Teacher-Student Interactions of Autistic Adolescents: Relationships between Teacher Autonomy Support, Structure, Involvement and Student Engagement

Research has demonstrated that autistic students are less engaged with classroom tasks and in the school setting (Jahromi et al., 2013; Roorda et al., 2021; Zajic et al., 2020). One predictor of their engagement is the quality of the teacher-student relationship (Losh et al., 2022; Roorda et al., 2021). Unfortunately, autistic students in primary education have a high risk of developing poorer teacher-student relationships (Blacher et al., 2014; Caplan et al., 2016; Feldman et al., 2019; Zee et al., 2020). This is worrisome because these studies were conducted at the primary level, where the teacher-student bond is thought to be stronger than at the secondary level, which often has rotating rather than permanent teachers (Breault, 2013). Whether poorer teacher-student relationships are a recurring phenomenon for autistic students during their school years remains to be investigated, as few studies have focused on autistic adolescents in secondary education (Hill, 2014), and even fewer have used observations to analyze the teacher-student interaction as it occurs in real time (Esqueda Villegas et al., 2024). The present study therefore looked into the observable interaction dynamics of autistic adolescents and their teachers, using the theoretical framework of Self-Determination Theory (SDT).

SDT establishes that all individuals have three basic psychological needs for autonomy, competence, and relatedness (Ryan & Deci, 2000, 2018). Teachers can help to fulfill these needs in the classroom context by providing autonomy-support, structure, and showing involvement during their lessons and in the interactions with their students (Snikkers-Mommer et al., 2024). Autonomy-supportive teachers acknowledge students’ needs and interests (Reeve & Jang, 2006); they provide students with choices and implement classroom activities that are meaningful (Jang et al., 2010; Ryan & Deci, 2018). Teachers can foster structure by providing step-by-step guidance to their students during a lesson, communicating clearly what is expected from them, and giving constructive feedback (Jang et al., 2010; Skinner & Belmont, 1993). Additionally, teachers can offer structure when they present an overview of the lesson organization and build on previous lessons to introduce new knowledge (Maulana et al., 2012). To demonstrate involvement, teachers can show responsiveness by listening actively, showing support, and taking time to address students’ questions (Ryan & Deci, 2018). The assumption is that when the three basic psychological needs are met, students will feel engaged in their learning (Ryan & Deci, 2018). Notably, previous studies suggest that each component (of need-supportive teaching) contributes uniquely to students’ engagement (Jang et al., 2010; Loopers et al., 2024; Maulana et al., 2013; Snikkers-Mommer et al., 2024).

Some teaching methods or classroom strategies can be counterproductive and thwart students’ need for autonomy, competence, and relatedness. For example, teachers may thwart students’ autonomy by using controlling language, or by giving commands about what students “must” do during the lesson (Reeve & Jang, 2006). In addition, they may give unclear or vague instructions and be unavailable to answer students’ questions due to various reasons (Stroet, 2014), which could be considered as chaos. Lastly, teachers can display disaffection when they address students in an unfriendly tone or fail to show commitment and support (Ryan & Deci, 2018; Stroet, 2014). All of these teacher behaviors can have a negative impact on student engagement.

Despite the literature emphasizing the relevance of both SDT and the teacher-student interaction to foster students’ engagement, the extent to which autistic students are provided with autonomy-support, structure, and involvement during the teacher-student interaction has been under-researched. However, the three dimensions of SDT have been (indirectly) addressed by a few studies. In terms of autonomy, research indicates that teachers perceive (inclusive) education as a means through which autistic students can become more autonomous and independent (Cassimos et al., 2015). In a more recent study, Tesfaye et al. (2023) found that autistic youth (aged 11–18 years) valued making their own decisions, particularly if these concern them directly. Yet, it seems like schools rarely provide opportunities for autistic students to experience autonomy, although this might increase their motivation and engagement (Heyworth et al., 2021; Shea et al., 2013). Regarding structure, this type of need-support is highly appreciated by autistic students (Saggers et al., 2011). Without guidance and clear instructions on what they have to do, autistic students have reported to experience ‘mental breakdowns’ and feel ‘panicked’ or ‘stressed’ in secondary schools (Esqueda Villegas et al., submitted). In terms of involvement, autistic students value teachers who are kind, check up on them throughout the day, show understanding and are active listeners (Saggers et al., 2011; Sproston et al., 2017; Stephenson et al., 2021). Lastly, these students value teachers who recognize their needs and care about establishing a personal connection (Goodall, 2019; Humphrey & Lewis, 2008). In short, although most studies do not observe what happens during teacher-student interactions in the classroom, previous research does underscore autistic students’ need for a classroom context that provides autonomy support, structure, and involvement. To capture what occurs during real-time classroom interactions in terms of teachers’ need-support and student engagement, we propose the teacher-(autistic) student interaction as a complex dynamic system.

The present study used classroom observations and State Space Grids to analyze the interaction dynamics between teachers and autistic students in secondary education in the Netherlands and Mexico. Overall, these interactions were highly variable, as the dispersion values of most of the teacher-student interactions (across all dimensions) were between 0.6 and 0.9. This is in line with previous research indicating that the teacher-student interactions (of typically developing students) tend to move in-and-out from one state into another throughout a lesson (Smit et al., 2022).

In terms of autonomy support, we observed that students were provided with few opportunities to make choices. This is in line with findings from Heyworth et al. (2021) and Shea et al. (2013), who reported that the school environment does not often take into account the (autistic) students’ needs and interests. Previous studies with non-autistic participants have found autonomy-support to be a crucial factor influencing engagement (Snikkers-Mommer et al., 2024). Although teachers cannot be expected to be in an autonomy-supportive state all the time, our results seem to indicate that this would be a good addition to their repertoire of instructional strategies, as more desirable engagement states occurred in students when autonomy support was provided. However, teachers’ opportunities to provide autonomy support may be affected by specific barriers within different educational contexts, such as curriculum constraints (Emam & Farrell, 2009). Future studies should explore these contextual barriers in greater depth, to better understand the challenges teachers face and to develop targeted, practical recommendations that support the implementation of autonomy-supportive practices across diverse educational environments.

In terms of structure, many teachers in this study often provided step-by-step guidance to their students. We frequently observed that this co-occurred with students being engaged with a task or listening. In the teacher-student pair with a high proportion of structure, we also observed that students showed more engagement compared to the teacher-student pair with low levels of structure, in which the interactions oscillated between engaged and disengaged states. Interestingly, we noted that some dyads had a disengaged student state regardless of teachers’ frequent provision of structure. This counterintuitive finding may be explained by the fact that some teachers implement structure in a more controlling manner (Domen et al., 2020), or it may simply be due to the fact that autistic students have difficulty staying engaged for long periods of time (Zajic et al., 2020).

In terms of involvement, teachers displayed low levels of warmth and affection during their interactions. Regardless of whether students were provided with high or low proportions of involvement, the teacher-student interactions fluctuated between engaged and disengaged student states.

Even though our coding conventions were the same for all lessons, notably, the subject of a lesson (for example, Math, English, Sociology) can influence teacher-student interactions and the need-supportive strategies used in the moment (Maulana et al., 2012). For example, the highly structured Physics lesson from teacher ‘Rafael’ in the Mexican context—focused on problem-solving—offered frequent step-by-step guidance, resulting in greater student engagement. In contrast, the Sociology lesson from teacher ‘Cindy’ discussing societal issues, which did not have a hands-on task, involved less structure and corresponded with higher levels of student disengagement. A lesson that relies solely on listening and paying attention, without hands-on tasks may be problematic, as some autistic learners struggle to process large amounts of oral information (Carrington et al., 2003). Similarly, certain teaching activities may limit the extent of need-support teachers can provide to students. To prevent disengagement in (inherently) less-structured lessons and activities, it is highly recommended to provide visual support, such as video explanations, images, and pictograms. These supports can help autistic students better understand and identify key information about a topic (Esqueda Villegas et al., submitted).

Last, although the educational systems of the Netherlands and Mexico seemed rather different, we found many similarities in the interactional patterns. First, there was a tendency for teachers (in both countries) to be more controlling than autonomy supportive. This is worrisome given that autonomy-thwart can cause students’ disengagement (Cents-Boonstra et al., 2021; Ryan & Deci, 2018). Second, for many teacher-student pairs the most common interaction pattern was teacher providing ‘Structure’ and an engaged student state, such as ‘Listening’. Indeed, this may be beneficial to many autistic students who have strong preference for predictable and structured environments (Saggers et al., 2011). A third dominant pattern across the Dutch and Mexican teachers was to display few levels of involvement and remain more ‘Neutral’ during the teacher-student interaction. Similar to the autonomy dimension, although it is not expected for teachers to show involvement during the whole lesson, interactions that contain relatively high levels of involvement have more engaged student states. Therefore, these findings give an indication on how to improve students’ engagement and well-being.

This is one of the first studies using real-time classroom observations and State Space Grids to analyze the teacher-(autistic) student interaction in secondary education. These types of studies are scarce (Esqueda Villegas et al., 2024), but provide many clues to the current and in-the-moment learning experiences of autistic students. We therefore encourage researchers in other countries to move beyond self-reports and questionnaires when studying teacher- (autistic) student interactions and instead incorporate classroom-based observational methods.

Our results show that there is great variation in the interaction between teachers and autistic students in secondary schools. However, Dutch and Mexican teachers seem to miss opportunities to provide explicit autonomy support; this contextual factor is minimally present in teacher-student interactions. This is consistent with previous research suggesting that the amount of autonomy support provided by teachers is not at a desirable level (Snikkers-Mommer et al., 2024). It was more common, however, for teachers in both countries, to provide their students with structure, which co-occurred (in most cases) with more engagement of autistic students. Yet, it should be noted that this desirable interaction pattern was not observed in all teacher-student pairs. Perhaps for these students, structure seems implemented in a more controlling form (Domen et al., 2020), or may simply not match their learning preferences. Since interaction patterns tend to stabilize across the school year, particular attention needs to be paid to disengaged student states that may be harder to break.