Current Issue

Human behavior is not merely a product of simple stimulus-response mechanisms, but rather is shaped by complex interactions between motivational states and environmental contexts. Particularly in educational contexts, understanding ‘why’ learners’ behaviors do not occur is essential for establishing effective intervention strategies. From this perspective, the concept of Motivating Operations has emerged as an important theoretical framework (Michael, 1993).

Motivating operations (MO) are defined as environmental variables that temporarily alter the reinforcing effectiveness of specific stimuli and regulate the likelihood of occurrence of related behaviors. In other words, they serve as antecedent conditions that ‘make behaviors happen’ rather than simply reinforcing behaviors, making them a core concept in behavioral intervention and instructional strategies (Laraway et al., 2003). MO can be broadly categorized into unconditioned motivating operations (UMO), based on biological states such as hunger and sleep deprivation, and conditioned motivating operations (CMO), formed through learning (Martin & Pear, 2011). In educational settings, strategies that intentionally manipulate these MOs to induce spontaneous behaviors in learners are widely utilized and have proven to be effective tools for eliciting functional communicative behaviors (Sundberg & Michael, 2001). While MO application is possible for all learners, its importance is particularly emphasized for children with developmental disabilities. Children with developmental disabilities typically show developmental delays in language, cognitive, and social domains, and natural behavior induction through intrinsic motivation or social stimuli is often difficult. While typical children can be sufficiently motivated by peer attention or teacher praise, children with developmental disabilities tend to be less responsive to such natural reinforcing stimuli (Koegel et al., 1997). Therefore, for these children, it is essential to artificially assign value to reinforcers through clear motivational manipulation and induce desired behaviors.

Indeed, numerous studies have reported increases in spontaneous requesting (mand) behaviors or information-seeking behaviors in children with developmental disabilities after utilizing MO (Betz et al., 2010; Roy-Wsiaki et al., 2010; Sundberg et al., 2017). This approach goes beyond simple temporary behavior induction, enabling children to recognize that they can control their environment through language, which can lead to expanded self-efficacy and social participation.

The motivational operations described above are antecedent stimulus changes or antecedent stimulus conditions that alter the reinforcing effects of other stimuli and subsequently change some dimension of responses or members of response classes associated with stimuli that establish or abolish the value of reinforcers (Belfiore et al., 2016; Laraway et al., 2003; McGill, 1999; Michael, 2000). Conditioned motivating operations are broadly classified into reflexive conditioned motivating operations (CMO-R), surrogate conditioned motivating operations (CMO-S), and transitive conditioned motivating operations (CMO-T) (Shin & Kim, 2022). CMO-R are environmental conditions that increase the value of conditioned negative reinforcement, eliciting behaviors that reduce or eliminate currently unpleasant situations. CMO-S are previously neutral stimuli that have acquired motivating operation effects by pairing with unconditioned stimuli (Chung et al., 2015), where subjects possess motivating operations that can cause behaviors when stimuli are present (Shin & Kim, 2022). CMO-T is an effective means for acquiring mand skills in children lacking verbal behavior, where one stimulus momentarily changes the reinforcing value of the next stimulus (Michael, 1988, 1993). In other words, in training mands that occur under functional control related to deprivation states or aversive stimuli, it is applied or adapted as an effective means to generate and enhance children’s needs, enabling them to produce verbal behavior and obtain desired reinforcers (Chung et al., 2017). Research on instructional strategies utilizing MO has been progressing. However, existing studies mostly rely on single-subject designs and show considerable heterogeneity in research methodology, subject characteristics, and measurement criteria. Particularly, comprehensive analysis is lacking regarding which types of MO are most effective for which behaviors, and which strategies are appropriate according to specific child characteristics. Additionally, verification of whether behaviors induced through MO manipulation generalize across various environments and people (e.g., parents, peers, teachers) and maintain over time has not been sufficiently conducted, which may serve as a limitation in actual educational field applications. Therefore, to objectively verify the practical effects of instructional strategies utilizing MO, it is necessary to systematically integrate and analyze accumulated research results rather than remaining limited to individual single-case studies. This approach can present the effectiveness of specific strategies in a more generalizable manner and is useful for supplementing limitations due to diversity in research methods and subjects (Glass, 1976; Oh, 2002). Meta-analysis, which comprehensively analyzes studies conducted on the same research topic to determine effect size trends, serves as an objective and comprehensive analytical method for identifying evidence-based practices by synthetically analyzing previous research on specific topics (Oh, 2002).

Recently, as the importance of evidence-based practice (EBP) has been emphasized in special education research, the need for establishing quality indicator standards for single-subject research has emerged. Horner et al. (2005) proposed quality indicators to ensure research quality and emphasized the necessity of conducting high-quality research that meets these criteria (Kim & Na, 2010). Subsequently, Jitendra et al. (2011) sought to supplement inconsistencies that appeared in the application process of Horner et al.’s (2005) indicators by specifying indicator descriptions and presenting a 3-point evaluation scale to confirm application validity (Kim, 2015). Additionally, previous research demonstrating that rigorously evaluated EBP interventions provide positive effects for children (Kim & Na, 2010) supports the importance of interventions based on scientific evidence. Therefore, the purpose of this study is to systematically review domestic and international single-subject studies related to motivating operations targeting students with autism spectrum disorder (ASD), and to evaluate EBP levels by applying a 3-point evaluation scale based on Horner et al.’s (2005) quality indicators and Jitendra et al. (2011) evaluation criteria. Through this, we aim to comprehensively synthesize the effects of behavioral interventions and instructional strategies utilizing MO through meta-analysis methods, thereby exploring effective intervention approaches based on scientific evidence for educational settings and providing practical assistance.

To achieve the research purpose, this study aims to systematically synthesize the effects of behavioral interventions and instructional strategies utilizing MO through meta-analysis methods. The specific research questions are as follows:

Fist, What are the research characteristics of studies targeting children with autism spectrum disorders?

Second, What are the moderating variables for the effects of motivating operations training targeting children with autism spectrum disorders?

Third, What are the quality indicators for motivating operations interventions targeting children with autism spectrum disorders?

In this study, a meta-analysis was conducted on the intervention effects of motivating operations (MOs) for children with autism spectrum disorder (ASD), using both domestic and international research.

First, a total of 31 participants were included across the analyzed studies, consisting of 25 boys (80.7%) and 6 girls (19.4%). In terms of school level, preschool was the most common setting with 9 studies (69.2%), followed by elementary school (2 studies, 15.4%), middle school (1 study, 7.7%), and preschool combined with elementary school (1 study, 7.7%). Regarding age distribution, 8 participants were aged 3 years (25.8%), 8 were aged 4 years (25.8%), 7 were aged 5 years (22.6%), 5 were aged 6 years (16.1%), 2 were aged 7 years (6.5%), and 1 was aged 14 years (3.3%). This supports previous findings that the earlier intensive interventions are initiated for children with ASD, the more positive the prognosis (National Research Council, 2001). Lovaas (1987) similarly reported that children who received early intensive interventions showed not only greater functional improvement but also successful integration compared to those receiving minimal intervention.

With respect to interventionists, most studies were conducted directly by researchers, many of whom were specialists in behavior modification working as therapists in clinical settings or as practitioners in the field. This suggests that the interventions closely reflect practical approaches to addressing problem behaviors in classrooms or schools. However, in contrast to Park and Shim (2021), who emphasized team collaboration (parents, researchers, and therapists) for the reduction and maintenance of problem behaviors, only 2 studies (14.4%) in the present analysis involved such collaborative efforts. This indicates a need for further studies adopting a team-based approach. Supporting this, Park and Kim (2022) reported that 58.9% of ASD students in Korea are educated in inclusive or special classes, and that their problem behaviors in school contexts often require intensive interventions. Considering that most participants in the present analysis were preschoolers, the findings suggest that motivating operation programs targeting mands were effectively implemented across diverse settings such as clinics, classrooms, homes, and special classes.

Second, analysis of single-subject research designs showed that the multiple probe design across behaviors was the most frequently used, reported in 4 studies (31.1%). This was followed by multiple baseline designs across participants (3 studies, 23.1%), multiple baseline designs across behaviors (3 studies, 23.1%), multielement designs (2 studies, 15.4%), multiple probe design across participants (1 study, 7.7%), and multiple baseline design across settings (1 study, 7.7%). The multiple probe design across behaviors is particularly suitable for examining intervention effects across diverse participants, behaviors, and contexts, while addressing issues of time and resource constraints inherent in continuous baseline data collection. This design also allows researchers to manage participant reactivity and ensures baseline stability assumptions (Lee et al., 2000), making it effective for evaluating the functional impact of independent variables (Kim & Na, 2010).

Third, regarding the types of motivating operations, CMO-T (transitive conditioned motivating operation) was the most frequently applied, used in 10 studies. This reflects the tendency of children with ASD to have limited or situationally specific natural motivation for social interaction and communication (Charlop et al., 2018), which can hinder the acquisition and generalization of new communication skills. CMO-T is a form of conditioned motivating operation in which specific stimuli or situations render other stimuli reinforcing and evoke behaviors that enable access to them (Michael, 1993)

In ASD interventions, the frequent use of CMO-T can be explained as follows. First, it creates artificial motivational contexts for children with limited natural motivation, thereby promoting communicative behaviors. Second, CMO-T provides functional communication opportunities by prompting children to use information-seeking mands such as “where” or “what” to obtain desired reinforcers (Endicott & Higbee, 2007; Shillingsburg et al., 2011). Third, applying CMO-T across diverse situations and contexts increases the likelihood of generalization and maintenance of acquired communication skills. Fourth, because reinforcers are provided immediately, the association between behavior and consequence is strengthened, thereby enhancing learning motivation. Finally, beyond simple requests for tangible items, CMO-T contributes to the development of cognitive and social skills such as problem-solving and higher-order question-asking (Marion et al., 2012).

Fourth, in order to examine the effectiveness of motivating operation (MO)-based interventions for children with autism spectrum disorder (ASD), a meta-analysis was conducted using a random-effects model. A total of 13 studies (k=31) were included, and the overall effect size was Hedges’g=5.12 (95% CI [2.81, 7.42]), indicating a very large and statistically significant effect (p<.001). This finding suggests that MO-based interventions exert a substantial and powerful influence on improving target behaviors in children with ASD (Marion et al., 2012; Michael, 1993).

The test of heterogeneity revealed a very high level of variability, with I²=94.85% (Q=232.92, p<.001), indicating substantial heterogeneity in effect sizes across studies. Such high heterogeneity is likely attributable not only to methodological differences — such as participants’ age, type of intervention, measurement tools, intervention duration, and frequency — but also to the diversity of individual characteristics among children with ASD (Endicott & Higbee, 2007; Shillingsburg et al., 2011). In particular, the Tau² value was estimated at 14.70, suggesting a considerable variance in effect sizes across studies. This indicates that although the overall effects of motivating operation (MO)-based interventions are large, the specific magnitude of effects may vary significantly depending on the context and conditions of individual studies (Viechtbauer, 2010).

The findings of this study support the conclusion that MO-based procedures, particularly the use of transitive conditioned motivating operations (CMO-T), can serve as effective strategies for enhancing communication and social behaviors in children with ASD (Endicott & Higbee, 2007; Marion et al., 2012). For this reason, CMO-T has been widely utilized as a core procedure in language interventions for ASD, especially in training mands for information, with consistent evidence of effectiveness across studies. However, given the high heterogeneity, future research should more clearly report intervention conditions and participant characteristics, and apply standardized procedures to reduce variability in effect sizes. Furthermore, to verify whether the large effect sizes identified through meta-analysis are replicable in practice, additional studies are needed across diverse settings and contexts.

In addition, the present study examined the quality of single-subject research conducted with children with ASD by applying the quality indicators proposed by Horner et al. (2005) and Maggin et al. (2014). Thirteen studies were reviewed, and each indicator was rated on a 3-point scale (1=not met, 2=partially met, 3=met). Overall, most indicators demonstrated a high level of adherence. In particular, fundamental elements of research reporting — such as participant descriptions (age, gender, disability type), description of settings, reliability of dependent variable measurement, description and implementation of independent variables, description of baseline conditions, internal validity, outcome reporting, replication of effects, and performance characteristics of the independent variable — were fully satisfied (100%) across all studies. These results suggest that the analyzed studies faithfully adhered to essential design and reporting standards.

A closer examination revealed that all studies clearly reported participant characteristics such as age, gender, and type of disability. Similarly, contextual descriptions, including intervention settings, locations, and interventionists, were presented in detail. In addition, the reliability of dependent variable measurement, procedures for manipulating independent variables, description of baseline phases, assurance of internal validity, clarity of outcome reporting, and replication of results were all rated as fully satisfied (score=3) across studies. This suggests that single-case design research involving children with ASD maintained a strong commitment to replicability and transparency throughout the research process.

However, some indicators demonstrated relatively lower levels of fulfillment. Participant selection criteria were fully satisfied in only 76.9% of the studies, with the remaining 23.1% rated as partially satisfied. This indicates that certain studies lacked sufficient detail regarding selection processes, diagnostic procedures, or inclusion and exclusion criteria. Similarly, descriptions of dependent variables (84.6% fully satisfied, 15.4% partially satisfied) and reporting of measurement validity (84.6% fully satisfied, 15.4% partially satisfied) were sometimes insufficient in providing operational definitions or evidence supporting the validity of measurement tools.

Fidelity of implementation was another area of concern, with 76.9% of studies fully meeting the criterion, while 23.1% failed to address it. Fidelity is a critical factor that verifies whether an intervention was carried out as intended, directly influencing the interpretation and replicability of intervention effects (Noell et al., 2005). In several studies, procedures or results related to fidelity were not reported, limiting the ability to evaluate the consistency of implementation and its applicability in real-world contexts.

The practicality and inefficiency reporting of independent variables also emerged as an important area for improvement. Practicality reflects whether an intervention is feasible in educational or clinical settings, and whether it may be ineffective under certain conditions. Yet, 38.5% of studies only partially satisfied this criterion, suggesting a need to strengthen reporting to enhance practical validity.

Regarding social validity, while descriptions of dependent variable importance, magnitude of change, and characteristics of independent variable implementation were largely satisfied, some studies lacked sufficient discussion about whether the observed behavioral changes were meaningful in real-life contexts. Social validity is critical in ASD intervention research, as it ensures that outcomes are not only statistically significant but also lead to positive and functional changes in everyday life (Horner et al., 2005).

In sum, the results of this quality indicator analysis demonstrate that single-case studies involving children with ASD generally maintain high reporting quality. Nonetheless, several domains — including fidelity of implementation, practicality/inefficiency of independent variables, participant selection criteria, and measurement validity — require more explicit and detailed reporting. Future research should systematically evaluate quality indicator adherence and align research design and reporting practices with internationally recognized standards. By adopting methodological rigor as emphasized by Maggin et al. (2014) and adhering to evidence-based practice criteria proposed by Horner et al. (2005), the quality of ASD intervention research can be further enhanced. Such efforts will ensure that research findings are more effectively translated into educational and clinical practice, ultimately contributing to improved quality of life for children with ASD.

Based on the conclusions and discussion drawn from this study, the following limitations and suggestions for future research on motivating operations (MOs) for children with autism spectrum disorder (ASD) can be identified:

First, this study analyzed 13 single-case design studies to evaluate quality indicators. However, the sample size was limited, and the inclusion and exclusion criteria during the selection process were not applied with sufficient rigor. In particular, the analyzed studies were heavily concentrated in certain countries and researcher groups, making it difficult to generalize the findings across diverse cultural contexts and environments. Future research should incorporate studies conducted in a wider range of cultural and linguistic settings to more comprehensively examine both the effects of MO-based interventions and the quality of reporting.

Second, the quality indicator analysis was conducted solely on the basis of information reported in published research articles. As such, details not explicitly stated in the reports or contextual variables in actual intervention settings could not be fully reflected. For example, items such as fidelity of implementation and the practicality or inefficiency of independent variables may have been carried out appropriately in practice, but were rated as “not met” or “partially met” due to insufficient reporting. Future research should therefore consider using multiple sources of information — such as researcher interviews, observational records of the intervention process, and reviews of raw data — to improve the accuracy of quality assessments.

Third, while this study focused on analyzing the extent to which quality indicators were satisfied in MO-related interventions, it did not examine in depth how specific types of motivating operations (e.g., CMO-T, CMO-S, CMO-R), variations in intervention procedures, or participant characteristics (e.g., developmental level, communication abilities) might influence both effectiveness and reporting quality. Future studies should conduct subgroup analyses by MO type and participant characteristics to determine under what conditions specific MO strategies are most effective. Such work will contribute to a more nuanced understanding of motivating operations and provide clearer guidance for practice.