. For Masie and Andrew, the target task was shape matching. Correct matches were defined as any instance in which the participant physically placed a shape cutout onto the corresponding shape in the binder. Incorrect matches were defined as any instance in which the participant matched a shape cutout to any shape other than the corresponding shape in the binder. For Josie and Ophelia, the target task was shape sorting. Correct sorts were defined as any instance in which the participant physically placed a shape cutout into the corresponding shape bin. Incorrect sorts were defined as any instance in which the participant placed a shape cutout into any bin other than the corresponding shape bin. Observers collected data on the frequency of correct matches or sorts, the frequency of incorrect matches or sorts, the duration of reinforcer access, and the terminal PR schedule (i.e., break point) in each session. Reinforcer access was defined as the time in seconds from when the experimenter provided access to the leisure item and social interaction (if applicable) to the moment access was removed. Experimenters determined the terminal PR schedule by identifying the last PR requirement that was successfully completed by the participant. For Spencer, the target task was in-square behavior, defined as any instance in which Spencer’s body was within one of the outlined areas of the session room. Observers recorded the duration of in-square behavior in seconds with an immediate onset and offset.

To address limitations that have been described in previous research, the experimenters also collected data on social consumption and leisure-item engagement during all reinforcement sessions. Social consumption was defined as any verbal interaction with the experimenter (e.g., initiating conversation, reciprocating conversation, manding for interaction, or gesturing), non vocal interaction with the experimenter (e.g., reciprocating toy play, head nodding, or sharing toys), or orientation toward the experimenter (e.g., eye contact or participant’s face and body within approximately 1 m of the experimenter and face angled 90 degrees or less toward the experimenter). Leisure-item engagement was defined as any instance in which the participant made physical contact with the leisure item (e.g., spinning a spin toy, building with blocks, pushing a toy car). If engagement with a leisure item did not require physical manipulation (e.g., watching a video on an iPad), engagement was defined as looking at the leisure item. Social consumption and leisure-item engagement were both recorded as duration in seconds with an immediate onset and 3-s offset.

Interobserver agreement

A second observer independently collected data on the participants’ item selection or interaction (Spencer only) for all SPAs. For the selection of leisure items, the observers calculated trial-by-trial interobserver agreement by dividing the number of trials with agreement (i.e., both observers recorded the same selection) by the total number of trials and multiplying the result by 100 to obtain a percentage. For interaction, the observers calculated agreement for total duration by dividing the smaller duration by the larger duration of interaction in seconds and multiplying the result by 100 to obtain a percentage. Mean agreement for selection was 100% for solitary, social, and combined SPAs for Masie, Josie, and Andrew. For Ophelia, mean agreement was 100% for the solitary and social SPAs and 97.73% (range: 95.45%– 100%) for the combined SPA. Mean agreement for duration of Spencer’s interaction with leisure items was 98.1% (range: 84.1%–100%) for the solitary SPA, 99.6% (range: 97%–100%) for the social SPA, and 99.1% (range: 75%– 100%) for the combined SPA.

A second, independent observer also collected data on correct matches or sorts during reinforcement sessions. For Andrew, Masie, Ophelia, and Josie, agreement was calculated using the proportional agreement method. That is, each 10-min session was separated into 10-s intervals, and the experimenters compared correct responses recorded across observers within each interval. The experimenters divided the smaller number of recorded responses by the larger number of recorded responses within each interval, summed the results, divided this result by the total number of intervals, and multiplied by 100 to obtain a percentage of agreement. For Andrew, interobserver agreement was calculated for a mean of 45.5% (range: 41.7%–50%) of sessions and averaged 99.1% (range: 96.9%–100%) for the iPad reinforcer assessment, 96.8% (range: 95.5%–99%) for the HP reinforcer assessment, and 96.1% (range: 93.7%–98.8%) for the LP reinforcer assessment. For Masie, agreement was calculated for a mean of 42.4% (range: 40%–55.6%) of sessions and averaged 91.5% (range: 86.8%–96.6%) for the iPad reinforcer assessment, 93.6% (range: 86.8%– 98.8%) for the HP reinforcer assessment, and 97.6% (range: 96.5%–100%) for the LP reinforcer assessment. For Josie, agreement was calculated for a mean of 37.13% (range: 33.33%–40%) of sessions and averaged 95% (range: 89.2%–100%) for the iPad reinforcer assessment, 95.6% (range: 92.5%–98.4%) for the HP reinforcer assessment, and 98.1% (range: 95.6%–100%) for the LP reinforcer assessment. For Ophelia, agreement was calculated for a mean of 70.75% (range: 61.5%–80%) of sessions and averaged 98.9% (range: 96.7%–100%) for the HP reinforcer assessment and 99.3% (range: 97.9%– 100%) for the LP reinforcer assessment. For Spencer, a second and independent observer also collected data on in-square behavior during reinforcement sessions. The experimenters calculated agreement for total duration of in-square behavior by dividing the smaller duration by the larger duration in seconds and multiplying the result by 100 to obtain a percentage. For Spencer, agreement was calculated for 54.9% (range: 33.33%–76.5%) of sessions and averaged 95.17% (range: 93%–100%) for the HP reinforcer assessment and 99.67% (range: 98.4%– 100%) for the LP reinforcer assessment.

Design

The frequency of correct matches or sorts was evaluated across a baseline and PR reinforcer assessment for Andrew, Masie, and Josie. For all participants except Spencer (see procedural modifications below), we conducted single-operant reinforcer assessments that included baseline, HP, and LP phases. For Andrew, Masie, and Josie, an iPad phase was also included prior to the HP and LP phases and the PR schedule gradually increased within session (see procedures below). For Ophelia and Spencer, an FR schedule was used throughout the reinforcement evaluation. Specifically, for Ophelia, the frequency of correct sorts was evaluated across a baseline and the FR reinforcer assessment. We implemented an FR-1 schedule for Ophelia throughout the reinforcer assessment given this schedule was more akin to reinforcement schedules used during her clinical programming. For Spencer, the duration of in-square behavior was evaluated across an HP and LP concurrent-operants reinforcer assessment. We implemented a concurrent-operants reinforcer assessment for Spencer due to observed levels of challenging behavior when preferred items were unavailable.

Within each reinforcer assessment, conditions alternated in a multielement design across several evaluations. First, an iPad-solitary, iPad-social, and social-interaction-only condition rapidly alternated within the iPad reinforcer assessment (if applicable). Then, the HP solitary, HP social, and social-interaction-only conditions rapidly alternated within the HP reinforcer assessment. Finally, the LP solitary, LP social, and social-interaction-only conditions rapidly alternated within the LP reinforcer assessment. If similar responding was observed across reinforcement conditions, the experimenters reversed to baseline and replicated the phase.

Procedures

Preassessment

All participants completed a preassessment prior to SPAs to determine the extent to which they could discriminate between the presentation of solitary and social leisure items. The experimenter placed three picture cards equidistant on the floor or desk in front of the participant. One card depicted the experimenter with a pleasant expression to be used in the presentation of social leisure items. Another card depicted a common object (e.g., a book), and the third card was a blank control card. The experimenter instructed the participant to touch one of the picture cards by saying, “Touch the picture of me” for the social-interaction card or “Touch book” for the object card. If the participant selected the correct or incorrect card, the experimenter removed all cards from the array, rotated the cards, and presented the next trial. The order of instructions given by the experimenter was determined via a random-number generator for each participant, and no programmed consequences were provided for correct or incorrect card selection. All participants selected the correct card for a minimum of 80% of trials across 15 trials and, therefore, continued to the SPAs.

Stimulus preference assessments

The experimenters conducted three separate SPAs for all participants to determine stimulus rankings for solitary leisure items (i.e., items presented alone) and social leisure items (i.e., items presented with social interaction) when presented in separate assessments (solitary and social SPAs) as well as when presented in the same assessment (combined SPAs). Consistent with previous research, the order of all participants’ assessments was solitary, social, then combined SPAs to reduce the likelihood of a previous programmed history of social interaction paired with leisure items influencing solitary SPA outcomes (Kanaman et al., 2022). As mentioned, a PSPA was conducted with four participants and a free-operant with response restriction preference assessment was conducted with Spencer (see procedural modifications below).

For all PSPAs, the experimenter labeled each stimulus and provided presession access (see specific procedures below). Immediately following presession access, the experimenter began the PSPA trials by presenting the first two stimuli equidistant on the desk or floor in front of the participant. The experimenter labeled each stimulus and allowed the participant to make a selection. Once the participant selected an item by touching or gesturing toward it, the experimenter removed the nonselected stimulus and provided 30 s of access to the selected stimulus (i.e., leisure item by itself or leisure item with continuous social interaction). After 30 s of access, the experimenter removed the selected stimulus (i.e., removed the item and discontinued social interaction, if applicable) and presented the next two stimuli. If the participant attempted to select two stimuli simultaneously, the experimenter blocked the selection by removing both stimuli from the array and re-presented the trial. If the participant did not select either stimulus within 5 s of presentation, the experimenter removed both stimuli and re-presented the trial. If the participant still did not select a stimulus, the experimenter removed both stimuli, recorded, “No selection,” and moved to the next pair of stimuli. No prompts or other feedback were provided throughout participants’ SPAs. This process was repeated until each stimulus was paired with every other stimulus twice to allow for each to be presented on either side of the participant.

Solitary paired-stimulus preference assessments Solitary PSPAs were conducted to determine the extent to which participants preferred leisure items when presented alone. Presession access included the experimenter presenting each leisure item alone and vocally labeling the item by saying, “This is the (leisure item name) to play by yourself” while pointing to the item. Then, the experimenter allowed the participant to interact with the item by themselves for 30 s. Immediately following presession access, the experimenter implemented the PSPA procedures as described above. If the participant attempted to interact with the experimenter at any point during the solitary PSPA, the experimenter indicated that social interaction was unavailable by stating, “I can’t talk right now.” If the participant made continued attempts to interact with the experimenter, the experimenter withheld attention and provided no other programmed consequence (i.e., the experimenter remained in their current position but avoided eye contact and any other interaction with the participant). The experimenter repeated the presentation and selection procedures outlined above for the duration of the solitary PSPA.

Social paired-stimulus preference assessments Social PSPAs were conducted to determine the extent to which participants preferred leisure items when presented with social interaction. Similar to the procedure for the solitary PSPAs, the experimenter first provided presession access; however, each leisure item was presented with the picture of the experimenter. That is, the experimenter placed their picture and the leisure item in front of the participant, pointed to the picture and the leisure item, and said, “This is the (leisure item name) to play with me” before providing 30 s of access to the item with continuous social interaction. Social interaction in this condition consisted of conversation, comments, and play surrounding the presented item. For example, if the presented item was a book, the experimenter would read the book to the participant, describe the pictures, or have other conversation surrounding the story. The experimenter initially guided engagement and conversation with the leisure item but allowed the participant to lead if applicable (i.e., if the participant continued or initiated play or conversation surrounding the leisure item). All social interaction was contextual to the participant- and experimenter-led engagement with the leisure item that was presented. In addition, if the participant requested a different type of social interaction than was being provided (e.g., said, “Let’s play chase” when the book was presented), the experimenter redirected the participant to the available social interaction (e.g., the experimenter said, “We’re reading our book right now”). If the participant requested to end social interaction, the experimenter offered a different comment or interaction within the same context of play for that leisure item. Following presession access to each leisure item with social interaction, the experimenter implemented the presentation and selection PSPA procedures as described above for the duration of the social PSPA.

Combined paired-stimulus preference assessments

The combined PSPAs were conducted to determine the extent to which participants preferred leisure items when presented with social interaction (as done in social PSPAs) or without social interaction (as done in solitary PSPAs) within the same assessment. To allow a single leisure item to be presented in both a social and solitary manner, six leisure items plus duplicates of those same six items presented without social interaction were included. The experimenter first provided presession access by presenting each item (either with or without continuous social interaction, depending on the trial) in the same ways as described in the solitary and social PSPAs. The experimenter then presented the first two leisure items (either with or without pictures of the experimenter, depending upon the trial). Once the participant selected an item, the experimenter removed the non-selected item (and picture, if applicable) and provided 30 s of access to the selected item with or without continuous social interaction, depending on the selection. After 30 s of access, the experimenter removed the materials, discontinued social interaction (if applicable), and presented the next two items. The same procedures as were used in previous PSPAs were used in the combined PSPAs if participants attempted to select more than one item or did not select an item within 5 s of its presentation.

Free-operant with response restriction (Spencer only) For Spencer, all SPAs were conducted in a free-operant-with-response-restriction format using procedures that were similar to those that were reported in Hanley et al. (2003). The experimenter first provided presession access in the same ways as described in the PSPA procedures. When beginning the solitary SPA, the experimenter presented the four leisure items by themselves equidistant in a semicircle on the floor in front of Spencer. The experimenter labeled each item and informed Spencer he could play with one, some, or none of the items and allowed him to interact with any item that he approached for the duration of the 5-min session without providing social interaction. For the social SPA, the procedures were identical to those for the solitary SPA, with the addition of a picture of the experimenter when presenting leisure items and the continuous delivery of social interaction (as described in PSPAs) with each item interaction. That is, during the social SPA, the experimenter placed their picture directly above each leisure item in the array; if Spencer interacted with an item, the experimenter provided continuous social interaction. For the combined SPA, the procedures were identical to those for the solitary and social SPAs, with the addition of duplicate items. That is, four leisure items were presented alone and four duplicate leisure items were presented with the picture of the experimenter to signal the availability of social interaction in the same array. To determine preferences, the experimenter recorded the item with which Spencer interacted for the longest duration at the end of each session and then removed that item from the array in the subsequent session. Throughout the assessment, the experimenter would have blocked any instance in which Spencer attempted to select multiple items from the array simultaneously (e.g., a social and solitary item); however, this never occurred. Following each session, the experimenter rotated the order of the remaining items and continued with the next session until the assessment was complete.

Reinforcer assessments

During the solitary conditions of the reinforcer assessment, reinforcement consisted of contingent access to the HP or LP leisure item that was identified from the combined SPA by itself (i.e., without social interaction). If the participant attempted to interact with the experimenter during a solitary condition, the experimenter stated, “I can’t talk right now.” During social conditions, reinforcement consisted of contingent access to the HP or LP leisure item from the combined SPA with continuous social interaction from the experimenter. During iPad, HP, and LP social conditions, social interaction consisted of comments, conversation, and play all relevant to the leisure activity or iPad (i.e., the same social interaction provided with social stimuli during SPAs). As in the SPAs, the experimenter initially guided then followed the participant’s lead with respect to conversation and play with the leisure items. During the social-interaction-only control condition, social interaction included general comments about the participant’s environment (i.e., neutral conversation) and a pleasant facial expression.

At the beginning of each reinforcer assessment session, the experimenter placed the target task materials and the alternative task materials equidistant on the table in front of the participant. The experimenter then provided a rule and presession exposure to the contingency in that condition using a three-step prompting sequence. During baseline sessions, no programmed consequences were delivered for correct responding. For Josie, Andrew, and Masie, correct responding was reinforced on a PR schedule. We doubled the PR schedule following the completion of two response requirements at a particular schedule (e.g., FR 1, FR 1, FR 2, FR 2, FR 4, FR 4, FR 8, FR 8) within the session to produce 1 min of reinforcement associated with that condition (Jarmolowicz & Lattal, 2010; Roane, 2008), and the PR schedule reset to FR 1 at the start of each new session (Harper et al., 2021). For Ophelia, correct responding was reinforced on an FR-1 schedule in which one correct response was required to produce 1 min of reinforcement associated with that condition throughout session. All reinforcer-assessment sessions ended after 2 min without responding toward the target task or after 10 min elapsed, whichever came first. Additionally, the duration of reinforcer delivery was removed from the total session time to control for opportunities to respond across sessions. The order of all sessions for each participant was quasi random. Specifically, prior to conducting sessions, the experimenter wrote the names of each condition on separate pieces of paper and placed all pieces of paper into a bowl. Then, the experimenter selected one piece of paper at a time and conducted sessions in the selected order for each phase of the reinforcer assessment.

Baseline

During baseline sessions, neither the preferred items nor the pictures of the experimenter were available. At the start of session, the experimenter provided presession exposure to the response requirement by first vocally informing the participant about the task and the consequence for that condition (e.g., “If you [match or sort], nothing happens”). Then, the experimenter prompted the participant to match or sort one shape and provided no programmed consequence. The participant was free to match (or sort) shapes or engage with the alternative task materials for the duration of the session.

iPad (solitary and social)

During the solitary and social conditions with the iPad, the iPad, target task materials, picture of the experimenter (social condition only), and alternative task materials were placed in front of the participant. The experimenter first vocally informed the participant about the task and the consequence for that condition (e.g., “If you match [or sort], you get to play the iPad by yourself”; “If you match [or sort], you get to play the iPad with me”). The experimenter then prompted the participant to match [or sort] one shape and delivered the iPad for 1 min with or without social interaction, depending on the condition. Then, the experimenter removed access to the iPad (and social interaction, if applicable). Following presession exposure, the experimenter started the session timer, withheld access to the iPad and social interaction, and allowed the participant to respond. Contingent on one correct match [or sort], the experimenter delivered 1 min of access to the iPad without social interaction (solitary) or with social interaction (social). After 1 min, the experimenter removed access to the iPad (and social interaction, if applicable) and allowed another opportunity to respond. For individuals under the PR schedule, the response requirement increased as previously described. For Ophelia, the response requirement remained at FR 1 for the duration of the session. Participants were free to match (or sort) shapes or engage with the alternative task materials for the duration of the session.

High preferred (solitary and social)

During the HP solitary and social conditions, the HP leisure item, target task materials, picture of the experimenter (social condition only), and alternative task materials were placed in front of the participant. The experimenter first vocally informed the participant of the task and the consequence for that condition (e.g., “If you match [or sort], you get to play [name of HP leisure item] by yourself”; “If you match [or sort], you get to play [name of HP leisure item] with me”). The remaining procedures were identical to those described in the iPad reinforcer assessment, except that the HP item was delivered in place of the iPad.

Low preferred (solitary and social)

During the LP solitary and social conditions, sessions were identical to those for the HP conditions described above except the LP leisure item was used.

Social-interaction-only control

During the social-interaction-only control condition, only the picture of the experimenter was placed in front of the participant. The experimenter first vocally informed the participant about the task and the consequence for that condition (e.g., “If you match [or sort], you get to talk to me”). The experimenter prompted the participant to match (or sort) one shape and provided 1 min of neutral social interaction. Next, the experimenter removed access to social interaction, started the session timer, and allowed the participant to respond. Contingent on one correct match (or sort), the experimenter delivered 1 min of continuous access to the same type of social interaction that was delivered in presession exposure (i.e., neutral conversation and general comments about the participant’s environment with a pleasant facial expression). After 1 min, the experimenter removed access to social interaction and allowed another opportunity to respond. The response requirement increased within session as described in previous conditions, and the participant was free to match (or sort) shapes or engage with the alternative task materials for the duration of the session.

Procedural modifications (Spencer)

As previously mentioned, Spencer’s reinforcer assessment was conducted using a concurrent-operants arrangement due to observed levels of Spencer’s challenging behavior when preferred items were unavailable. To conduct Spencer’s assessment, the session room was first separated into three concurrently available and condition-specific areas. Specifically, the room was separated into three large and equidistant squares (approximately 1  1.5 m) by placing tape on the floor to segment each area. The areas were arranged such that Spencer had enough space to walk between each square, stand within each square, and stand outside of the squares. Prior to each session, the experimenter arranged each square to depict solitary, social, and social-interaction-only conditions. That is, one square contained the HP or LP item (depending on whether it was an HP or LP condition) by itself (solitary), another square contained the HP or LP item and the picture of the experimenter (social), and the last square contained only the picture of the experimenter (social interaction only).

At the start of each session, the experimenter vocally informed Spencer about the condition and provided presession exposure to the stimulus in each square. That is, the experimenter prompted Spencer to step into each square and provided 30 s of access to the solitary, social, and social-interaction-only conditions. The experimenter informed Spencer that he could go back and forth between the squares or enter none of the squares. Contingent on Spencer entering one of the squares (i.e., once Spencer fully stepped into a square), the experimenter provided continuous access to the HP or LP item (with or without social interaction, depending upon the condition) or continuous access to neutral social interaction (social interaction only) for the duration that Spencer remained in the square. If Spencer entered none of the squares, the experimenter stood in the corner of the room and refrained from providing social interaction. The experimenter continued these procedures for the duration of the session. Following each completed session, the experimenter rotated the location of each stimulus within each square until the evaluation was complete.

Procedural fidelity

During the SPAs, secondary observers collected data on correct stimulus delivery (i.e., leisure item delivery with or without social interaction, depending on the programmed stimulus) during all SPA trials across participants. During PSPAs, observers scored stimulus delivery as correct if the experimenter correctly delivered the selected leisure item and social interaction (if applicable) on a trial. The observer scored the delivery of stimuli as incorrect if the experimenter failed to deliver a programmed stimulus for a selection or delivered the stimulus in a way that was different from what was programmed (e.g., delivered social interaction following selection of a solitary leisure item, delivered social interaction in a way that differed from the definition). For Spencer, the observer scored stimulus delivery as correct if the experimenter correctly provided access to the leisure item with which Spencer interacted and social interaction (if applicable) for the duration of Spencer’s interaction. The observer scored stimulus delivery as incorrect if the experimenter failed to provide access to the selected stimulus for the duration of Spencer’s interaction or provided access to a stimulus in a way that was different than programmed. Observers recorded correct or incorrect stimulus delivery following each selection or period of interaction and divided the number of opportunities with correct stimulus delivery by the total number of selections to obtain a percentage. The mean percentage of correct stimulus delivery was 100% for solitary, social, and combined SPAs for Masie, Josie, Ophelia, and Spencer. For Andrew, the mean percentage of correct stimulus delivery was 97% (range: 93.33%–100%) for the solitary SPA and 100% during social and combined SPAs.

During the reinforcer assessments, data that were collected on target responding and reinforcer access were used to calculate procedural fidelity across participants. Data collectors scored procedural fidelity as correct if the experimenter delivered the condition specific stimulus (i.e., leisure item only, leisure item with social interaction, or social interaction only) in the correct way (i.e., presence or absence of leisure item with or without social interaction, depending upon the condition) within 3 s of completion of the current PR requirement (i.e., when the required number of correct matches or sorts under the current schedule was completed) or within 3 s of entering the square (Spencer only). Data collectors scored procedural fidelity as incorrect if the experimenter incorrectly delivered the condition-specific stimulus (e.g., delivered a leisure item with social interaction during a solitary condition) or delivered no stimulus within 3 s of completion of the response requirement. Data collectors also scored procedural fidelity as incorrect if the experimenter delivered a stimulus at any time other than at the completion of the response requirement (e.g., delivered the item after seven correct matches during the PR-8 schedule requirement). The experimenters divided the number of correct stimulus deliveries by the total number of opportunities and multiplied by 100 to obtain a percentage. For Andrew, procedural fidelity was assessed for a mean of 61.07% (range: 58.3%–66.6%) of sessions and averaged 97.6% (range: 83%–100%) for the iPad reinforcer assessment, 97.9% (range: 85.7%– 100%) for the HP reinforcer assessment, and 100% for the LP reinforcer assessment. For Masie, procedural fidelity was assessed for a mean of 42.4% (range: 40%– 55.6%) of sessions and averaged 83.8% (range: 70%– 100%) for the iPad reinforcer assessment, 95.0% (range: 87.5%–100%) for the HP reinforcer assessment, and 97.1% (range: 85.7%–100%) for the LP reinforcer assessment. For Josie, procedural fidelity was assessed for a mean of 43.63% (range: 33.33%–50%) of sessions and averaged 94.2% (range: 70%–100%) for the iPad reinforcer assessment, 98.9% (range: 90%–100%) for the HP reinforcer assessment, and 98.2% (range:

90.9%–100%) for the LP reinforcer assessment. For Ophelia, procedural fidelity was assessed for a mean of

93.35% (range: 86.7%–100%) of sessions and averaged 96.9% (range: 85.7%–100%) for the HP reinforcer assessment and 99.2% (range: 93.33%–100%) for the LP reinforcer assessment. For Spencer, procedural fidelity was assessed for a mean of 83.35% (range: 66.7%–100%) of sessions and averaged 93.26% (range: 87%–100%) for the HP reinforcer assessment and 91.4% (range: 77%–100%) for the LP reinforcer assessment. Across participants, errors in procedural fidelity were observed during relatively high PR requirements (e.g., PR 16, PR 32) when the experimenter miscounted the participant’s matches or sorts and delivered the reinforcer prior to or soon after the response requirement was met (e.g., the experimenter delivered the reinforcer following 33 instead of 32 sorts).

Results

Figure 1 displays the results of the solitary, social, and combined SPAs for Josie, Ophelia, and Andrew given the similarities in their outcomes. Specifically, participants’ preference rankings remained relatively stable across assessments and the same two items were ranked “1” and “2” across solitary and social assessments for these participants. Additionally, for participants for whom the iPad was included (Josie and Andrew), the iPad was ranked “1” across all assessments. However, the combined SPA data showed some displacement of preference rankings across participants. For Josie, all items that were presented with social interaction were ranked higher than items that were presented without social interaction in the combined SPA. For Ophelia, almost all leisure items that were presented without social interaction were ranked higher than those that were presented with social interaction in the combined SPA. For Andrew, combined SPA outcomes were mixed; however, several leisure items that were presented without social interaction were ranked higher than items that were presented with social interaction.

Figure 2 displays the results of solitary, social, and combined SPAs for Masie and Spencer. Similar to the rankings of the other participants, the same two items were ranked “1” and “2” across solitary and social assessments and participants’ preference hierarchies remained relatively stable across all three assessments. For Masie, the iPad was ranked as “1” across all assessments, like other participants for whom the iPad was included. For both participants, the combined SPA data showed mixed preferences for solitary and social stimuli. That is, preference rankings of solitary and social stimuli depended on the leisure item.

Figure 3 displays the results of the reinforcer assessments for Andrew, Masie, and Ophelia given similarities in their outcomes during one or more social conditions.

FIGURE 1 Stimulus preference assessment results for Josie, Ophelia, and Andrew showing overall higher (Josie and Ophelia) or lower (Andrew) preference for a single stimulus category. SI = social interaction.

FIGURE 2 Stimulus preference assessment results for Masie and Spencer showing mixed preferences across stimulus categories. SI = social interaction.

FIGURE 3 Reinforcer assessment results for Andrew, Masie, and Ophelia showing differentiated responding in one or more social conditions. BL = baseline; SR = reinforcement; HP = high preferred; LP = low preferred; SI = social interaction; “a” indicates sessions that were terminated due to lack of responding.

Specifically, Andrew and Masie engaged in low levels of correct responding in baseline and higher levels of correct responding in the solitary and social test conditions relative to the social-interaction-only control condition during the iPad phase. Given that we observed overall higher levels of correct responding and differentiation between both test conditions relative to the control condition with limited variability, we moved to the HP-item phase. During this phase, Andrew and Masie engaged in higher levels of correct responding during the social test condition relative to the solitary test condition and differentiation was observed between test conditions and the social-interaction-only control condition. Finally, during the LP-item phase, Andrew continued to engage in high levels of correct responding in the social test condition relative to the solitary test condition and social-interaction-only control condition, with a high level of differentiation; however, Masie engaged in low levels of correct responding across all conditions, with minimal variability. For Masie, these results corresponded to her combined SPA. For Andrew, the combined SPA results were mixed; however, the reinforcer assessment results show that Andrew completed more responses when access to the HP and LP items was presented with social interaction. These results suggest that Andrew’s combined SPA produced at least a partial false-negative outcome (DeLeon et al., 1997; N. M. Goldberg et al., 2023).

Ophelia engaged in low levels of correct responding in baseline with minimal variability and higher levels of correct responding during both solitary and social test conditions relative to the social-interaction-only control condition in the HP-item phases. During the first LP-item phase, Ophelia engaged in much higher levels of correct responding during the social test condition relative to the solitary test condition and social-interaction-only control condition, suggesting that social interaction increased the reinforcing value of the LP item. These results suggest a false-negative outcome for Ophelia’s combined SPA given that almost all solitary items outranked social items. In Ophelia’s second HP-item phase, we observed a higher degree of variability; however, levels of responding across all conditions were similar to those that were observed during the first HP-item phase. During the second LP-item phase, Ophelia engaged in much higher levels of correct responding during both social and solitary test conditions relative to the social-interaction-only control condition; however, high levels of correct responding were observed in the final social-interaction-only control condition relative to previous control conditions. Overall, Ophelia’s responding during the social test condition in the second LP-item phase mirrored her responding during the first LP-item phase, whereas her responding during the solitary test condition was much higher in the second LP-item phase than it was in the first LP-item phase.

Figure 4 displays the results of the reinforcer assessment for Josie. Josie engaged in low levels of correct responding in baseline and high levels of correct responding across all conditions (including the social-interaction-only control condition) in the iPad and HP item phases with slightly higher levels of correct responding in the social test conditions relative to solitary and social-interaction-only conditions. In the LP-item phase, Josie engaged in higher levels of correct responding in the social and solitary test conditions relative to the social-interaction-only control condition. These results partially correspond to Josie’s combined SPA, which showed that all stimuli that were presented with social interaction outranked those that were presented without social interaction. Overall, the results of Josie’s reinforcer assessment suggest that leisure items that are presented with and without social interaction functioned as effective reinforcers.

FIGURE 4 Reinforcer assessment results for Josie. BL = baseline; SR = reinforcement; HP = high preferred; LP = low preferred; SI = social interaction; “a” indicates sessions that were terminated due to lack of responding

Figure 5 displays the results of the concurrent-operants reinforcer assessment for Spencer. During both the HP- and LP-item phases, Spencer allocated the majority of session time to either the solitary or social test conditions relative to the social-interaction-only control condition, with more time spent in the social test condition overall. Preferences for treatment conditions relative to control conditions are not uncommon (e.g., Dozier et al., 2007), and these results correspond to Spencer’s combined SPA outcomes for the HP item. However, Spencer’s combined SPA data indicated that the LP item was less preferred when presented with social interaction. Similar to Andrew and Ophelia, the results of Spencer’s reinforcer assessment suggest a false-negative outcome in that Spencer allocated more responding toward the LP social test condition relative to the LP solitary test condition despite a lower ranking of the LP social item in the combined SPA.

FIGURE 5 Reinforcer assessment results for Spencer. SR = reinforcement; HP = high preferred; LP = low preferred; SI = social interaction

Figures 6 and 7 display participants’ data for social consumption and leisure-item engagement, respectively. The data for social consumption show that participants generally consumed the social interaction that was provided for most of the sessions across all conditions. Additionally, some participants showed higher levels of social consumption in social test conditions relative to the social-interaction-only control condition (e.g., Andrew during the iPad, HP, and LP conditions; Masie during the HP social condition). The data for leisure-item engagement show varying levels of engagement across conditions and participants. For example, Andrew’s data show high levels of engagement in both solitary and engagement during the social condition than during the social conditions during the iPad and HP phases; how- solitary condition. Similar results are observed in Masie’s ever, during the LP phase, Andrew showed much higher and Ophelia’s HP and LP phases. For the other participants (i.e., Josie, Spencer), high levels of leisure item engagement were observed during the social and solitary conditions of all reinforcer assessment phases.

FIGURE 6 Social consumption. HP = high preferred; LP = low preferred; SI = social interaction. White and gray bars represent the mean of values. Black error bars represent the range of values.

FIGURE 7 Leisure-item engagement data. HP = high preferred; LP = low preferred. White and gray bars represent the mean of values. Black error bars represent the range of values.

Discussion

Overall, the results of the current study replicated and extended those of previous research by demonstrating the influence of social interaction on preference and reinforcing efficacy of leisure items for children with autism. For all participants, both HP and LP items that were identified from the combined SPA functioned as reinforcers and social interaction increased the reinforcing efficacy of one or more leisure items. Interestingly, the combined SPA produced false-negative outcomes for three participants in that items that were identified as LP functioned as reinforcers, producing the same or higher levels of responding as HP items did when the LP items were presented with social interaction. This finding replicates the results of N. M. Goldberg et al. (2023), indicating that combined SPAs that include social interaction as a stimulus category may not fully capture the reinforcing efficacy of LP social stimuli.

These findings extend research on variables that influence preference for leisure items and reinforcer efficacy for children with autism in several ways. First, we found that the iPad functioned as a reinforcer regardless of the provision of social interaction for all participants for whom the iPad was included. This finding supports those of Conine and Vollmer (2018), suggesting that the inclusion of screen-based media influences preference hierarchies. Second, the HP item from the combined SPA functioned as an effective reinforcer for all participants and the reinforcing value of the HP item increased when presented with social interaction for four participants. Interestingly, only the reinforcing value of the LP item increased when presented with social interaction for Ophelia. This finding suggests that social interaction likely increases the value of some leisure items; however, certain items may remain more preferred and reinforcing when presented alone. Therefore, continued research is warranted to determine variables that affect the reinforcing function of social interaction within the context of leisure activities. Third, combined SPAs failed to accurately predict the reinforcing value of HP and LP items (presented with and without social interaction) for some participants. For example, almost all solitary stimuli outranked social stimuli in Ophelia’s combined SPA; however, Ophelia engaged in more correct responses to access the LP item with social interaction. Additionally, Ophelia’s responding to access the LP solitary item increased substantially from the first LP item phase to the second LP item phase. Anecdotally, the experimenters observed Ophelia playing with the LP item by herself in ways she had not during the first LP solitary phase (e.g., she began imitating the experimenter’s vocalizations and motions with the LP item). Therefore, it is possible that Ophelia’s increased level of responding in the second LP solitary phase was a result of learning. This finding suggests that exposure to social interaction with LP items can function to condition preferences and increase the reinforcing value of items by teaching new ways to engage with items (Hanley et al., 2006). Fourth, the order of all participants’ SPAs was solitary, social, then combined to decrease the potential influence of social SPAs on subsequent solitary SPAs; however, researchers may consider changing that order in future investigations to evaluate the stability of preferences following exposure to leisure items with social interaction.

Although results from the current study extend the literature in several ways, there are limitations worth noting. First, we did not control for access to or restriction of HP items, LP items, or social interaction outside of participants’ research sessions due to clinical programming needs. It is possible that stimulus access outside of research sessions influenced motivation for reinforcers (Hanley et al., 2006; Vollmer & Iwata, 1991). Second, although we used target tasks that were similar to those that are used in participants’ clinical programming, the extent to which similar reinforcement effects would be obtained with other responses (e.g., more complex curriculum tasks, activities of daily living) is unknown. Thus, future research might involve evaluating the reinforcing efficacy of leisure items that are presented with and without social interaction within the context of skill acquisition (i.e., novel tasks) and mastered tasks. It is possible that reinforcing efficacy depends on not only the degree of preference and inclusion of social interaction but also the difficulty associated with the task and the presence or absence of a reinforcement history for the target response. Relatedly, the PR schedules that were used during Andrew, Masie, and Josie’s reinforcer assessments all started at PR 1 and increased (i.e., progressively doubled) within session; thus, the extent to which similar results would be obtained if we had started sessions with a less dense schedule of reinforcement (e.g., PR 10) or maintained a consistent schedule (e.g., FR 10) within sessions is unknown. In the future, researchers may evaluate the influence of social interaction on reinforcement effects using schedules of reinforcement that are commonly implemented in early intervention or classroom settings (e.g., completion of one worksheet to access reinforcement).

Third, although the social interaction provided during social conditions was programmed to be item specific, this did not allow the researchers to evaluate preference or reinforcing efficacy of other types of social interaction. Anecdotally, some participants (Andrew, Ophelia, and Spencer) occasionally requested non-programmed forms of social interaction (e.g., tickles, bounces, chase). Given that the experimenters did not provide non-programmed interactions, the extent to which individually identified and preferred forms of social interaction would have influenced outcomes remains unknown. In the future, researchers may include individualized forms of social interaction as a point of comparison. Furthermore, all sessions were conducted by the same experimenter for each participant. Therefore, it is unknown whether social interaction provided by other individuals (e.g., peers, less familiar adults) would be equally preferred and reinforcing. Thus, future research might involve investigating other parameters of interactions (e.g., provider, duration, presence or absence of instructions and modeling, compliance with requests).

Overall, results indicate that preference for and reinforcing efficacy of leisure items are influenced by the inclusion of social interaction for children with autism and that the degree of influence may be idiosyncratic. That is, for some individuals, social interaction may increase preference for and reinforcing efficacy of many or all leisure items; for other individuals, social interaction may increase preference for and reinforcing efficacy of only certain leisure items. In the future, researchers may conduct quantitative analyses to further describe the degree of displacement from solitary and social SPAs to combined SPAs or conduct correspondence analyses between SPAs and reinforcer-assessment outcomes to facilitate a detailed interpretation of the results. Thus far, our findings suggest that clinicians who are programming leisure items and activities as reinforcers should consider (a) their presentation during SPAs (i.e., whether leisure items are presented with or without social interaction during the assessment) and (b) the presentation of leisure items as programmed reinforcers (i.e., whether the leisure items are presented with or without social interaction when delivered following a target behavior for increase). Furthermore, given that we observed higher levels of engagement with leisure items during social conditions relative to those observed during solitary conditions for several participants, clinicians should strongly consider providing social interaction during reinforcer-access periods to promote item engagement and potentially condition social interaction to be reinforcing. Additionally, presenting leisure items and activities with social interaction may enhance the social and ecological validity of reinforcement procedures given that interactions commonly occur in the natural environment. Furthermore, the incorporation of social interaction may be a useful procedural adaptation to differential reinforcement procedures that are used for skill acquisition or to decrease challenging behavior, particularly for individuals whose challenging behavior is maintained by access to attention or tangibles. It is likely worthwhile for clinicians to determine which items are more reinforcing when presented with and without social interaction to program reinforcement most effectively.

Acknowledgments

We would like to thank McKenna Reilly, Abigail Rains, and Nicole Dowell for their assistance with data collection.

Conflict of Interest Statement

The authors have no conflicts of interest to disclose regarding the current manuscript.

Data Availability Statement

Data are available from the corresponding author upon request.

Ethics Approval

This study received institutional review board approval and was conducted in accordance with established ethical guidelines for the treatment of human participants. Caregivers provided informed consent for all participants. Prior to all sessions, assent from the participant was required.

References

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). https://doi.org/10.1176/appi.books.9780890425596

Call, N. A., Shillingsburg, M. A., Bowen, C. N., Reavis, A. R., & Findley, A. J. (2013). Direct assessment of preferences for social interactions in children with autism. Journal of Applied Behavior Analysis, 46(4), 821–826. https://doi.org/10.1002/jaba.69

Carr, J. E., Nicolson, A. C., & Higbee, T. S. (2000). Evaluation of a brief multiple-stimulus preference assessment in a naturalistic context. Journal of Applied Behavior Analysis, 33(3), 353–357. https://doi.org/10.1901/jaba.2000.33-353

Conine, D. E., & Vollmer, T. R. (2018). Relative preferences for edible and leisure stimuli in children with autism. Journal of Applied Behavior Analysis, 52(2), 557–573. https://doi.org/10.1002/jaba.525

DeLeon, I. G., Frank, M. A., Gregory, M. K., & Allman, M. J. (2009). On the correspondence between preference assessment outcomes and progressive-ratio schedule assessments of stimulus value. Journal of Applied Behavior Analysis, 42(3), 729–733. https://doi.org/10.1901/jaba.2009.42-729

DeLeon, I. G., Iwata, B. A., & Roscoe, E. M. (1997). Displacement of leisure reinforcers by food during preference assessments. Journal of Applied Behavior Analysis, 30(3), 475–484. https://doi.org/10.1901/jaba.1997.30-475

Dozier, C. L., Vollmer, T. R., Borrero, J. C., Borrero, C. S., Rapp, J. T., Bourret, J., & Gutierrez, A. (2007). Assessment of preference for behavioral treatment versus baseline conditions. Behavioral Interventions, 22(3), 245–261. https://doi.org/10.1002/bin.241

Fahmie, T. A., Iwata, B. A., & Jann, K. E. (2015). Comparison of edible and leisure reinforcers. Journal of Applied Behavior Analysis, 48(2), 331–343. https://doi.org/10.1002/jaba.200

Fisher, W., Piazza, C. C., Bowman, L. G., Hagopian, L. P., Owens, J. C., & Slevin, I. (1992). A comparison of two approaches for identifying reinforcers for persons with severe and profound disabilities. Journal of Applied Behavior Analysis, 25(2), 491–498. https://doi.org/10.1901/jaba.1992.25-491

Francisco, M. T., Borrero, J. C., & Sy, J. R. (2008). Evaluation of absolute and relative reinforcer value using progressive-ratio schedules. Journal of Applied Behavior Analysis, 41(2), 189–202. https://doi.org/10.1901/jaba.2008.41-189

Goldberg, M. C., Allman, M. J., Hagopian, L. P., Triggs, M. M., Frank-Crawford, M. A., Mostofsky, S. H., Denckla, M. B., & DeLeon, I. G. (2017). Examining the reinforcing value of stimuli within social and non-social contexts in children with and without high-functioning autism. Autism, 21(7), 881–885. https://doi.org/10.1177/1362361316655035

Goldberg, N. M., Roscoe, E. M., Newman, Z. A., & Sedano, A. J. (2023). Single- vs. combined-category preference assessments for edible, leisure, and social-interaction stimuli. Journal of Applied Behavior Analysis, 56(4), 787–803. https://doi.org/10.1002/jaba.1007

Graff, R. B., Gibson, L., & Galiatsatos, G. T. (2006). The impact of high- and low-preference stimuli on vocational and academic performances of youths with severe disabilities. Journal of Applied Behavior Analysis, 39(1), 131–135. https://doi.org/10.1901/jaba.2006.32-05

Gutierrez, A., Fischer, A. J., Hale, M. N., Durocher, J. S., & Alessandri, M. (2013). Differential response patterns to the control condition between two procedures to assess social reinforcers for children with autism. Behavioral Interventions, 28(4), 353–361. https://doi.org/10.1002/bin.1372

Hagopian, L. P., Long, E. S., & Rush, K. S. (2004). Preference assessment procedures for individuals with developmental disabilities. Behavior Modification, 28(5), 668–677. https://doi.org/10.1177/0145445503259836

Hanley, G. P., Iwata, B. A., Lindberg, J. S., & Conners, J. (2003). Response-restriction analysis: I. Assessment of activity preferences. Journal of Applied Behavior Analysis, 36(1), 47–58. https://doi.org/10.1901/jaba.2003.36-47

Hanley, G. P., Iwata, B. A., & Roscoe, E. M. (2006). Some determinants of changes in preference over time. Journal of Applied Behavior Analysis, 39(2), 189–202. https://doi.org/10.1901/jaba.2006.163-04

Harper, A. M., Dozier, C. L., Briggs, A. M., Diaz de Villegas, S., Ackerlund Brandt, J. A., & Jowett Hirst, E. S. (2021). Preference for and reinforcing efficacy of different types of attention in preschool children. Journal of Applied Behavior Analysis, 54(3), 882– 902. https://doi.org/10.1002/jaba.814

Hodos, W. (1961). Progressive ratio as a measure of reward strength. Science, 134(3483), 943–944. https://doi.org/10.1126/science.134.3483.943

Jarmolowicz, D. P., & Lattal, K. A. (2010). On distinguishing progressively increasing response requirements for reinforcement. The Behavior Analyst, 33(1), 119–125. https://doi.org/10.1007/bf03392207

Jones, B. A., Dozier, C. L., & Neidert, P. L. (2014). An evaluation of the effects of access duration on preference assessment outcomes. Journal of Applied Behavior Analysis, 47(1), 209–213. https://doi.org/10.1002/jaba.100

Kanaman, N. A., Hubbs, A. L., Dozier, C. L., Jones, B. A., Foley, E., & Ackerlund Brandt, J. (2022). Evaluating the effects of social interaction on the results of preference assessments for leisure items. Journal of Applied Behavior Analysis, 55(2), 430–450. https://doi.org/10.1002/jaba.897

Kelly, M. A., Roscoe, E. M., Hanley, G. P., & Schlichenmeyer, K. (2014). Evaluation of assessment methods for identifying social reinforcers. Journal of Applied Behavior Analysis, 47(1), 113–135. https://doi.org/10.1002/jaba.107

MacNaul, H., Cividini-Motta, C., Wilson, S., & Di Paola, H. (2021). A systematic review of research on stability of preference assessment outcomes across repeated administrations. Behavioral Interventions, 36(4), 962–983. https://doi.org/10.1002/bin.1797

Morris, S. L., & Vollmer, T. R. (2019). Assessing preference for types of social interaction. Journal of Applied Behavior Analysis, 52(4), 1064–1075. https://doi.org/10.1002/jaba.597

Nuernberger, J. E., Smith, C. A., Czapar, K. N., & Klatt, K. P. (2012). Assessing preference for social interaction in children diagnosed with autism. Behavioral Interventions, 27(1), 33–44. https://doi.org/10.1002/bin.1336

Pace, G. M., Ivancic, M. T., Edwards, G. L., Iwata, B. A., & Page, T. J. (1985). Assessment of stimulus preference and reinforcer value with profoundly retarded individuals. Journal of 260 KAMLOWSKY ET AL. 19383703, 2025, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.2919 by Behavior Analyst Certification Board, Wiley Online Library on [14/04/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Applied Behavior Analysis, 18(3), 249–255. https://doi.org/10.1901/jaba.1985.18-249

Paden, A. R., & Kodak, T. (2015). The effects of reinforcement magnitude on skill acquisition for children with autism. Journal of Applied Behavior Analysis, 48(4), 924–929. https://doi.org/10.1002/jaba.239

Penrod, B., Wallace, M. F., & Dyer, E. J. (2008). Assessing potency of high- and low-preference reinforcers with respect to response rate and response patterns. Journal of Applied Behavior Analysis, 41(2), 177–188. https://doi.org/10.1901/jaba.2008.41-177

Piazza, C. C., Fisher, W. W., Hagopian, L. P., Bowman, L. G., & Toole, L. (1996). Using a choice assessment to predict reinforcer effectiveness. Journal of Applied Behavior Analysis, 29(1), 1–9. https://doi.org/10.1901/jaba.1996.29-1

Roane, H. S., Lerman, D. C., & Vorndran, C. M. (2001). Assessing reinforcers under progressive schedule requirements. Journal of Applied Behavior Analysis, 34(2), 145–166. https://doi.org/10.1901/jaba.2001.34-145

Roane, H. S. (2008). On the applied use of progressive-ratio schedules of reinforcement. Journal of Applied Behavior Analysis, 41(2), 155–161. https://doi.org/10.1901/jaba.2008.41-155

Roscoe, E. M., Iwata, B. A., & Kahng, S. (1999). Relative versus absolute reinforcement effects: Implications for preference assessments. Journal of Applied Behavior Analysis, 32(4), 479–493. https://doi.org/10.1901/jaba.1999.32-479

Saini, V., Retzlaff, B., Roane, H. S., & Piazza, C. C. (2021). Identifying and enhancing the effectiveness of positive reinforcement. In W. W. Fisher, C. C. Piazza, & H. S. Roane (Eds.), Handbook of
applied behavior analysis (2nd ed., pp. 175–192). The Guilford Press.

Taravella, C. C., Lerman, D. C., Contrucci, S. A., & Roane, H. S. (2000). Further evaluation of low-ranked items in stimulus-choice preference assessments. Journal of Applied Behavior Analysis,
33(1), 105–108. https://doi.org/10.1901/jaba.2000.33-105

Vollmer, T. R., & Iwata, B. A. (1991). Establishing operations and reinforcement effects. Journal of Applied Behavior Analysis, 24(2), 279–291. https://doi.org/10.1901/jaba.1991.24-279