Investing in preschool children's language development is critical to adressing educational inequality (Cunningham et al., 2019; Degotardi & Gill, 2017). However, opportunities for interaction are not shared equally (Black, 2004). Teachers’ interactional behaviour is strongly influenced by their perceptions of children (Black, 2004) and the pupil’s social background (Peleman et al., 2020).  A narrative review study on language learning across early childhood education (ECE) shows that overheard speech that is not directed to the child does not contribute effectively to children’s language development (Rowe & Snow, 2020). Direct interaction between teacher and child with the full attention of the teacher is therefore a critical factor in language development (Weisleder & Fernald, 2013). The study of teacher attention is an emerging field in educational research, due to the innovative technique of mobile eye tracking. These studies have shown that teacher gaze is not evenly distributed across pupils (İnan-Kaya & Rubie-Davies, 2022; Smidekova et al., 2020). In particular, novice teachers tend to give their undivided attention to a limited number of pupils rather than to all children in the classroom (Cortina et al., 2015; Dessus et al., 2016). Explanations for the uneven distribution are inconsistent. For example, Chaudhuri and colleagues (2022) found that teachers focused primarily on the lowest-achieving pupils in their classrooms, while Dessus and colleagues (2016) found that primary teachers focused significantly more on a 'steering group' consisting primarily of middle- and high-achieving pupils.

Mobile eye tracking has also been used to investigate unconscious mechanisms of inequality.  While teachers can largely control their verbal messages, they cannot control their non-verbal communication in the same way (İnan-Kaya & Rubie-Davies, 2022). For instance, eye tracking has been used in previous research to measure teachers' implicit biases, including in early childhood settings (Gilliam et al., 2016; İnan-Kaya & Rubie-Davies, 2022).  We hypothesise that implicit bias also influences the teacher’s gaze, given that previous mobile eye tracking research by Gilliam and colleagues (2016) has shown that teachers’ implicit biases based on race and gender influenced their eye gaze.

In attempting to explain teacher attention, a distinction can be made between intentional top-down processes, such as intentionally looking at a particular child during a particular exercise to gather information about their learning processes, and unintentional bottom-up processes, such as student behaviour that disrupts classroom activities (Goldberg et al., 2021; Theeuwes et al., 2000), or unconscious teacher mechanisms, such as teacher bias (Gilliam et al., 2016). Theeuwes and colleagues (2000) write that the intentional top-down processes only secondarily influence the direction of attention. Unintentional processes are of bigger influence.

This paper presents a single case study that examines a preschool teacher's distribution of attention and interaction opportunities during two moments of language learning, with a dual aim. First, we want to investigate the teacher's attentional distribution in the context of language learning in early childhood education, with the aim of unravelling the unconscious mechanisms of inequality at the teacher level. Second, we sought to examine the ecological validity of mobile eye tracking by triangulating eye tracking metrics with data collected from alternative sources.

RQ1.   How is the teacher attention distributed among preschool children during a formal and an informal language learning activity?

RQ2.   How does teacher attention relate to the quality and quantity of teacher-child interactions?

RQ3.   What influences the teacher’s attention during language learning, as perceived by the teacher?

RQ4.   How do the eye tracking metrics relate to the attentional distribution in a real classroom?

This study uses a data triangulation approach to realise an explanatory sequential mixed methods single case study design (Onghena et al., 2019). The school was selected on the basis two criteria: being located in an at-risk neighbourhood, which is characterised by socio-economic and linguistic diversity, and having a social mix in the school. Within the classroom four focal children were selected based on two criteria: age and language background. Two monolingual and two multilingual 4-year-olds were randomly selected (n=4). In order to describe these pupils, the teacher’s perceptions and expectations were investigated. The teacher was asked to describe the focal children based on three dimensions: expectations about language development, perceptions about pupils' sense of belonging (Laevers & Heylen, 2013), and expectations about pupils' social skills (Cassidy & Asher, 1992). Mobile eye tracking was used to answer RQ1, using the Tobii Pro Glasses 2 with a one-point calibration system and a data rate of 50 Hz. The four children are individually identified as the teacher's areas of interest (AOIs). Two classroom activities were videotaped, in order to capture both formal and informal language learning: interactive book reading and fruit eating. To answer RQ2, these activities were transcribed, and the interactions were coded and analysed. In RQ3, the eye tracking metrics are accompanied by a stimulated recall interview (SRI) with the teacher. The teacher was asked to watch her own video recordings made by the glasses immediately after the eye tracking data collection. The purpose of the SRI was to explore why the teacher's attention was drawn to certain children at certain times and to give deeper meaning to the eye tracking data . To answer RQ4, video observations were conducted over two school days. Eye-tracking analysis software, Imotions, was used to perform fixation mapping in combination with manual mapping by the researcher. Raw eye tracking metrics are reported (RQ1), such as dwell time, which represents the number of seconds the teacher focuses on the child. Video recordings of language learning moments were transcribed verbatim and coded using a literature-based coding scheme (RQ2) (Justice et al., 2018; Tsybina et al., 2006; Vanparys et al., 2023; Verhallen & Walst, 2011). A qualitative content analysis was conducted to analyse the SRI (RQ3). To investigate the ecological validity (RQ4), the real-classroom video observations were coded using the coding scheme described above.

Eye tracking revealed an uneven distribution of teacher attention (Chaudhuri et al., 2022; Dessus et al., 2016; Haataja et al., 2021). The data show a complex picture of what influenced teacher attention. At the centre of this picture is the child whose initial language skills, courage to speak and teacher’s perceptions and expectations all contribute to uneven teacher behaviour. Connections emerged between the quality of interactions and the distribution of attention. The quality and quantity of interactions, measured by the number of strategies such as open and closed questions, recasts or expansions are related to the dwell time. Results of the interview data suggest an explanation for the uneven distribution. Intentional processes, such as pursuing a learning goal, and unintentional processes, such as responding to a child-initiated interaction, could be identified (Goldberg et al., 2021; Theeuwes et al., 2000). Cross-coding revealed a contrast in the use of intentional and unintentional processes. Unintentional processes were more frequently used to explain the focus on the proficient child, whereas intentional processes were used to explain the focus on the less proficient children. This suggests a conscious effort to regulate conversations and achieve a balanced distribution of attention. However, in line with previous research (Theeuwes, 2010), the results show that unintentional processes may direct attention more than intentional processes. This highlights the need for teachers to become aware of attentional processes and to promote awareness of inequalities that teachers may be unconsciously contributing to (Breese et al., 2023). This single case study provides valuable insights into the underlying mechanisms that contribute to unequal language development opportunities in ECE. The similarity of results between mobile eye tracking and real classroom observations, suggests that eye tracking is an ecologically valid data collection method that can be used to investigate teacher attention and preschoolers' opportunities for interaction.

Cunningham, J. E., Zimmerman, K. N., Ledford, J. R., & Kaiser, A. P. (2019). Comparison of measurement systems for collecting teacher language data in early childhood settings. Early Childhood Research Quarterly, 49, 164–174. https://doi.org/10.1016/j.ecresq.2019.06.008 Black, L. (2004). Differential participation in whole-class discussions and the construction of marginalised identities. Journal of Educational Enquiry, 5(1), 34–54. Peleman, B., Vandenbroeck, M., & Van Avermaet, P. (2020). Early learning opportunities for children at risk of social exclusion. Opening the black box of preschool practice. European Early Childhood Education Research Journal, 28(1), 21–42. https://doi.org/10.1080/1350293X.2020.1707360 Rowe, M. L., & Snow, C. E. (2020). Analyzing input quality along three dimensions: Interactive, linguistic, and conceptual. Journal of Child Language, 47(1), 5–21. https://doi.org/10.1017/S0305000919000655 İnan-Kaya, G., & Rubie-Davies, C. M. (2022). Teacher classroom interactions and behaviours: Indications of bias. Learning and Instruction, 78(101516), 1–13. https://doi.org/10.1016/j.learninstruc.2021.101516 Smidekova, Z., Janik, M., Minarikova, E., & Holmqvist, K. (2020). Teachers’ gaze over space and time in a real-world classroom. Journal of Eye Movement Research, 13(4). https://doi.org/10.16910/jemr.13.4.1 Cortina, K. S., Miller, K. F., McKenzie, R., & Epstein, A. (2015). Where Low and High Inference Data Converge: Validation of CLASS Assessment of Mathematics Instruction Using Mobile Eye Tracking with Expert and Novice Teachers. International Journal of Science and Mathematics Education, 13(2), 389–403. https://doi.org/10.1007/s10763-014-9610-5 Cortina, K. S., Miller, K. F., McKenzie, R., & Epstein, A. (2015). Where Low and High Inference Data Converge: Validation of CLASS Assessment of Mathematics Instruction Using Mobile Eye Tracking with Expert and Novice Teachers. International Journal of Science and Mathematics Education, 13(2), 389–403. https://doi.org/10.1007/s10763-014-9610-5 Gilliam, W. S., Maupin, A. N., Reyes, C. R., Accavitti, M., & Shic, F. (2016). Do Early Educators’ Implicit Biases Regarding Sex and Race Relate to Behavior Expectations and Recommendations of Preschool Expulsions and Suspensions? Yale University Child Study Center. Goldberg, P., Schwerter, J., Seidel, T., Müller, K., & Stürmer, K. (2021). How does learners’ behavior attract preservice teachers’ attention during teaching? Teaching and Teacher Education, 97, 103213. https://doi.org/10.1016/j.tate.2020.103213 Theeuwes, J. (2010). Top–down and bottom–up control of visual selection. Acta Psychologica, 135(2), 77–99. https://doi.org/10.1016/j.actpsy.2010.02.006 Chaudhuri, S., Muhonen, H., Pakarinen, E., & Lerkkanen, M.-K. (2022). Teachers’ visual focus of attention in relation to students’ basic academic skills and teachers’ individual support for students: An eye-tracking study. Learning and Individual Differences, 98, 102179. https://doi.org/10.1016/j.lindif.2022.102179