Teachers’ instructional practices during pupils’ individual seatwork in Norwegian Language arts

ABSTRACT This study examines teachers’ instructional practices during seatwork in primary Norwegian Language arts classrooms. Although seatwork is a common activity in classrooms, our knowledge about what teachers actually do during this activity is limited. In this qualitative study, six teachers in third grade primary school were videotaped while pupils worked individually. The study gives examples from writing instruction. An analytical framework was developed based on existing research and used for coding of the video data. The analyses reveal that the participating teachers utilised their time during seatwork very differently. Two of the six teachers spent a majority of the time on instructional support (for instance challenging pupils’ thinking), while the other four spent the majority of the time on organizational support (for instance practical help) or monitoring (for instance listening to pupil read). Emotional support was almost non-existent in these classrooms. Teachers’ use of challenging versus telling as a way of supporting pupils is a key difference among the observed teachers.


Introduction
Seatwork, defined as pupils working alone on an assigned task (Stright & Supplee, 2002), is a common activity in classrooms around the world (Alexander, 2001;O'Keefe, Xu, & Clarke, 2006). Research from Nordic classrooms shows, for example, that the amount of individual seatwork is increasing (Carlgren, Klette, Myrdal, Schnack, & Simola, 2006;Klette et al., 2008;Lindblad & Sahlström, 1999). For the teacher, seatwork often means walking between the desks, helping and supporting the pupils in their individual work. Alexander (2001) uses the term monitoring to describe teachers' activities during seatwork, covering both teachers' surveillance and teachers' help. While monitoring and teacher support during seatwork is a frequent activity in primary classrooms (Alexander, 2001;Klette, 2003), it is also often described as 'the invisible teaching method' (Alexander, 2001, p. 407), and there is a lack of knowledge about teachers' instructional practices during seatwork (Alexander, 2001;Haug, 2006;Klette, 2003;Sahlström, 1999;Tanner, 2014). Instructional practices (Grossman and McDonald (2008) Pressley, Rankin, and Youkoi (1996)) here refers to the different approaches teachers use in their conversations with individual pupils. The term covers academic, social and practical help given by teachers to individual pupils during seatwork. Scholars seem to use different terms when trying to explain teachers' instructional activities during seatwork, for instance individualised teaching (Bergem, 2015), monitoring (Alexander (2001), Kikan-Shido ("walking between the desks") (O'Keefe et al., 2006) or desk interaction (Tanner, 2014). While the term monitoring has certain connotations of control, the term desk interaction may be understood as purely descriptive. None of these, however, covers the breadth of teachers' activities during individual seatwork; in the following, we use the term instructional practices when trying to untangle teachers' activities and approaches during seatwork.
In the article that follows we explore teachers' instructional practices during seatwork in Norwegian Language arts lessons at the primary school level. The purpose of the article is to increase the depth of understanding about teachers' instructional practices during this type of teaching. Six teachers from third grade (8-to 9-year-old pupils) in primary classrooms participated in our study. These teachers were videotaped while pupils were doing seatwork over the course of one week; all the analysed examples are from periods of writing instruction.
Research reveals that teachers' activities during pupils' seatwork have different functions. O'Keefe, Xu and Clarke (2006) distinguish between guidance of pupils, organizational work, social work and monitoring as four different functions based on analyses of between-desk instruction (Kikan-Shido). Alexander (2001) uses the term monitoring to refer to a low level activity during which the teacher keeps an eye on the students and observes their working process without talking to them (Alexander, 2001). Pianta and Hamre (2009) and Klette (2013) distinguish between instructional support, organizational support and emotional support as three key elements in teachers' instructional practices. Although they originate from studies of whole class instruction, these elements are recognizable as teachers' instructional practices during seatwork and are thus relevant for the analyses in this article. Grossman and McDonald (2008), Klette (2013), and Schoenfeld (2014) argue that organizational and emotional support, such as classroom management and classroom climate (Pianta & Hamre, 2009), contribute favourably to pupils' learning. They all claim, however, that systematic research on teachers' instructional support, such as teachers' use of modelling and telling versus challenging, is still in its initial phase, and further research is needed. Teachers' use of instructional, organizational and emotional support, including monitoring practices, are key questions investigated in the present study, and we are especially interested in teachers' use of instructional support in these primary classrooms. Instructional support here refers to instructional approaches the teachers use to help their pupils while walking between desks, such as giving explanations (telling), confirming that the student is on the right track or modelling possible ways of solving the task at hand. Organizational support covers task and activity management and practical help, while emotional support refers to social, non-academic talk. The following research questions will be explored: (1) What characterises each teacher's instructional practices during pupils' seatwork in language arts lessons? (2) Which approaches do teachers make use of in their instructional support?
We developed a conceptual framework that enabled us to analyse different aspects of teachers' instructional practices during individual seatwork. Below, we summarise relevant theoretical perspectives and key concepts contributing to the framework. We then present data sources and methodological design, followed by key findings and results.

Theoretical framing
As a theoretical framing of this study, we see teachers' instructional practices as part of their scaffolding repertoire. Scaffolding as an approach to teaching was introduced by Wood, Bruner andRoss in the late 1970s (1976). They defined scaffolding as support 'that enables a child or novice to solve a problem, carry out a task or achieve a goal which would be beyond his unassisted efforts' (Wood et al., 1976, p. 90). The term scaffolding is closely linked to sociocultural perspectives on teaching and learning, and the teacher must adapt their scaffolding to the individual pupil (Myhill & Warren, 2005). Scaffolding is also temporary (Mercer & Littleton, 2007). The Vygotsky tradition in sociocultural theory includes two processes which Vygotsky calls intrapsychological and interpsychological (Vygotsky, 1978, p. 57). This refers to the notion that a child's cognitive development happens both in and through social interaction (Rogoff, 1990), as, for example, through conversations between the pupil and the teacher during seatwork in classrooms. A key challenge of such interaction is how involved or how distant the teacher should be in these conversations and the degree of direct and explicit instruction versus the degree of facilitation.
Scaffolding research includes research on pupils' personal development, for example, linked to emotions (Meyer & Turner, 2007) or motivation (Guthrie, Wigfield, & Perencevich, 2004), as well as pupils' academic development. Concerning the latteracademic development-researchers distinguish between different means of scaffolding (Van de Pol, Volman, & Beishuizen, 2010). Scaffolding means include instructional practices based on constructivist approaches, such as challenging the pupil through hints and questioning. It also includes cognitive approaches, such as giving the pupil the answer through explanations and direct instruction. The teacher may also use modelling or feedback as a means of scaffolding (Van de Pol et al., 2010). In the next sections, we will expand on these scaffolding means by first synthesising research on teachers' instructional practices and then considering more specifically research on teachers' feedback practices in primary school. This review also serves as a grounding for our analytical framework; thus, key concepts are marked in italics throughout the review.

Teachers' instructional practices during seatwork
When discussing instructional practices in language arts instruction, Taylor and colleagues (2002) make a distinction between telling and different ways of challenging pupils (see also Grossman et al., 2010). In this research telling means that the teacher gives the answer to the pupil, without apparently requiring the pupil's understanding. Challenging is used as an opposite to telling, meaning that the teacher motivates the pupil to think for himself, for instance through asking questions to the pupil. According to the present study, we stress that the concept is used to describe the teachers' instructional practices, as a way of teaching. Whether the pupil really is challenged cognitively or not is not investigated in this project. Terms with similar meanings, also from language arts instruction, are 'telling' versus 'coaching' (Taylor et al., 2002), and 'direct instruction' versus 'constructivism' (Pressley et al., 2003). Studies of expert language arts teachers argue that accomplished teachers prefer coaching over telling in their literacy instruction (Allington & Johnston, 2002;Taylor, Pearson, Peterson, & Rodriguez, 2005). Taylor et al. (2005) find that results in both reading and writing achievement are positively related to coaching as an instructional practice, and they see challenging pupils' thinking as decisive for pupils' cognitive engagement in language arts. Using story writing as a reference point, Dunn and Finley (2010) underscore teachers' encouragement of pupils' reflection before and during the writing process. Building on this research, telling and challenging constitute key categories in the analytical framework in the present study. Challenging as an instructional practice includes teachers' use of questioning, (Alexander, 2001;Cazden, 1983;Myhill, 2006), which is frequently used during writing instruction (Davidson, 2007;Dunn & Finley, 2010). Recent research into questioning as an instructional activity has focused in particular on the function of teachers' questions rather than the form (e.g. open versus closed questions). Based on literacy and numeracy teaching in primary classrooms, Myhill (2006, pp. 26, 27), for example, argues that different teacher questions have different functions, such as classroom management, checking for prior knowledge, developing vocabulary, and developing reflection. Teachers' questioning as an instructional practice may thus be linked to all three support elements, instructional, organizational and emotional. However, in the present study, the most relevant aspect is teachers' use of questions as a part of instructional support and challenging.
Teachers' instructional activities during seatwork discussed so far are based on teachers' oral feedback and interactions. Research also underscores that adults are decisive role models in children's literacy development (Cazden, 1983). Modelling is claimed to be an effective tool in language arts instruction (Grossman et al., 2010;Schunk & Zimmerman, 2006), for both reading (Fisher, Frey, & Lapp, 2008;Methe & Hintze, 2003) and writing instruction (Martin, Christie, & Rothery, 1987). To be a good writer, pupils must see the teacher write, argue Gerde, Bingham and Wasik (2012). Modelling is thus a vital category when analysing teachers' instructional practices during seatwork. Thus far, telling, challenging, and modelling are identified as key concepts in teachers' instructional practices during seatwork within the subarea of instructional support. We now turn to research on teachers' feedback practices.

Teachers' feedback practices in primary school
As already indicated, oral feedback from teacher to pupil forms a central part of teachers' instructional activities during seatwork. Feedback research has received significant attention in recent years; Hattie (2008) ranks feedback in his top ten list of elements influencing pupils' achievement. However, the most influential and cited reviews of feedback research do not distinguish between oral and written feedback (Black & Wiliam, 1998;Hattie & Timperley, 2007) or between individual and collective feedback (Kluger & DeNisi, 1996;Shute, 2008). Neither do any of these reviews include pupils' age or specify the subject area at hand when discussing feedback. Using primary classrooms as their reference point, Tunstall and Gipps (1996) distinguish between feedback related to socialization and feedback related to academic assessment. They further separate evaluative feedback (e.g. rewarding and punishing) and descriptive feedback (e.g. specifying attainment or improvement). In the case of writing instruction, most feedback research has focused on teachers' written feedback on pupils' written texts. There may, however, be differences between teachers' carefully prepared written feedback and the oral, 'on the spot' feedback taking place in classrooms. Research on oral 'writing conferences' (Harris, 1986) at the primary level emphasises, for example, the complexity and dynamic character of such dialogues about writing. Glasswell and Parr (2009) identify three crucial factors necessary to exploit what they describe as 'teachable moments': the teacher must know the pupil's writing level and how the pupil can improve the writing. In addition she should be able to create meaningful classroom talk that could help the pupil in his writing. Hawe, Dixon, and Watson (2008) find that teachers struggle with providing feedback that addresses the substantive and deep features of writing. They conclude that oral feedback makes significant demands on teachers' competence in language arts education. Hattie and Timperley (2007) and Shute (2008) underscore that high-quality feedback should be linked to clear learning goals and should be provided on one of four levels: task level, process level, self-regulation level or self-level (Hattie & Timperley, 2007). Feedback on the self-level is not effective for learning, Hattie and Timperley argue, as non-targeted praise linked to the person does not lead to learning (Kluger & DeNisi, 1996). To lead to learning, praise must be specific, confirming what is good and why (Hattie & Timperley, 2007). For us, this finding is noteworthy as non-specific praise is rather common in primary classrooms (Burnett & Mandel, 2010;Haug, 2006) internationally as well as in Norway. Pianta and Hamre (2009), on the other hand, see positive climate, which includes such non-specific praise, as a key element in their category emotional support, while they define teachers' feedback practices as a part of teachers' instructional support. Voerman, Korthagen, Meijer, and Simons (2014) claim that all kinds of praise may create positive emotions and may thus, indirectly, positively influence learning. In this study, specific praise is defined as a part of instructional support and linked to the subcategory confirming.
To summarise, giving learning goals and confirming are key concepts in teachers' feedback practices. Together with telling, challenging and modelling, these are key concepts for our analysis of teachers' instructional practices during seatwork. These five dimensions constitute what we call teachers' instructional support in this study. Scaffolding as a theoretical framework involves a broad view of teaching. Thus, in addition to instructional support, we also include emotional support, organizational support (Pianta & Hamre, 2009) and teachers' monitoring (Alexander, 2001) in our analyses as important dimensions for pupils' learning; these are discussed further in the next section. Combined, we see these dimensions as a part of teachers' scaffolding repertoire during seatwork.

Methods and Design
The data for the present study are based on video observations of six primary school teachers' third-grade classrooms. Each of the teachers was observed for one week. The design of the study was inspired by video research projects such as LPS (Clarke, 2006) and the PISA + study (Klette, Bergem, & Roe, 2015;Klette et al., 2008).

Sample
Our sample consists of six teachers, who were chosen to provide as much variation as possible in terms of professional background and experience. We asked school authorities in a medium-sized Norwegian city for permission to observe six third-grade teachers. The authorities asked the principals, who in turn asked their teachers. The six teachers represent three different schools, with three teachers from school A, two teachers from school B and one teacher from school C. Five of the teachers were female and one was male. Teachers varied in age from 37 (Ellen) to 55 (Maria). All six teachers were experienced, with between 12 and 31 years of teaching experience. They varied, however, in professional background. Four of the teachers had general teacher education, while two of them (Peter and Julie) were educated pre-school teachers. All of the teachers had attended different professional development courses, and Anne and Ada had specific professional qualifications targeted to Norwegian Language arts primary education. Table 1 describes the teachers, schools and classrooms involved in this study, including the school and class size.

Data collection
To obtain data from relevant lessons that could shed light on teachers' instructional practices during seatwork, the teachers selected lessons that included planned individual seatwork. At the lower primary levels, seatwork is often only a small part of a  lesson. Thus, the entire lesson was videotaped to capture all of the teachers' class instruction, including instruction prior to seatwork. The whole lesson was seen as relevant for understanding the teacher's actions during the seatwork. The specific period devoted to individual seatwork during these lessons was then used for systematic coding and analyses (see next subsection). A two-camera approach was used. The cameras covered the whole class and the teacher walking between the desks. The teacher was wearing a lapel microphone that captured the words of both the teacher and the pupil in focus. The total videotaped data material consists of 32 hours and 10 minutes of Norwegian Language arts instruction. The seatwork portion used for systematic analyses includes 11 hours and 47 minutes of videotaped data material. The time spent on seatwork varied from 1 hour and 31 minutes (Anne) to 2 hours and 42 minutes (Ada) over the week that each teacher was filmed.

Analysis
As described in the theory section, we created a framework building on existing research for the analysis of the videotaped data, comprising concepts relevant for analysing teachers' instructional practices during seatwork. The framework covers the main categories of Instructional support, Emotional support, Organizational support, Monitoring and the category No direct interaction with pupils. For organizational support, we distinguish between practical help, correction/redirection, and task and activity management. Emotional support refers to social, non-academic talk, for instance, about spare-time activities. Organizational support includes practical help, task and activity management and teachers' correction or redirecting of students' work (getting them back on track). Because we were interested in covering the whole range of teaching activities and approaches during individual seatwork, we also included the category monitoring (e.g. overlooking working progress and listening to pupils read). The category called no direct interaction with pupils covers the teacher talking to another adult, sitting at the desk, or clearing/organizing the classroom. These five main categories, divided into 15 subcategories, serve as an analytical framework for classifying teachers' instructional practices during seatwork linked to instruction in Norwegian Language arts classrooms. Table 3 summarises the key dimensions of the categories, including their subcategories (for a more detailed description of the codes, see Appendix 1).
The selected material was coded by using the software Videograph (http://www. dervideograph.de/enhtmStart.html) and the conceptual framework described above. The five main categories and 15 subcategories were designed to be mutually exclusive. Consequently, dilemmas occasionally arose in assigning practices to subcategories, for instance, when the teacher was providing instructional support and practical help at the same time. For example, the teacher might sharpen pencils while talking to a pupil about an academic task. In such situations, the coding was assigned based on what the teacher said rather than on what the teacher did. In the coding process, we used sequence coding, meaning that every second of seatwork is coded (as opposed to interval coding, where, for instance, every fifth second is coded). Once the material was coded, we calculated the time spent on the different subcategories within the five main categories. These results, and figures presented in the following section, create the basis for the analyses and discussion in this article. A research assistant coded 10% of the material. A software program compared this coding with the first author's coding, second per second. The results of this reliability testing showed 79% compliance for the main categories and 72% for the subcategories. In addition to the aspects covered by the categories in the framework, we also tried to describe the overall social climate and relations between pupils and teachers in the observed classrooms.
There are some limitations of this study. The sample is small, and the observations are limited to third-grade classrooms. We observed classroom practices for only one week in each classroom, and the type of activities scheduled during this week might influence the teacher's instructional practices during seatwork. Lastly, we want to make clear that we cannot generalise these results to the instruction practices of all teachers on the basis of such a selective sample. As such, this study serves as an attempt to understand possible differences in teachers' instructional practices during seatwork in primary classrooms.

Results
This section presents the results of coding the material. We begin by capturing teachers' main activities during seatwork and then move to teachers' approaches during instructional support specifically. Finally, we present two teachers' (Anne's and Ellen's) use of challenging as instructional support during seatwork.

Teachers' main activities during seatwork in Norwegian Language arts lessons
As described above, our analytical framework distinguishes between five main categories for analysing teachers' instructional practices during seatwork: Instructional support, Emotional support, Organizational support, Monitoring and No direct interaction with pupils. To illustrate similarities and differences between the teachers, we show the results as pie charts. Figure 1 shows large variations between the teachers in the amount of time spent on the different categories: The time each teacher spent on the different activities varied widely. During the seatwork, the teachers spent between 14% and 52% of the total time on instructional support. Ellen spent the most time, more than half, on instructional support, while Ada spent the least time on instructional support. The amount of time spent on organizational support also varied significantly, from 12% to 58%. Ada and Peter spent more than half of their time on such activities, while Julie spent 12% of her time on organizational support. In addition, the time spent on monitoring varied significantly (from 4% to 50%). Julie spent half of the time on monitoring, specifically coded as listening to pupils read.
Different contextual factors may explain some of these varying results. A closer look at organizational support indicates that it specifically covaries with classroom management and type of activity and task. If the teacher provided unclear instructions, subsequent support included a large amount of help coded as task and activity management. The situation is often, however, more complex, as in Ellen's case: her classroom management was usually very clear. Most of Ellen's organizational support, however, occurred when she taught a turbulent class, of which she was not the primary class teacher. She therefore provided a great deal of support coded as task and activity management during these lessons. In addition, practical activities often required organizational support (subcategory practical help). Julie was the teacher with the fewest practical activities in her lessons and thus provided primarily teacher-centred instruction. She also spent the least time on organizational support.
The degree of monitoring reflects how much time the teachers spent on overseeing pupils' work, including listening to pupils read. At Julie's school (School C), for example, the school leadership had decided that teachers should check the pupils' reading homework daily. This may have influenced the amount of monitoring in Julie's classroom. The emotional support category, in comparison, was rather homogenous: all of the teachers spent only small amounts of time on emotional support (0%-4%). The classroom atmosphere, however, generally appeared good, and teachers treated their young pupils with friendliness and respect. Finally, there were differences in how much time teachers spent in direct interaction with pupils: time coded as no direct interaction with pupils varied from 2% to 23%. Two of the teachers, Ada and Julie, spent more time on no direct interaction with pupils than on instructional support. In contrast, Peter and Ellen spent almost all their time on interaction with pupils. It is worth noting that Peter and Ellen were the only adults in their classroom during the week. No other teachers or assistants were present.
In summary, except for emotional support, we may so far conclude that there are no distinct patterns in the type of support given by teachers during seatwork. To scrutinise these results further, we analysed in more detail the category instructional support in the cases of Anne and Ellen.

Investigating instructional support in depth: The cases of Anne and Ellen
As indicated previously, there was enormous variation in time spent on instructional support in the observed classrooms. Within instructional support, we distinguish between telling/giving answer, challenging, giving learning goals, modelling and confirming as five subcategories. We selected two of the teachers, Anne and Ellen, for more indepth analyses. These two were chosen because they are the only teachers who worked with story writing during the observed week, a topic that often demands significant help from the teacher (Davidson, 2007;Dunn & Finley, 2010). These two teachers also have the most similar profiles with regard to the main categories of the study, including time spent on instructional support, making a comparison possible. Both Anne and Ellen primarily engaged in instructional support (45% and 52%). Figure 2 below illustrates the subcategory profiles of instructional support for Anne and Ellen (for corresponding profiles for all six teachers, see Appendix 2).
We see from Figure 2 that even though Anne and Ellen spent almost the same amount of time on instructional support, their specific instructional support approaches differed substantially. Both Anne and Ellen used confirming, but to different extents (Anne, 11%; Ellen, 4%). Anne praised the pupils for specific elements they had

Anne's instructional support Ellen's instructional support
Telling/giving answer  mastered, for instance, regarding orthography and punctuation. The use of non-specific praise was more frequent than specific praise for both teachers. For instance, when a pupil had an idea regarding what to write about, Ellen answered: "Yes! Write that. Yes! Excellent!" Modelling was a part of Ellen's instructional support but not of Anne's. For instance, Ellen demonstrated the meaning of specific words: when a pupil asked what 'toddle' meant, she showed how a toddler moves. Ellen also used learning goals as a part of her instructional support during the observed week; she was the only one out of the six teachers who did this during the observations. Anne's instructional support did not include modelling or giving learning goals but were spread more evenly between telling/ giving answer, challenging and confirming. Ellen showed a broad repertoire of instructional practices, dominated by challenging.
The main difference between the two teachers lay in the ratio between telling/giving answer and challenging as instructional approaches. During her instructional support, Ellen spent 72% of the time on challenging the pupil's thinking, while Anne used challenging 45% of the time. Anne used telling/giving answer 43% of the time compared to 21% of the time for Ellen. Ellen used challenging as her main approach, while Anne used a combination of telling/giving answer and challenging.
So far, we have shown that the teachers' instructional activities during seatwork varied significantly as a whole as well as by subcategory within instructional support. In the section that follows, we will look more closely at the subcategory challenging and how it is used by Anne and Ellen as a part of instructional support. Challenging is chosen because teachers seem to use this approach often in writing instruction (Davidson, 2007), with good results (Taylor et al., 2005).

Anne's and Ellen's use of challenging in pupils' story writing
To illustrate how challenging may appear in writing instruction, we show two excerpts of Anne's and Ellen's support during pupils' story writing. The two situations were chosen because they have many common features, and one can therefore assume that the instructional practices were fairly similar. Both situations are taken from lessons in which the pupils were writing stories and were at the beginning of their work. Both pupils struggled with getting started on their writing.
The pupil in the first excerpt (excerpt 1) was instructed to write a story based on a picture selected from a box of illustrations and photos. The pupil in excerpt 2 had a model text with the title 'When the dog got its cold nose'. She was asked to write her own equivalent story about why the giraffe has such a long neck.

Excerpt 1: Ellen's way of challenging
The pupil had raised her hand, and the teacher came to her desk.
(1) T: What do you see in the picture?
(2) P: [Whispering] A girl. (6) T: Scissors. Mm. What do you think the girl is going to do with the scissors? (7) P: Cut out a sun. (8) T: Cut out a sun, maybe. Mm. Why? (9) T: What do you think she is going to use the sun for? (10) P: Glue on a piece of paper.
(11) T: Maybe she's going to glue it on a piece of paper. Where is she, then? (12) P: At school. [The sequence lasts 50 seconds.] The teacher left. The pupil immediately took her pencil case from her school bag and took out a pencil and rubber. Two minutes and 10 seconds later, she started writing.

Excerpt 2: Anne's way of challenging
The pupil had raised her hand, and the teacher came to her desk.
(1) T: How did it happen that the giraffe got long neck? [Pause] (2) T: No . . . What may have happened? Something happened. [The situation lasts 1 minute and 30 seconds.] The teacher left. The pupil sat still for five seconds. Then she started writing. Both teachers used a friendly voice when talking to the pupils. Ellen repeated what the pupil said. Anne squatted down physically to the pupil's level and smiled when she talked. Excerpt 2 is, however, a pure teacher monologue. The pupil looked at her text and did not have eye contact with the teacher at all. Excerpt 1 is more of a dialogue, but the pupil was whispering and the teacher took the lead during the conversation. Both of the teachers used questions (Ellen, n = 7; Anne, n = 7) to challenge the pupil. The use of open questions dominates. Open questions are questions with indeterminate answers (Nystrand, Gamoran, Kachur, & Prendergast, 1997), and the use of this kind of questions may be due to the type of task: there are no definitive answers in these writing tasks. When we look more closely at the use of questions, we see that Anne asked seven open questions in a row. She seemed to appeal to the pupil's imagination, thus leaving it to her to develop the study further. Ellen first asked three closed questions based on the pupil's picture. Then she continued with four open questions. Ellen also developed questions that went beyond the picture, appealing to the pupil's imagination. From these excerpts, we can say that both teachers provided support linked to the pupils' imagination in order to nurture their writing, and both conversations created action by the pupils.

Discussion
Our analyses of the observed teachers showed vast differences among the teachers in terms of their instructional practices during seatwork; interestingly, there are few common characteristics across the teachers. In answer to our first research question ("What characterises the teachers instructional practices during pupils' seatwork in language arts classrooms?"), Peter, Maria and Ada spent their time mainly on organizational support, Anne and Ellen on instructional support and Julie on monitoring. However, taking the results together, organizational support is the most frequently occurring category (34% of the coded material), and emotional support is the least frequently occurring category (2%). As our second research question focussed specifically on the teachers' use of instructional support ("Which approaches do teachers make use of in their instructional support?"), Anne and Ellen were chosen as illustrative cases to examine how instructional support might vary across different classrooms. The analyses show that even teachers with similar teaching practice profiles may differ in how they provide instructional support during seatwork. Both Anne and Ellen used telling/giving answer and challenging in particular, but to different degrees. From the perspective of scaffolding, the teachers in this study use different scaffolding means (Van de Pol et al., 2010), including both constructive and cognitive approaches. The function of instructional support during seatwork indicates that the scaffolding is temporary (Mercer & Littleton, 2007): the teacher provides as much help as she thinks is necessary for the individual pupil, and then she moves on to the next pupil. In the discussion that follows, we elaborate on patterns and possible variations across the teachers with regard to the five main categories of teaching practices.

Instructional support
Our data show how teachers use challenging during writing instruction to help pupils start writing; researchers agree that challenging is a core element of fostering pupils' learning through instructional support. There may, however, be both academic and pragmatic reasons for usingtelling/giving answer as well. As an example, we may imagine a pupil who asks the teacher how to correctly write a word. We argue that it may sometimes be more effective to tell the pupil the answer so he can continue writing rather than to always ask him to look up the word in a dictionary. This is in accordance with Pressley and colleagues (2003), who argue that strong evidence exists for balancing direct instruction (telling/giving answer) and constructivism (challenging) in literacy instruction. With regard to scaffolding theory, the most important aspect may not be how to scaffold but rather that the support works as a scaffold-that can later be removed.
The teachers' lack of instructional support that was directly linked to learning goals is noteworthy, as this kind of feedback is highlighted as a factor contributing to success in feedback research (e.g. Hattie & Timperley, 2007). The lack of feedback linked to learning goals in literacy instruction supports observations by Hawe et al. (2008), who recommend professional training in this area. This also illustrates that some recommended feedback strategies may be challenging for teachers to carry out in a complex classroom setting (Svanes & Skagen, 2016).

Organizational support
The vast difference in how much time the teachers spent on organizational support (varying from 12% to 58%) is worth discussing. We might argue that the amount of organizational support should be minimal, as a large proportion of organizational support may originate from lack of clarity in classroom management. However, the amount of organizational support is also closely linked to the type of activity and task taking place. Therefore, teachers who include practical activities in their lessons in addition to more traditional academic tasks may provide more organizational support than those utilising a greater proportion of more traditional teacher-led instruction. In a recent study, Clarke and Mesiti (2014) show how transition phases in mathematics, for example, also include content-relevant teaching, contrary to what is often described in earlier studies.

Emotional support
The small, almost non-existent, amount of emotional support provided by all teachers is also noteworthy. This lack indicates that the teacher's focus is academic, even in lower primary classrooms. The finding may seem a bit surprising, as we may assume that emotional support may be a part of a positive classroom climate, which is seen as a pre-condition for learning (Pianta & Hamre, 2009). Positive atmosphere and teachers' friendliness were not coded in this study, nor was non-specific praise. These kinds of teacher actions occurred together with actions from other categories and thus could not be coded mutually exclusively, which was a key goal when developing the categories for the present study. The frequent use of nonspecific praise, such as 'Good work, John', is often described as non-productive in research from primary classrooms (Burnett & Mandel, 2010;Haug, 2006). In feedback research, nonspecific praise has been found not to influence pupils' learning positively (Kluger & DeNisi, 1996), but this finding is criticised in recent research because praise (including non-specific praise) may lead to positive emotions, which may indirectly influence the pupils' learning positively (Voerman et al., 2014). From our observations, we suggest that even though the use of non-specific praise may be criticised, in practice it may contribute to what we observe as positive relations between the teachers and the pupils.

Monitoring
The variation in time spent on monitoring by the observed teachers (from 4% to 50%) may be due to several factors. Teachers' monitoring is divided into two subcategories in the framework used in the present study: monitoring working process and listening to pupils read. When Pressley and colleagues (2003) stress the importance of teachers' monitoring pupils, they seem to refer to monitoring working process, as they claim that the best teachers are quickly at the pupil's side when he needs help. This suggests that the instruction, followed by the monitoring, is critical for pupils' learning. The subcategory listening to pupils read is linked to the reading instruction. As we saw in the case of Julie, monitoring may also be influenced by contextual factors, in addition to the teacher's own choices of instructional practices.

No direct interaction with pupils
As emphasised in the section on theoretical framework, four of the main categories in the framework are variously linked to pupils' learning. Interestingly, the category no direct interaction with pupils is an exception. Peter and Ellen, as the only adult in their respective classrooms, spent almost all their time on interaction with the pupils; this may be considered in light of Haug's (2006) findings on number of adults per class in primary classrooms. He finds that two adults in the classroom may relieve the teacher but may not necessarily benefit pupils. Based on our observations, we suggest that spending significant time on non-interaction activities (for instance, sitting by the computer or being out of the classroom) might influence the quality of teaching provided. These classrooms may lack what Pressley and colleagues (2003, p. 165) call 'instructional density'; they stress that the best teachers 'are constantly teaching and providing instruction'.

Conclusions and Directions for Further Research
This study aimed to collect in-depth information about teachers' instructional practices during seatwork in primary language arts classrooms. The analysis reveals considerable differences in the participating teachers' instructional practices during individual seatwork in general and specifically regarding instructional support. Due to the qualitative character and the small sample size of the present study, more research is needed to confirm or challenge the results of this study. More in-depth analysis is also needed of the different instructional practices used in the present study in order to scrutinise our understanding of teacher actions and pupils' reactions. Further research could, for example, include studies of pupils and of their work to understand how teachers' instructional practices support pupils' learning outcomes. Further research should also examine what kind of professional expertise teachers use in their instructional activities. In particular, Ellen's and Anne's support shows how teachers use creativity and imagery instruction to nurture pupils' writing (Egan, 1992;Graham, McKeown, Kiuhara, & Harris, 2012). We also still know little about modelling as an instructional tool in writing classrooms. Research on teachers' instructional practices during seatwork should therefore include studying the pupils' texts to understand how teacher support influences the writing. Teachers' instructional practices during seatwork is part of a careful conversation between the teacher and the pupil for the sake of the pupil's learning. Building on this research, these conversations should be further studied to understand even more of teachers' teaching and pupils' learning.