Pre-service teachers as learners of formative assessment in teaching practice

ABSTRACT This research focussed on the use of formative assessment in Finnish initial teacher education. Pre-service teachers’ perceptions of using formative assessment to promote learning were explored. A convenience sample (N = 60) was taken from two preservice teacher cohorts in which two methods of formative assessment were used during a period of teaching practice. A total of 102 learning reports prepared by the participants were analysed using a three-dimensional framework. The pre-service teachers were found to approach formative assessment from a teacher’s viewpoint; however, it was also used to enhance learning discussions and ensure learners’ active involvement. The objectives and peers’ viewpoints tended to be silenced. Our findings suggest that it is beneficial to provide pre-service teachers with opportunities to practice formative assessment under the guidance of mentors to demonstrate its advantages for students’ learning processes.


Introduction
This study focusses on Finnish pre-service teachers who were given the opportunity to participate in guided practice in using formative assessment (FA) during the teaching practice period of their initial teacher education (ITE).Because the assessment training that occurs within ITE is often quite limited (in hours or credits), focuses on quantitative testing and exams has often neglected FA (DeLuca et al. 2013(DeLuca et al. , 2019;;Jyrhämä 2021;McGee and Colby 2014;OECD 2005), we decided to focus on the formative approach within this research.
The key authorities on FA have thus far been P. Black and Wiliam (1998, 2009, 2018), whose improvement-oriented description of FA is important for us: Practice in a classroom is formative to the extent that evidence about student achievement is elicited, interpreted, and used by teachers, learners, or their peers, to make decisions about the next steps in instruction that are likely to be better, or better founded, than the decisions they would have taken in the absence of the evidence that was elicited.(P.Black and Wiliam 2009, 9) The description integrates assessment as an inseparable part of pedagogy, i.e. teacher's informed and accurate observation, decision-making and feedback regarding students' learning paths.Not only feedback but 'feedback that moves learners forward' (P.Black and Wiliam 2018, 569) is needed.Whilst it is easy to see oral classroom dialogue as the core of FA, formative interaction is also realised in relation to various types of written work, albeit in asynchronous form.Black and Wiliam (2018) designed a three-dimensional model of FA where three actors (i.e.teacher, peer and learner), three processes (i.e.where the learner is going, where the learner is right now and how to get there) and five strategies (clarifying intentions, engineering classroom discussions, providing feedback, two types of student activation) cross-tabulated (see Table 1 below).This model plays a crucial role in our data analysis.
Much debate exists in the literature regarding the power of FA to improve students' learning results, motivation, or progression (Bennett 2011;Lee et al. 2020;Phelan et al. 2011).Although the initial optimism about FA should not be so strong, FA is still very promising from the perspective of pedagogical processes (Pastore, Manuti, and Fausta Scardigno 2019).
International research on the subject has indicated that both teachers and students have positive experiences of FA (e.g.Burner 2016;Cotton 2017;Harrison 2013;Ni Chróinin and Cosgrave 2013).Also, Kyttälä et al. (2022) and Atjonen et al. (2022) recently reported that pre-service teachers in Finland prefer formatively oriented assessment conceptions.However, both these Finnish explorations called for additional education of assessment and FA to be given to pre-service teachers.In the current study, we undertake this challenge by conducting an empirical experiment to examine the use of FA within ITE.
The potential power of FA is in the hands of teachers who can allocate the energy and time necessary to implement it successfully.In contrast, FA may be neglected amidst the pressures of national summative assessment policies and centralised structures (P.Black and Wiliam 2018;Burner 2016;Chen and Brown 2016;Smith 2011).The most recent amendments to the Finnish national assessment norms (FNAE 2020a(FNAE , 2020b) ) simultaneously emphasised support for pupils' individual learning processes (i.e.formative) and established comprehensive subject-based lists of assessment criteria for the final assessment (i.e.summative) at the end of their 9-year basic education.
FA can also be jeopardised by internal pressures.According to Hamodi, López-Pastor and López-Pastor (2017), schools' assessment cultures (e.g.norms, customs, beliefs and values) are not always favourable to FA. Assessment innovations may be rare, and support for FA from principals, colleagues and parents often remains weak.Based on the most recent overview of assessment in education, these concerns are relevant also in the Finnish context (Atjonen et al. 2019).
From a broader theoretical perspective, research on FA also involves teacher assessment literacy (Xu and Brown 2016) that is a part of preservice teachers' professional pedagogical learning; however, our focus is on FA rather than all facets of assessment.

FA and teacher education
We found the concepts of 'FA' and 'assessment for learning' (AfL) to be used largely interchangeably within the existing literature.Some researchers see them as synonymous (see Wiliam 2010) but according to Frey and Schmitt (2007), FA looks at assessment mainly from the teacher's perspective, while AfL looks at it from the student's viewpoint (see also P. Black et al. 2004).Cowie et al. (2018) call AfL as an umbrella term that incorporates many aspects of classroom assessment and activities to generate feedback to support student learning.In our own work, when communicating with pre-service teachers and their mentors during their teaching practice, FA was used because it is the term used in the Finnish curricular guidelines (FNAE 2020a).Black and Wiliam (2018) illustrated a FA-model (Table 1) that introduces three actors and three processes in relation to five strategies (cells 1-5).This model will be used to analyse our empirical data.

Student involvement, feedback, and methods in FA
The table identifies several important topics of assessment; however, due to the scope of our research questions (outlined below), we discuss here only student involvement, feedback, and methods.
One of the key principles of FA is student involvement (Alonzo 2018;P. Black and Wiliam 2018;Leighton 2019).Students are invited to become assessors of their own learning (i.e.self-assessment) and to become resources to other students (i.e.peer feedback).This also means that students must be provided with guided opportunities to practise their self-assessment skills (P.Black and Wiliam 1998), including responding when the results show that they are struggling to meet certain objectives.
A review by Lee et al. (2020) indicated that student-initiated FA seems to be most effective, as measured by effect sizes.On the other hand, Bader et al. (2019) studied Norwegian students and reported that teachers' authority and skills were seen as crucial in the exchange of useful feedback.Taken together, the findings indicate that pre-service teachers need to pay attention to their own FA skills during ITE.
In addition to student involvement, feedback is a key facet of FA (P.Black and Wiliam 2018;Wiliam 2010).Moments of assessment (i.e. a teacher observing a student's action and making an objective-based pedagogical decision) and feedback (i.e.oral or written comments on strengths or weaknesses) follow each other, sometimes rapidly, in the flow of the learning processes.These 'moments of contingency', as Black and Wiliam (2009) call them, are important for the purposes of FA.For example, lower-level skills require different types of feedback compared to higher-level skills, immediate feedback serves different purposes than does delayed feedback, and prospective feedback is more effective than is retrospective feedback (Wiliam 2010).Regarding numerical grades as a form of feedback, Hattie and Timperley (2007) and Shepard et al. (2018) argue that the teacher's task of determining the next steps to take within a lesson requires information that is qualitative rather than quantitative.
Although FA is like a process or a pedagogy, its practical methods are useful to know.If a lesson involves full-time frontal teaching, teachers' cognitive capacity may be focussed only on their own actions.Therefore, teachers need methods or tools that enable students' engagement and learning to become more visible during the lesson (see Table 1: 'elicit evidence of student understanding').These methods must be seen as embedded in the daily pedagogy and not as 'small exams' or 'quizzes', as they were in the first and narrowest interpretation of FA (Bennett 2011).
Several general textbooks, reports (e.g.In the following sections, we will first focus on in-service teachers' pedagogical perceptiveness regarding classroom processes and then explore pre-service teachers' learning about FA.

The role of FA in promoting the pedagogical perceptiveness of in-service teachers
FA can be categorised into three types (Dante and Worrell 2016;Heritage 2007): on-the-fly assessment, planned-for interaction assessment and curriculum-embedded assessment.The first and third are easier for experienced teachers.The first -also called real-time or synchronous assessment (P.Black and Wiliam 2018;Wiliam 2010)-occurs spontaneously during a lesson.In the second and third types, a teacher decides beforehand how to elicit students' thinking during instruction (e.g.questioning techniques for classroom discussion or checking points from broader curriculum topics).
From the viewpoints of synchronous and planned FA, our earlier notion of FA being a process-based approach becomes even more important: FA is not a specific tool or event -it is more like a pedagogy.FA means gathering the information during the everyday pedagogical practices to inform teaching and learning.However, because assessment and pedagogy are not synonyms, it is important to maintain a clear focus on FA as an assessment -that is, as a procedure for making inferences (Bennett 2011;P. Black and Wiliam 2018).In this respect, the theory of FA is narrower than the theory of pedagogy (Wiliam 2010).
Many teachers are not perceptive enough when using process data to make informed pedagogical decisions (G.Black 2014;Heritage et al. 2009).Lyon et al. (2010) illustratedusing a case study of six in-service teachers in the United States -that some teachers were very successful in gathering meaningful data from students but lack the content knowledge necessary to adjust their instruction to meet students' needs.Using sequential analysis, Furtak et al. (2017) also revealed crucial differences in in-service teachers' abilities to utilise so-called 'talk turns' pedagogically during FA processes.Based on the review of 54 research articles, Schildkamp et al. (2020) emphasise teachers' ability to collect different types of data for effective use of FA.Harrison (2013) conducted a qualitative case study of seven in-service teachers in the UK.After reflecting collaboratively on videos, audio recordings, annotated lesson plans and samples of student work, teachers had to strongly reconceptualise their FA practices.A similar rethinking of FA practices was necessary among five Irish physical education teachers whose perspectives on assessment changed dramatically because they understood that 'the children learned more in their physical education classes and that they themselves became better teachers of physical education' (Ni Chróinin and Cosgrave 2013, 230).
Among critical viewpoints to FA, Pastore et al. (2019) reminded that teachers' conceptions of FA may be superficial, and its methods are often confused with summative assessments.Teachers have also been blamed for becoming stuck using assessment methods based on rote, recitation and remembering, instead of dialogue to clarify the state of students' learning and to understand how to enhance it (P.Black and Wiliam 2018).

Previous studies of pre-service teachers as users of FA
In the studies by Nilsson (2013) in Sweden and Tang (2010) in Hong Kong, pre-service teachers' lessons were videotaped and commented on in collaborative seminars.Their AfL interventions improved students' pedagogical content knowledge and contributed to students' learning processes by means of renewed questioning techniques, increased learner engagement and focussed feedback.Throughout the process, many pre-service teachers 'developed an understanding and practice of AfL in a way qualitatively different from their previous understanding and practice' (Tang 2010, 674).
The results of three Spanish studies of pre-service teachers (Cañadas 2021;Cañadas, Santos-Pastor, and Castejón 2019;Hamodi, López-Pastor, and López-Pastor 2017) indicated that FA promoted teaching competences if clear criteria and processfocussed feedback were used.Within ITE, FA-centred assessment courses were appreciated by teachers both during their pre-service education and later, when working as in-service teachers.Although it is crucial that pre-service teachers be provided with research-based FA education, theoretical education alone is not necessarily powerful enough to change conceptions; hands-on practise is also needed (Levy-Vered and Alhija 2018).
von Aufschnaiter and Alonzo (2018) summarised three types of difficulties that preservice teachers face in interpreting their students' learning processes: they may lack familiarity with common student ideas, be unable to identify students' learning needs and be likely to make holistic judgements about student thinking as 'right' or 'wrong'.As Hopster-den Otter et al. ( 2019) noted, correct-incorrect categorisation works against the idea that FA must provide fine-grained knowledge rather than judgemental reflection on students' capabilities.
Two studies of pre-service teachers of physical education were also found.In their follow-up of three Swedish preservice teachers, Tolgfors et al. (2022) found that various contextual dimensions of schools provided different conditions that either support or hinder the use of AfL.Macken et al. (2020) reported from Ireland that the use of AfL during school placement required time to practice and develop.

Research questions
Only two groups of researchers in Finland have yet shed valuable light on our research topic.Atjonen et al. (2022) found that pre-service teachers desire opportunities to see how FA works in practice.Kyttälä et al. (2022) identified several assessment-critical preservice teachers who disagreed with the AfL approach and had quite a one-sided conception of assessment, viewing it as purely involving summative grading.These recent Finnish explorations motivate our empirical experiment related to teaching practice.
We formulated our research question (RQ) as follows: What kinds of perceptions of FA do pre-service teachers report in promoting the learning of a) themselves as teachers and b) their students?To answer the research question, we provided the pre-service teachers with practical guided opportunity to put FA into practice.

Participants
The experiment took place in a research-based Finnish university that has its own teacher training school (TTS), as all ITE units in Finland have.The FA-focussed experiment was included as part of the last teaching practice period of the five-year ITE.The participants were in the final phase of their MA degree in a two-tier ITE programme (Bachelor 3 years + Master 2 years).During a period of eight weeks, pre-service teachers used FA during two lessons and reflected on them afterwards.
In total, 67 pre-service teachers participated in the FA-focussed training period in 2021 and 2022 (March -May in the spring terms), and 60 of them eventually submitted their learning reports.This accounted for approximately one third of all pre-service teachers enrolled in these teaching practice periods.The total number of learning reports included in the data was 102.
The process for recruiting participants in this study went as follows: Seven voluntary mentors at the TTS accepted our invitation to participate in the spring term of 2021.In 2022, six of them wanted to participate again, and nine new mentors joined in.The preservice teachers who were randomly placed on teams led by these mentors became participants in the FA experiment.All mentees of the teams used two FA-methods, but it was not obligatory to participate in the research.Before the practice period, the mentees were asked to give written permission if they allowed their learning reports to be used anonymously for our research purposes.Only two mentees refused.

Data collection
The pre-service teachers received a list of 49 FA methods that could be used (HEA 2015) and written advice about completing their learning reports.They were advised to choose two FA-methods that were suitable for the objectives of the lessons, the school subjects being taught and the students' ages.The TTS mentors facilitated planning when necessary.Because Oldfield et al. (2012) and Looney (2019) see digital methods particularly suitable for FA purposes, the first method was suggested to be non-digital and the other digital.
The data were gathered by means of structured items and open-ended questions, which formed a document called 'Learning Report'.Several background items were presented first, including the school subject, names of the FA methods being used and assessment target or learning objective of the lesson.These background items were followed by the open-ended questions as follows: 1) Why did you choose this method?2) How did the method work? 3) What kind of knowledge did you gain from it?4) How did FA effect your pedagogical decisions?Additional knowledge gained from questions 5 and 6 (i.e.How successful was the digital version of FA? Did this experiment change your conception of assessment?) is also used in our analysis.
Because the pre-service teachers were asked to use two FA methods and to report on them separately, the participants replied to the above questions twice.In addition to preservice teachers' own replies, their TTS mentors also expressed key observations regarding the FA experiment.During the final mentoring session of the teaching practice, the experiences were discussed between mentors and their mentees.Finally, the pre-service teachers sent their learning reports to the second author of this article.

Data analysis
Our analysis focussed on written data obtained from the 102 learning reports.The longest report consisted of 514 words, while the shortest contained 114 words.The scope of the fully anonymised data was 28,025 words.
All the text from the open-ended questions 1-4 of the learning report was first analysed as a whole, to gain a full account of the experiences.Keeping in mind the key theoretical ideas presented earlier (section 'FA and teacher education'), comprehensive inductive analysis was done first to dig in the qualitative diversity of perceptions.Thereafter, we were able to use the three-dimensional model designed by Black andWiliam (2009, 2018; Table 1) to fully analyse the data.
Our results are based on deductive, theory-based analysis where the main emphasis is on qualitative description of FA-perceptions but also frequencies of utterances are mentioned (see Azungah 2018).The analysis unit was defined as an utterance with a single meaning.The units often consisted of a sentence or two.The content analyses were done by means of the ATLAS.tiWeb software.Based on negotiations among the authors regarding the coding protocol and principles of interpretation, the second and third authors of this article took on the main responsibility of the analysis.The abbreviation 'LR' in data citations refers to 'learning report'.

Introductory information of the experiment
An initial analysis of the subjects and age groups showed that FA was used most often in mathematics and science (including environmental studies; mentioned in 37 reports) and expressive arts (including crafts, music and physical education; 22).The subjects of Finnish language (26) and the humanities (including history and social sciences; 17) were also well represented in the data.Half of the 60 pre-service teachers used FA when teaching first and second graders (i.e.students aged 7-8 years), one fourth used it among third and fourth graders (9-10 years old), and one fourth used it among fifth and sixth graders (11-12 years old).
Regarding FA-methods, non-digital tools were mentioned in 49 learning reports.Most of them were quick FAs that were easy to do 'on the fly', such as the activities 'traffic lights' and a thumbs-up vote.Also ABC cards and exit slips were used as much as interactive discussions.Digital applications were described in 53 reports.They were divided into games (e.g.Kahoot or Socrative), producing (e.g.Padlet, Seesaw or Book Creator), surveys (e.g.Survey Monkey, Mentimeter or Google Forms) and learning analytics; of these, producing was the most popular (31 reports) type compared to surveys (12) and gamification (9).

Overview of FA-perceptions
Following the framework of Black and Wiliam (Table 1), Table 2 presents our key results regarding pre-service teachers' perceptions of FA.The first quantitative notion in Table 2 indicates the strong dominance of the teacher's perspective (row = 325).Simultaneously, the pre-service teachers were able to look at the future of learning processes (column = 241) Among the five FA strategies, giving feedback (i.e. the light grey cell in Table 2) and engineering and eliciting learning discussions (the black cell) were most often mentioned by our participants.Regarding students' activation, the pre-service teachers were able to promote students' ownership in learning (the lined cells) but almost totally ignored the students as each other's educational resources (the dark grey cells).
We will next look qualitatively at each cell in detail.From a broader perspective, the following sections shed light on the developing state of pre-service teachers' assessment literacy (Atjonen et al. 2022;Xu and Brown 2016).

Where the learner is going
Only a few pre-service teachers initially set goals for learning, addressed them with the students and assessed their implementation, particularly towards the end of teaching.Some also highlighted the need to open up ways for learners to achieve objectives.Otherwise, the prospective teachers neglected the feedforward purpose of the assessment.

Where the learner is right now
The pre-service teachers perceived the identification of the students' knowledge base as a function of FA.This was typically named as the assessment of student competence in a certain topic (31 utterances).For example, one report included the following: 'The students came up with newspaper headlines about things studied in the lesson.Chromebooks were used here.The subject of the assessment was how well the students had internalised the topics studied in the lesson' (LR22).
The pre-service teachers also aimed to determine how the learned content was assimilated (24).They noticed that FA helped them to gain knowledge about individual students' understanding (10) and learning (6).They perceived the purpose as getting a broad picture of the group's learning or knowledge (15).FA also revealed how students' knowledge base was evolving ( 4) and what kinds of preconceptions of the topic they had (4).
The pre-service teachers raised the identification of students' experiences as an important aim of FA.They learned about students' experiences of their individual (10) and group work (3), their learning (4) and their know-how (2).The identification of students' feelings (4) was also mentioned in the reports, as in the following example: 'Information was also provided on the students' perceived feeling about the fluency of their own work regarding crafts lessons' (LR30).
Connected to the knowledge base, the identification of difficulties in learning certain content was mentioned (13).In some cases (3), FA facilitated observations of whether a certain task was difficult for a learner.Some of the students had not been able to specify what they have learned, and they wrote that everything was easy.On the other hand, some reports included very interesting and accurate information -for example: 'I have understood the connection between percent, fractions and decimal numbers' or 'I can't do task number 7 from page 6'.(LR41)

How to get there
Among the five FA strategies, future-oriented feedback was the most often discussed (the dark grey cell in Table 2).When anticipating and reviewing pedagogical solutions, the pre-service teachers emphasised FA's importance for teachers' work planning (50 utterances).They explained that 'With the help of selfassessments, it is possible to make pedagogical solutions for the following lessons, such as how [when teaching the Finnish language] to practice and recap plurals and units' (LR04) or 'This FA method made it easier for me to plan for the next lesson because I could see immediately if any of the essential learning objectives had been eclipsed' (LR18).
Along with these, the pre-service teachers noted that FA facilitated information about reaching the objectives of lessons ( 7) and gaining an overview of the lesson (7).FA informed the teachers of students' opinions regarding the various teaching methods used (10).All in all, the pre-service teachers were motivated to create rich and versatile assessment situations and to try new assessment methods.In one learning report, FA was perceived within the context of professional development: 'With the help of observation, the teacher is also able to gather information about what works and what doesn't, so that she can develop herself professionally in a better direction' (LR62).
Helping teachers to understand learning processes was also highlighted as one of FA's advantages.This included simply collecting information ( 13) for pedagogical decisionmaking and, later, consideration.Some pre-service teachers described FA as helpful for the knowledge-intensive follow-up of working (10) and learning (9) processes and their phases (4): 'The method allows the student's reading to be recorded and to be revisited.Recording makes it easier to track reading fluency progress' (LR06).Understanding and reviewing students' self-assessment skills were also often (22) discussed as valuable advantages of FA.
The pre-service teachers gained information about the support needs (43 utterances) on both the classroom and student level.These varied from task-and content-specific tutoring and helping individual students to the classroom-level design and how to match the teaching to emerging student needs: 'The teacher was able to guide students to recap the learned things and to reteach some information in a more manageable pieces' (LR24).Additionally, some pre-service teachers reported the use of FA to create ability groups based on the gained learning results.

Where the learner is going
As in the teacher point of view, setting goals for learning was almost disregarded when the FA was analysed from the learner point of view.Some pre-service teachers noted that, with the help of FA, students learned to set goals for learning, or they saw it as important that learners revisit the set objectives and assess their success according to the objectives: 'I felt that . . . in self-assessment, they can reflect on their own behaviour and notice their own successes and points of improvement' (LR55).
Additionally, some pre-service teachers emphasised that the FA experiment promoted the planning of students' work and the observation of how students' preconceptions of the learned topic may change, as seen in the following two quotations: 1) 'Students assessed their own success and planned their own work through the project.They set goals to be completed and evaluated their own success relative to the goals they set' (LR15); 2) 'In the beginning, students pondered how well they think they know the topic of the lesson according to their preconceptions' (LR20).

Where the learner is right now
Identifying and developing learners' knowledge was emphasised as an aspect of FA with remarkable potential.Through FA, learners can be involved in assessing their own competence in certain topics (26 utterances).In some experiment cases, learners assessed competence with either 'I can' or 'I can't', while in other cases, it was labelled 'challenging', 'suitable' or 'too easy'.Students also self-assessed their competence more freely with the help of some specific tools, as is seen in the next data excerpt: After reading, students listened to their own recordings and reflected on the fluency of their own reading using the 'Assessment Land' [large imaginary painting printed on a poster].Students videorecorded the 'Assessment Land' and the sign [plastic symbol] they had placed in it and explained how they thought the reading had gone.(LR17) Learners were also asked to describe verbally their knowledge of certain topics: Students raised their thumbs up if they could explain what the term meant.Thumbs down meant they couldn't explain, and thumbs to the side meant they were able to partially explain . . .The correct answer was still verified/asked from a student whose thumb was up.(LR21) The learners' motivation was often reflected in the learning reports (14 utterances).The pre-service teachers chose FA methods because they 'work as an inspiring and motivating activity' (LR12) for learners.The digital applications used within FA, such as the Kahoot game and digital portfolio were perceived as a particularly motivating element for students.
The prospective teachers exemplified how the FA experiment provided an opportunity for the students to review what they had learned and to receive tips about what should be practiced more (two quotations): 1) 'It was wonderful to see how, when reflecting on the word, they recounted the concepts they learned in the lesson as well as the skills to find the word' (LR42).2) 'The assessment was carried out in the last lesson before the test, at the end of the lesson.As a result, students were left with the opportunity to practice and recap their own weaknesses and to reinforce strengths in the final refresher lesson' (LR52).

How to get there
FA as a tool for identifying a learner's own progress was mentioned often.In many reports, this was expressed as encouraging learners to share their overall experiences of work success (20 utterances).In contrast, some reports highlighted, that learners noticed their own progress in learning or working (7) or understood their learning process better (4): 'In soft materials in crafts education, students document at the end of every lesson their unfinished product by photographing it and adding it to digital notebook in the BookCreator application' (LR31).
Promoting learners' self-direction was perceived as a key feature of FA.Learners gained opportunities to reflect on their own work during lessons (8 utterances).FA was also a means by which to document the working process (6): 'to document one's own working as a process in crafts education is also important, as it gives data about the progress and fluency of one's own working for students and teachers' (LR17).Additionally, developing self-and peer-assessment skills and promoting writing skills (or verbalising skills) were seen as important motives for using FA.
The ability of FA to give opportunities for participation was also recognised by the preservice teachers.They wanted to encourage learners to reflect on their own participation (7): 'A thumbs-up vote is a so-called low-threshold assessment tool for students that still makes students reflect on their own behaviour, work and learning' (LR52).They wanted learners to feel that they had participated in the learning process and had been engaged in the work.The development of students' agency and student-centredness were also discussed in few reports.

Peer or significant person point of view
The peer perspective of FA was seldom discussed (8 utterances) in the pre-service teachers' learning reports.These comments emphasised the importance of peer feedback in a relevant way: 'students get the opportunity to rehearse giving peer feedback' (LR54).The FA experiment also facilitated the sharing of ideas ( 2) between peers about how to develop their learning according to the objectives.In some cases, students informed their peers by means of an FA method that they needed help or wanted to discuss the topic (3).Only two pre-service teachers noticed how FA promoted the sharing of knowledge about students' learning with guardians.

Discussion
Our research focussed on Finnish pre-service teachers' use of FA during their teaching practice in ITE.According to the results, the pre-service teachers' perceptions of FA tended to focus on the teacher's viewpoint, and they did not analyse peer assessment much.They understood the power of FA in giving feedback on their teaching and were able to discuss learners' processes and need for support.They dealt with the question of 'how to get there' mainly from the viewpoint of teacher's decision-making process but not so often from the learner's viewpoint.They fluently used both digital and non-digital tools.Regarding digital applications, they also used broader publishing of students' artefacts.

Reflections of findings
As indicated in the literature (Atjonen et al. 2019;DeLuca et al. 2021), FA is not yet well known or used due to the dominating summative practices in the countries of highly centralised assessment system.This was also true among our experiment participants, although Finland relies on low-stakes assessment and local development based on broad national assessment guidelines.Our participants reported on FA as a 'new' assessment perspective that challenged their previous personal conceptions of summative assessment.However, our focus on FA does not mean that we see it as standing opposite to summative assessment -rather, it has its own important functions.
Although FA is not predominantly a fixed set of methods, the pre-service teachers' perceptions of FA indicated that even a single method may be valuable in opening their eyes to see students' learning more accurately.It would be a methodological challenge to measure this change in a study applying any research design, however justified the demand would be from those who are interested in learning results and effect sizes (see Bennett 2011;Lee et al. 2020;et al. 2011).
In our study, the participants certainly faced challenges -such as those related to lesson time allocation and technical obstacles with iPads -but these did not destabilise their trust in FA.Many were even relieved to use it, finding that FA is not as complex activity in practice as one may anticipate.They did not report any cultural contradictions (Hamodi, López-Pastor, and López-Pastor 2017) or external summative pressures (Chen and Brown 2016), probably because the FA experiment took place in a TTS as part of Finland's innovation-oriented, research-based ITE.Researchers (e.g.Maclellan 2004;Pastore, Manuti, and Fausta Scardigno 2019;von Aufschaiter 2018 andAlonzo 2018) have previously noted naïve or inaccurate conceptions of FA among pre-and in-service teachers, but our self-reporting method revealed no such misconceptions among our participants.This could be an interesting avenue for further research on FA.
There were remarkable differences between individuals in terms of their ability to look carefully at teaching-learning processes.This is in line with several international studies (Cañadas 2021;von Aufschneider and Alonzo 2018;Kyttälä et al. 2022;Lyon, Nabors Oláh, and Wylie 2010).It was challenging to gather pedagogical knowledge spontaneously or to recognise 'moments of contingency' because the rapid ongoing events were complex in the novices' eyes (P.Black and Wiliam 2009;Luckritz Marquis 2021).Despite this, planned FA (P.Black and Wiliam 2009;Dante and Worrell 2016) is also valuable, and our participants reported how their observations of their most recent lessons helped them in planning the next teaching-learning event.
Although we did not focus our research questions on the teachers' assessment literacy, as such, the results can be discussed within a broader framework.For example, Xu and Brown (2016) identified seven elements that should be part of a teacher's knowledge base in relation to their assessment literacy (e.g.content knowledge, assessment methods, feedback, self-and peer assessment and communication).Practically, all of them were touched upon in the FA perceptions of the pre-service teachers in this study.
As a part of assessment literacy, the perceptiveness to observe and activate learners is highly valuable in the ITE phase.New qualitative explorations (see Furtak, Ruiz-Primo, and Bakeman 2017) are needed to demonstrate how teachers learn not only to give feedback (i.e.directed backwards) but also to use techniques of feedforward (i.e.directed onwards) in facilitating students' learning.This may result in new insight into learners' assessment literacy as well.
From national viewpoint of Finland, our experiment encourages teacher educators to provide preservice teachers with versatile opportunities to learn about the power of FA in ITE.The in-service assessment education of teachers must also be invested in as is already done nationally (see https://www.kaaro.fi/en9).From European or global perspective, FA deserves special attention from the viewpoint of pedagogical support of students' learning in the era of growing supranational waves of evidence-based assessment by means of summative measurements.

Limitations and suggestions for further studies
Several explorations have indicated that pre-service teachers' views of FA are somewhat contradictory (Cotton 2017;Harrison 2013;Kyttälä et 2022).In our data, 19 learning reports (out of 102) discussed its easy use, 11 about the enhanced understanding of assessment, 13 about new assessment methods and 18 about interesting options for digitalised FA.In contrast, 16 included the idea that the FA experiment did not have an impact on the understanding of assessment; however, several reasons were simultaneously listed about why it was still useful to participate in.We did not analyse the mentors' comments in detail, but they confirmed their mentees' self-assessments of the use of FA.Because it was not possible to protect the pre-service teachers' anonymity during mentoring, the most critical experiences may have been silenced.
Although the use of a convenience sample limits options for recommendations, our experiment seemed to promote an understanding of FA options among prospective teachers who do not yet have much teaching experience.Their enhanced perceptions of FA's power to increase students' self-assessment and self-reflection (P.Black and Wiliam 2018;Chen and Brown 2016;Leighton 2019;Luckritz Marquis 2021) are important findings.While our experiment was not resource intensive, its positive results are in line with those of Tang (2010) and Nilsson (2013), who used more intensive interventions.We touched upon digital FA-methods only lightly, but digitalisation definitely deserves more focused explorations (see e.g.https:// www.assessforlearning.eu/).