The importance of structural characteristics for interaction quality in Austrian preschools

ABSTRACT Many studies have found associations between structural characteristics and interaction quality of children in ECEC settings. However, findings are inconsistent and since previous research has been conducted outside Austria, transferability to the Austrian country context is limited. By addressing this gap, the study aims to identify relations between structural characteristics and interaction quality (measured with the Individualized Classroom Assessment Scoring System, inCLASS) of 161 three- and four-year-old children from 55 preschools in Austria. After including control variables, main findings revealed positive relations between adequate equipment and teacher interactions and negative relations between child-staff ratio, proportion of children with immigration background per preschool class, preschool teachers’ work experience and peer interactions. Preschool teachers’ job security was positively related to task orientation and poorer child-staff ratio corresponded with more conflict interactions. Higher work experience led to fewer conflict interactions. The results indicate that specific structural characteristics are associated with a higher interaction quality in preschools in Austria.


Introduction
Educational quality in preschools is often conceptualized within a framework of a 'structure-process model of quality' (Kluczniok and Roßbach 2014, 147), which includes different components of quality: the quality of interactional processes between children, peers, preschool teachers, and tasks (termed 'process quality' or 'interaction quality') (Schmidt et al. 2018); the quality of structural characteristics such as group size or childstaff ratio ('structural quality') (Slot 2018); the quality of educational beliefs; and collaboration with families (Kluczniok and Roßbach 2014;Kluczniok 2018). This paper refers to interaction quality and structural quality. Interaction quality stands out as the central component of educational quality (Tietze et al. 1998), and research has shown that the quality of children's interactions with preschool teachers, peers, and their engagement with tasks in preschools predicts the development of children's competencies (Tietze et al. 1998;NICHD ECCRN 2006;Burger 2010;Ulferts, Wolf, and Anders 2019).
Theoretical framework models assume that characteristics of structural quality (e.g. group size, staff qualification) predict interactional quality in preschools (Essa and Burnham 2001;Kluczniok and Roßbach 2014). There is a substantial body of research supporting the idea of predictive importance of structural quality for interaction quality (Vandell and Wolfe 2000;Manning et al. 2019). However, it must be pointed out that conclusions cannot simply be applied to individual countries with their specific history and conditions of preschool education. For example, Love et al. (2003) discussed the limited generalizability of findings of the US-American NICHD Study of Early Child Care to contexts in other countries. One consequence is that it is necessary to consider country-related specifics when investigating issues on educational quality in preschools.
In Austria, there have been few findings to date on educational quality in preschools (Smidt 2018;Hartel et al. 2019 for overviews; for recent studies: 2023). As in many other countries (OECD 2019b), the attendance rates for 3-5-year-old children are high in Austria at 92.6% (Statistik Austria 2021). However, the Austrian preschool system is also characterized by some peculiarities, such as a comparatively low attendance rate for children under three years old in institutional day care (OECD 2019b) and the non-academic professional qualification of pedagogical preschool staff (Hartel et al. 2019). To achieve reliable conclusions, scholarly work indicates that connections between structural characteristics and interaction quality should be regarded in light of countryspecific and context-specific conditions, which the present study applies for the case of Austria.

The Austrian preschool system
The preschool system in Austria is regulated by the nine federal states in terms of personnel, structures, and technical supervision, while the training of pedagogues and guidelines on interaction quality is primarily the responsibility of the federal government (Hartel et al. 2019). The attendance rates for the three to five-year-old children are around 93% (Statistik Austria 2021), and attendance is compulsory for the final year of preschool (Smidt 2018). A preschool teacher and at least half an assistant are responsible for up to 20-25 children, and staff-child ratios usually range between 1:12 and 1:17. This does not meet international recommendations because assistant staff are always included in the staff-child ratio (Hartel et al. 2019). The regular vocational qualification of the educational staff in Austria is non-academic (Eichen and Krenn-Wache 2020). Recently, first study programmes for further education in early childhood education have been established (Hartel et al. 2019;Holzinger and Reicher-Pirchegger 2020). In 2009, an educational plan with specific guiding principles (e.g. individualization) and educational domains (e.g. language and communication) was established (Ämter der Landesregierungen der österreichischen Bundesländer, Magistrat der Stadt Wien, Bundesministerium für Unterricht, Kunst und Kultur 2009). There is a pronounced distinction between preschool education and compulsory schooling (Smidt 2018).
Interaction qualitytheoretical background, conceptualization, and the importance of structural characteristics Interaction quality, which is in the focus of the present study, is often referred to as a component of a 'structure-process model of quality' (Kluczniok and Roßbach 2014, 147), that includes structural variables, which are assumed to predict interaction quality. Preschools teachers' beliefs ('orientational' (Bäumer et al. 2011, 96) quality; e.g. educational goals) and cooperation with families (e.g. counselling parents) are further components which are sometimes included (Kluczniok and Roßbach 2014;Kluczniok 2018). Interaction quality in preschools can be theoretically framed with ecosystemic approaches (Bronfenbrenner and Morris 2006;Booren, Downer, and Vitiello 2012) that emphasize the view on preschool children and preschool teachers belonging to the 'preschool class' microsystem and being involved in a wide range of activities and interactions that take place in the preschool class. In addition, interaction quality in preschools can be framed based on social-constructivist approaches related to the thoughts of Vygotsky (Winsler and Carlton 2003;Bodrova and Leong 2018), which emphasize the role of preschool teachers and peers as competent co-constructors. Domain-specific theories (Carey and Spelke 1993) are sometimes used to describe the assignment of interaction quality to specific domains such as language or math (Smidt and Roßbach 2016).
In terms of implementing 'good' quality interactions in preschools, children should be engaged in developmentally appropriate interactions and activities with their preschool teachers and peers according to best practice standards (Copple and Bredekamp 2009), and pedagogical opportunities in preschools should cover a variety of domains (e.g. language, mathematics, and social relationships). The role of preschool teachers should be supportive by enriching pedagogical tasks for children's development; supporting appropriate relationships among children, preschool teachers, and peers; and promoting healthful and safe care (Tietze et al. 1998;Cryer et al. 1999;Smidt and Roßbach 2016).
Findings on interaction quality in preschools vary according to a number of methodological aspects, such as the instrumental measurement of quality with specific foci, the level of aggregation, and the focus on children (e.g. NICHD ECCRN 2000;Pianta, La Paro, and Hamre 2008;Downer et al. 2010;Harms, Clifford, and Cryer 2015). Findings on the predictivity of structural characteristics for interaction quality differ according to the instruments used for observing and assessing interaction quality (e.g. Pianta et al. 2005;Tietze et al. 2013), as well as on dimensions of measured interaction quality (e.g. Løkken et al. 2018;Ramirez and Linberg 2022). In addition, country contexts (e.g. Cryer et al. 1999) may also influence relations between structural characteristics and interaction quality as several structural characteristics are politically regulated (Kluczniok and Roßbach 2014).
Numerous studies have found associations between structural characteristics and interaction quality of children in ECEC settings (e.g. Phillipsen et al. 1997;NICHD ECCRN 2002;Pianta et al. 2005;Tietze et al. 2013;Melhuish and Gardiner 2019). However, findings are inconsistent (Slot 2018). With Kluczniok and Roßbach (2014), it can be stated that no single variable has a very strong association with interaction quality, and small to moderate effect sizes (in terms of r = .10 to r = .30; Cohen 1988) are expectable at best. It must also be considered that several characteristics are predictive of interaction quality other than structural aspects, such as characteristics of the children, including immigration background, and language competencies (Cabell et al. 2013;Kluczniok and Schmidt 2020;Smidt, Embacher, and Kluczniok 2020;Ramirez and Linberg 2022).
The number of children per preschool group and the ratio between children and preschool staff are frequently used as structural variables. According to an overview of international studies by Vandell and Wolfe (2000), smaller group size has often been found to corresponded with better interaction quality in preschool and toddler groups. The same is true for lower child-staff ratio, (i.e. less children per staff) which is frequently associated with better interaction quality. However, for both predictors, there are also a few studies that detected no associations (Vandell and Wolfe 2000). This is even true with regard to more recent research. In a Dutch study, Slot et al. (2015) found no relations between group size, child-staff ratio, and interaction quality, which was attributed to a lack of variation in these structural variables. In the United Kingdom, Melhuish and Gardiner (2019) found that nonfamilial settings for three to four-year-old children with comparatively low staff-to-child ratios also achieved higher interaction quality. This finding is seen in connection with a high number of well-qualified staff in these groups.
However, research has also shown that relations between staff qualification and interaction quality are inconsistent. In an overview of seven studies, Early et al. (2007) found almost no relations. In a meta-analysis, Manning et al. (2019) found that higher formal education of staff corresponded with better interaction quality with small to medium effects. The proportion of children with an immigration background in institutional child care settings is also the subject of research. Findings from Germany show that lower proportions of children with an immigration background in preschool classes corresponded with better interaction quality (Kuger and Kluczniok 2008). Regarding work experience, summaries of studies by Slot (2018) and Smidt (2012) indicate that research has not been consistent, with some studies finding more work experience related to better interaction quality, and other studies reporting the opposite. These findings could be even explained with routineness and familiarity (positive relations), as well as a tendency toward burnout and stress (negative relations) (Smidt 2012).
Other occupation-related variables have been considered as important structural characteristics as well, such as contractual weekly working hours of the preschool teacher, job security, and preschool teachers' wages (OECD 2019a). The use of space and materials has also been used as predictors. In a German study, Smidt (2012) found that better equipment with adequate (learning) material corresponded with higher interaction quality in preschools, and a similar pattern emerged in an Austrian study Embacher 2020, 2023).
In Austria, there is a lack of research regarding relations between structural characteristics and interaction quality in preschool. In addition to positive relations between adequate equipment and interaction quality with teachers and tasks,  found that a higher child-staff ratio tends to be negatively related to the interaction quality with peers of three-and four-year-old children in preschools in Austria. Geißler et al. (2022) show positive relations of staffs' educational level and their years of experience with certain aspects of interaction quality of children under three years old in ECEC centres in Austria.

The current study
Previous studies highlight the importance of structural characteristics for interaction quality in preschools. However, findings regarding relations between structural characteristics and interaction quality are mixed and differ depending on the measurements used for interaction quality or country contexts. As studies have usually been conducted outside of Austria, there may be limited transferability of previous research results to the context of Austria with its specific preschool system. The aim of the present study is to address this research gap by identifying relationships between interaction quality of children in preschools in Austria and structural characteristics. The focus of the study lies in the interaction quality at the level of the individual child, which was measured with the inCLASS (Downer et al. 2010). Child characteristics were considered as control variables.

Participants
The study draws on data from the first wave of the project 'Quality of Children's Interactions in Preschool', which was funded by the Austrian Science Fund (FWF). The sample included 261 three and four-year-old children attending 81 preschools in Tyrol, a federal state of Austria. Preschools, preschool classes, and children in each preschool were randomly selected. Complete data were available for 161 children from 55 preschool classes (from 55 preschools), including 85 girls and 76 boys, who were included in the current analyses. Data collection took place from April to June 2019 at the end of the children's first year of preschool. At that time, the children were between 39.82 and 57.00 months old with a mean of 49.60 months (SD = 4.05).
A total of 14.29% of the children had an immigration background. In the case of this study, this means that they have a family language other than German, which is the official language in Austria. Overall, 55 preschool teachers participated in the study and completed a questionnaire. To ensure adherence to ethical guidelines, the preschool teachers and parents of the participating children were informed about the study prior to their participation and signed an informed consent form (parents acted as representatives of their child). Participation was voluntary, and anonymity was guaranteed throughout data processing.

Measures
Interaction quality Interaction quality at the level of the individual child was measured using the Individualized Classroom Assessment Scoring System (inCLASS; Downer et al. 2010), which was developed to assess children's competencies in interactions with teachers, peers, and tasks/learning activities in preschool classrooms (e.g. the degree to which a child seeks proximity with peers or teachers). The inCLASS consists of 10 dimensions (Downer et al. 2010, and based on previous research (Downer et al. 2010Slot and Bleses 2018;Bohlmann et al. 2019), they were assigned to the following four factors (domains): teacher interactions (α = .82, containing positive engagement with the teacher and teacher communication), peer interactions (α = .90, containing peer sociability, peer communication, and peer assertiveness), task orientation (α = .71, containing task engagement and self-reliance), and conflict interactions (α = .58, containing teacher conflict, peer conflict and behaviour control [reverse coded]) (see Embacher 2021, 2023 for further information). Descriptive results and intercorrelations are shown in Table 1.
Observations were carried out within one observational visit per preschool class. On a regular preschool morning (usually between 8 am and 12 pm) up to four children per preschool class were observed with three or four alternating observation cycles (M = 3.69; SD = 0.54) for each child. One observation cycle lasts 15 min and includes a 10minute section for observation and taking notes, followed by a 5-minute section where the level of observed interactions was rated on a seven-point scale (1-2 = low level, 3-5 = intermediate level, and 6-7 = high level) (Downer et al. 2010. Data collectors rotated between the selected children until the end of the visit to observe each child in different situations (e.g. free choice, planned and led activity). The rating scores of all cycles were averaged in each dimension to obtain the final dimension scores for each child.
Data were collected by 14 observers (students of educational science and psychology) who completed a two-day training by a certified inCLASS trainer. The training included information regarding the dimensions and domains of the inCLASS and practical guidance for usage of the inCLASS manual. At the end of the training, all observers had to code five reliability clips independently and to score within one point of a master code on 80% of their scores to pass the reliability test (Downer et al. 2010). To examine inter-rater reliability, a total of 47 observation cycles (7.91% of the observations) were double coded by two observers who independently observed and rated the same children. Intraclass correlation coefficients (ICCs) of the single domains ranged between .92 and .96, indicating excellent inter-rater agreement (Cicchetti 1994).

Structural characteristics
Paper questionnaires were administered to preschool teachers near the time of the observations to obtain data on structural characteristics (child-staff ratio, adequate equipment of the preschool class, proportion of children with immigration background per preschool class, work experience of the preschool teacher, contractual weekly working hours of the preschool teacher, and job security of the preschool teacher). Table 1 provides the descriptive results and intercorrelations.
The child-staff ratio was calculated by dividing the child-group size by the number of staff (teachers and assistants) in the preschool class. Teachers' perceptions regarding adequate equipment in the preschool class (e.g. sufficient writing or drawing materials, books, or games) was measured with seven items of a scale used in the German BiKS study, and based on the Home Observation for Measurement of the Environment (Caldwell and Bradley 1984). The internal consistency of this scale (α = .73) is acceptable (Nunnally 1978). The proportion of children with an immigration background in each preschool class was calculated using the information provided by the preschool teachers. Preschool teachers also provided information regarding their work experience in ECEC settings (in years) and their contractual weekly working hours. Furthermore, teachers' perception of job security was captured with a reverse-coded item (original: 'My job security is poor') of the Effort-Reward Imbalance Questionnaire, which is rated on a 4-point scale ranging  from 1 ('strongly disagree') to 4 ('strongly agree') (Rödel et al. 2004;Siegrist et al. 2004Siegrist et al. , 2009). Since the item was reverse coded, higher scores indicate better job security.

Child characteristics
In line with previous studies Linberg and Kluczniok 2020;Smidt, Embacher, and Kluczniok 2020;Ramirez and Linberg 2022;Smidt and Embacher 2023), child characteristics were included as control variables. The gender of the children was captured through telephone interviews with the parents. Children's language skills were measured with a mean score (T-value, α = .75) of the three subtests 'understanding sentences', 'morphological rule formation' and 'phonological working memory' of the Sprachentwicklungstest für drei-bis fünfjährige Kinder (SETK 3-5; 'language development test for three to five-year-old children') (Grimm 2015).
Child personality was captured through preschool teacher ratings using the Fünf Faktoren Fragebogen für Kinder -Kurzform (FFFK-K; 'five factor questionnaire for children short version') (Asendorpf 2007). This short questionnaire was developed within the framework of the Socio-Economic Panel (SOEP) and consists of 10 bipolar items originally rated on a 11-point scale ranging from 0 to 10 (Asendorpf 2007). The 10 items were assigned to the following five factors: neuroticism (α = .59), extraversion (α = .77), intellect (α = .71), agreeableness (α = .60), and conscientiousness (α = .53). We then followed the method described by Asendorpf et al. (2001) and conducted a twostep clustering procedure to derive the following child personality types: 'resilient', 'overcontrolled' and 'undercontrolled' (see Smidt and Embacher 2023 for a detailed description). These child personality types are predictive of various social outcomes and problem behaviours in childhood (Robins et al. 1996;Asendorpf and van Aken 1999;Hart et al. 2003;van den Akker et al. 2013). The 'Big Five' patterns of the three resulting types show that 'resilients' are characterized by below-average scores on neuroticism; above-average scores on extraversion, intellect, and conscientiousness; and average scores on agreeableness. 'Overcontrollers' score below average on extraversion and above average on neuroticism, conscientiousness, agreeableness, and intellect. 'Undercontrollers' are characterized by below-average scores on intellect, conscientiousness, agreeableness, and extraversion and above-average scores on neuroticism. To evaluate the consistency of this three-cluster solution across different subsamples, we applied a double-cross validation (see Smidt and Embacher 2023 for further information). The median κ-value across 10 random splits of the original sample was above .60 (κ = .63), which is considered acceptable (Asendorpf et al. 2001).

Data analyses
To examine relations between interaction quality and structural characteristics, data were analysed using regression analyses in Stata. The inCLASS factors served as dependent variables, whereas the independent variables were structural characteristics. Child characteristics were used as control variables. We computed robust standard errors as preschool children were clustered in preschool classes (Williams 2000). To assess the practical significance, we transformed the beta coefficients into correlation coefficients (Peterson and Brown 2005;Lenhard and Lenhard 2016) with thresholds for small (r = .10), medium (r = .30), and large (r = .50) effect sizes (Cohen 1988).

Results
In this chapter, we summarize the findings on structural characteristics as predictors of teacher interactions, peer interactions, task orientation and conflict interactions. Detailed results of the regression analyses are presented in Table 2. Except for the model on teacher interactions, the regression models were significant. The amount of explained variance ranged between 10% and 22%.

Teacher interactions
The results on teacher interactions show that adequate equipment of the preschool class was positively related to teacher interactions (β = .33; transformed into correlation coefficient: r = .38) with a medium effect size. In addition, working hours per week (r = −.26) tended to be negatively associated to teacher interactions with a small effect size. Childstaff ratio, proportion of children with immigration background per preschool class, work experience, job security, children's gender, language skills and personality types were not related to teacher interactions.

Peer interactions
The findings of peer interactions show that the child-staff ratio (r = −.28), the proportion of children with an immigration background per preschool class (r = −.23), and work experience (r = −.27) were negatively related to peer interactions with small effect sizes. Moreover, working hours per week (r = −.20) tended to be negatively related to peer interactions, whereas job security (r = .19) tended to be positively related to them. Both effect sizes were small. Children's gender (r = .20) also tended to be positively associated to peer interactions. This small effect indicates that boys tended to be rated better than girls in their interactions with peers. Regarding child personality types, both 'overcontrollers' (r = −.26) and 'undercontrollers' (r = −.30) were negatively related to peer interactions. These small to medium effects indicate that 'overcontrollers' and 'undercontrollers' were rated significantly lower in their interaction quality with peers compared to 'resilients'. Adequate equipment and children's language skills were not related to peer interactions.

Task orientation
Regarding the prediction of task orientation, job security (r = .31) was positively related to task orientation with a medium effect size. Furthermore, children's language skills (r = .30) were positively associated to task orientation with a medium effect size. Child-staff ratio, proportion of children with immigration background per preschool class, adequate equipment, work experience, working hours per week, children's gender and personality types were not related to task orientation.

Conflict interactions
The results of conflict interactions show that child-staff ratio (r = .29) was positively related to conflict interactions, whereas work experience (r = −.28) was negatively related to conflict interactions. These small effects indicate that a poorer child-staff ratio (i.e. more children to care for) and less work experience of the preschool teacher were related to more conflict interactions of the children. Language skills of the children (r = −.20) tended to be negatively related to conflict interactions with a small effect size. Furthermore, 'overcontrollers' (r = −.22) were negatively associated to conflict interactions. This small effect indicates that 'overcontrollers' show fewer conflict interactions than 'resilients'. Proportion of children with immigration background per preschool class, adequate equipment, working hours per week, job security, children's gender and the personality type 'undercontrollers' were not related to conflict interactions.

Discussion
The aim of the present study was to examine relations between structural characteristics and interaction quality in Austrian preschools. After controlling for child variables, there were positive relations between adequate equipment and teacher interactions, as well as negative relations between child-staff ratio, proportion of children with an immigration background per preschool class, preschool teachers' work experience, and peer interactions. Preschool teachers' job security was positively related to task orientation. A poorer childstaff ratio led to more conflict interactions, whereas greater work experience of the preschool teachers led to fewer conflict interactions. Effect sizes were small to medium.

Structural characteristics as predictors of interaction quality
Teacher interactions Adequate equipment of the preschool class was positively related to the quality of children's interactions with preschool teachers, and the effect size was medium. Similar results have been found in a German study, where interaction quality was captured with a focus on preschool teachers (Smidt 2012). The finding might indicate that preschool children use equipment of the preschool class to initiate and sustain interactions with the preschool teacher. However, the role of the preschool teacher remains unclear because the inCLASS does not directly capture the activities of the preschool teachers. Recent analyses on relations between activity settings and interaction quality were conducted with the sample that was used in the present study revealed that preschool teachers' scaffolding behaviour positively predicted teacher interactions . However, it must also be noted that preschool teachers' support is less likely to occur through teacher-planned activities, and free choice is the dominant activity setting as previous findings for Austria/Tyrol suggest . Similar results have also been found in some older studies conducted in Germany (Tietze et al. 1998;Smidt 2012), and it has been discussed that preschool teachers may well provide adequate equipment but tend to be less actively involved in educational activities (Bredekamp and Rosegrant 1992;Smidt 2012).
The number of preschool teachers working hours per week negatively predicts the quality of children's interactions with preschool teachers, but the association was only significant by trend, and the effect size was small. This finding might be explained by more working hours corresponding with increased work-related stress to some extent, such as higher levels of depersonalization (Trauernicht, Besser, and Anders 2022). Work-related stress might lead to lower quality of interactions (Zinsser et al. 2013;Penttinen et al. 2020; see also Embacher 2021).

Peer interactions
With the exception of adequate equipment, the relations between structural characteristics and peer interactions are significant or significant by trend with small effect sizes. More children per staff corresponded with lower quality of peer interactions. Conceivable reasons for this are the following: (1) A poorer child-staff ratio might relate to confined space and limited resources (e.g. play materials) for each child, which may lead to lower interaction quality with peers. (2) A poorer child-staff ratio probably relates to a higher noise level and increased discomfort, which may impede children's interactions with peers to some extent. (3) A poorer child-staff ratio may result in preschool staff having less time to enhance and support peer interactions. In particular, preschool teachers' scaffolding behaviour to support peer interactions has been mentioned as crucial (Acar, Hong, and Wu 2017).
A higher proportion of children with an immigration background per preschool class corresponded with a decrease of the quality of peer interaction. This is in line with other studies, which found that lower proportions of children with an immigration background in preschool classes corresponded with better interaction quality (Kuger and Kluczniok 2008; see Kluczniok and Roßbach 2014 for an overview). When the focus is on the role of the preschool staff, one explanation is that a high proportion of children with an immigration background attracts preschool teacher's attention to the support of children's language competencies rather than other domains (Kuger and Kluczniok 2008). In the present study, the focus is on the children, indicating that a high proportion of children with an immigration background may lead to increased language barriers and communication problems, which may inhibit peer interactions.
Working experience was related to lower quality of children's interactions with peers. Reasons for this result may be multi-fold. It could be that preschool teachers with less working experience whose training was completed more recently are more familiar with current professional contents and strategies to support peer interactions. Alternatively, increased work experience may be accompanied by decreased work motivation, which leads to less involvement in promoting peer interactions (see Tietze et al. 1998).
More working hours per week were negatively related with peer interactions, whereas greater job security tended to be positively related to peer interactions. In both cases, work-related stress, which relates to lower quality of interaction (Zinsser et al. 2013;Penttinen et al. 2020), may play an important role. Better working conditions (e.g. job security) may relate to a decrease of work-related stress (Viernickel, Voss, and Mauz 2017), whereas a higher number of working hours per week and dissatisfaction with the working hours are associated with increased work-related stress (Trauernicht, Besser, and Anders 2022).

Task orientation
Of the structural characteristics, only job security of the preschool teacher predicts preschool children's task orientation. Higher job security probably serves as a resource for lower work-related stress and higher working motivation (Bakker and Demerouti 2007), which may lead to higher engagement in preparing preschool classes adequately to foster children's task engagement (see Penttinen et al. 2020 for similar findings). However, previous findings conducted with the sample  indicate that preschool teachers obviously tend to not actively support children' s task orientation by using social-constructivist approaches such as scaffolding (Winsler and Carlton 2003;Bodrova and Leong 2012). The other structural characteristics do not play a prominent role in the current study. This is also true for adequate equipment, where significant relations with task orientation have been found in previous analyses by using the same sample . However, this relation disappears when further structural characteristics are used as predictors.

Conflict interactions
Only a few structural variables revealed a significant influence, and the effect sizes were small. A poorer child-staff ratio was related to more conflict interactions. This finding seems plausible since a poorer child-staff ratio probably corresponds with limited space and resources for each child, as well as increased noise and discomfort, which may lead to increased conflict interactions with peers, preschool teachers, and reduced behaviour control. However, it should be noted that there were relatively few observations of conflictual interactions, which is in line with previous studies (von Suchodoletz, Gunzenhauser, and Larsen 2015;Slot and Bleses 2018). The low level of conflict interactions and the lack of variance (including distributional problems such as high skewness) have also been critically discussed (von Suchodoletz, Gunzenhauser, and Larsen 2015; Slot and Bleses 2018;Smidt and Embacher 2021), and it has been suggested that conflictual items of the inCLASS be revised (Smidt and Embacher 2021).
Distributional problems may also impede the interpretation of the finding that greater work experience of preschool teachers led to less conflict interactions. Previous research on work experience has been inconsistent (Smidt 2012;Slot 2018). In the current study, preschool teachers with more work experience might have developed habits and routines to deal with conflicts in the daily preschool life in such a way to ensure some calm and discipline more successfully. This is, however, a vague assumption, which should be investigated in future studies.

Control variables
We used children's gender, language competencies, and personality as control variables, and there were some small to medium effects on peer interactions, task orientation, and conflict interactions. One finding is the predictive role of children's personality in terms of the personality types, particularly for peer interactions, as well as conflict interactions to a lesser extent. This pattern of results has been intensively discussed in a recent study conducted with the same sample (Smidt and Embacher 2023). The present study can emphasize the robustness of these findings.
Another finding is related to the predictive importance of children's language competencies for their task orientation and, to a lesser extent, conflict interactions. This pattern of results has been detected and discussed in recent analyses with the same sample (Smidt, Embacher, and Kluczniok 2020;Embacher 2020, 2023). The present analyses underscore the robustness of the findings on the predictivity of children's language skills.
Finally, boys revealed a slightly higher quality of peer interactions than girls. This finding is only significant by trend but persists after accounting for structural variables. Thus, some robustness of this slight difference can be assumed. However, in sum, previous research implies that children's gender does not seem to play a central role in the prediction of interaction quality in preschool (Downer et al. 2010;Vitiello et al. 2012;Slot and Bleses 2018;Smidt, Embacher, and Kluczniok 2020).

Study limitations and implications for educational policy and practice
This study has some limitations. First, it must be considered that this study was conducted in the Austrian federal state of Tyrol. Because important elements of the preschool system are legislated at the federal state level (Smidt 2018), transfer to the other Austrian federal states is limited to some degree. Second, it cannot be ruled out that a deviating pattern of findings may emerge in the second and third years of preschool. Third, interaction quality captured with the inCLASS puts the focus on child-initiated interactions at the individual child level and not on interactions initiated by preschool teachers on preschool class level, as with some other instruments (e.g. CLASS). It is possible that different patterns of findings would emerge if interaction quality had been measured with other instruments. Fourth, the factor conflict interactions indicates distributional problems, which may influence the findings (Smidt and Embacher 2021). Fifth, it was not possible to consider the level of preschool teachers' training (e.g. non-academic, bachelor's degree) as a predictor because preschool teachers in Austria usually receive non-academic training at the BAfEPs (Eichen and Krenn-Wache 2020).
With regard to implications the present findings could raise further awareness at the level of educational policy (at the national, federal state, and municipal level) about the necessity to ensure good framework conditions related to factors that have been shown to be important in the present study (e.g. adequate child-staff ratios, well-equipped preschool classes, and preschool teachers' job security). Some structural characteristics of the preschool classes can be regulated by educational policy and law (Kluczniok and Roßbach 2014), and with regard to Austria, such regulations could be implemented nationwide with consistent standards.
The present study provides first empirical indications for the importance of structural variables for interaction quality in the Austrian (Tyrolian) preschool context. That is, by revealing small to medium effect sizes, child-staff ratio, proportion of children with immigration in the preschool class, adequate equipment, preschool teachers work experience, and job security play a predictive role, depending on the domain of interaction quality. Since the inCLASS focuses on the individual child level, other measures of quality such as CLASS and ECERS-R (which focus on the level of the preschool class) should be included to gain a more comprehensive picture about predictors of interaction quality in Austrian preschools (see Smidt 2012 for a discussion on educational quality in Germany). This is particularly true because until recently there were very few (older) study results on interaction quality in Austrian preschools (Smidt 2018). Furthermore, it seems important to consider the level of education of preschool teachers. In the medium term, this could be possible since some academic courses in early childhood education (advanced training) have been implemented recently (Hartel et al. 2019;Holzinger and Reicher-Pirchegger 2020).
Despite the importance of structural characteristics, it should not be overlooked that child characteristics such as personality types and languages skills are important predictors of interaction quality as well. It is important to note that the findings on child characteristics persist even after accounting for structural variables and can be considered relatively robust (see Smidt, Embacher, and Kluczniok 2020;Smidt and Embacher 2023 for earlier analyses). In sum considering structural aspects (e.g. adequate child-staff ratio, rich and good material resources in preschool classes, job security of the preschool teachers) and awareness of influencing child variables (e.g. focusing on children identified as 'undercontrollers', focusing on children with poor language skills) highlight important ways to improve children's interaction quality in preschool.

Disclosure statement
No potential conflict of interest was reported by the author(s).