Group work assessment intervention project—A methodological perspective

Abstract The assessment of individual knowledge and abilities should be frequently undertaken when learning is developed in interactions with other students, such as in group work and/or cooperative learning. Previous research reveals that group work assessment is a neglected research area, and this applies in particular to group work assessment interventions studies. The focus of this article is methodological, and its aim is to provide a reflective and critical account of a group work assessment intervention project, and the implications of the different choices made in this process. The intervention project that was scrutinized had a mixed-method longitudinal quasi-experimental design, and interventions in the form of shorter educational sessions were central to the project. Both qualitative and quantitative data were collected, analyzed, and compiled. The methodological issues discussed and problematized were the importance of (a) establishing collaboration with teachers; (b) well-thought-out and delimited methodological choices, and subsequent consequences; and (c) including both teachers and students to secure successful effects of the interventions. As a result of the study, it was concluded that intervention could be beneficial as a means of increasing the scientific knowledge in relation to intervention studies, and also to the emerging discourse on group work assessment.


ABOUT THE AUTHOR
Eva Hammar Chiriac is a professor of Psychology at the Department of Behavioural Sciences and Learning, Linköping University, Sweden. Her scientific activity lies within the social psychological research field with a strong focus on group research, mainly connected to groups, group processes, learning and education. Her current research project concerns group work assessment, school climate and relations in schools, and problem-based learning. Karin Forslund Frykedal, PhD, Professor of Education at the Department of Social and Behavioural Studies, University West. Her scientific activity lies within educational research with a focus on leadership, group processes and learning in small educational groups. Her current research project concerns group work and group work assessment in educational context and parental education groups in primary health care.

PUBLIC INTEREST STATEMENT
Many teachers express feelings of uncertainty about how to make group work assessments and experience group work assessment as a complex and challenging issue. The teachers are uncertain about the purpose of the assessment, what should be assessed and by whom, and whether the assessment should be made at individual or group level. One special challenge concerns the difficulty of disentangling one individual student's knowledge and abilities from the groups joint work when assessing. The focus of this article is methodological, and its aim is to provide a reflective and critical account of a group work assessment intervention project, and the implications of the different choices made in this process.

Introduction
Since as early as 1990, there has been scientific support for intervention in educational settings as a means for developing pedagogical practice in cooperative situations (Ashman & Gillies, 1997;Gillies & Ashman, 1996). Employing short educational interventions yields positive results in both group work efficiency and productivity. These results have since been confirmed by a number of studies showing that a relatively small effort generates positive results for work and learning in groups or cooperative leaning (CL), (Gillies, 2016;Johnson & Johnson, 2002, 2015Martinez Lirola, 2016a, 2016bRoseth et al., 2008;Slavin, 2014). Correspondingly, research has shown that group work assessment greatly influences both efficiency and productivity in group work/CL Hammar Chiriac et al., 2019b;Johnson & Johnson, 2021;R. A. Lotan, 2003;Slavin & Madden, 2021). Although we have not been able to find any intervention studies demonstrating increased knowledge and ability in relation to group work assessment, we argue that the same applies to it. Hence, our starting point is that intervention can also be beneficial as a means of developing scientific knowledge and the pedagogical practice vis-á-vis group work assessment.

Background
On many occasions in educational settings, an assessment of individuals' knowledge and abilities need to be accomplished in situations where learning is developed in interactions with other students in a social context, such as in group work/CL. While group work is customarily defined as "pupils working together as a group or a team" (Blatchford et al., 2003, p. 155), CL is a specific approach to group work where cooperation with clear structures, enhanced student activity, and common goals are the basic prerequisites for learning (Ferguson-Patrick & Jolliffe, 2018;Gillies, 2007;Kagan & Stenlev, 2013). Although we acknowledge the differentiation between group work and CL, they are both pedagogical methods for organizing classroom activity where students' learning is developed from a joint activity (Barkley et al., 2005;Davidson, 2021;Davidson & Major, 2014;Frey et al., 2009) and thereby faces similar complexities and challenges when it comes to assessment (Barblett & Maloney, 2010;Brookhart, 2013;Johnson & Johnson, 2004). Hence, group work and CL can be discussed interchangeably in connection with assessment.
A comprehensive requirement in several countries, and at several levels in the education system (e.g., Brookhart, 2013;Johnson & Johnson, 2004), is that teachers are expected to assess each student's level of knowledge and ability in relation to knowledge requirements. This goal-and criteriarelated assessment system is also required as means for group work assessment. However, reconciling the demands of individual assessment while fulfilling the requirement to teach cooperative abilities (Johnston & Miles, 2004;Ross & Rolheiser, 2003;Swan et al., 2006;Webb, 1997) poses a challenging and complex task for teachers at all levels in the educational system (Brookhart, 2011(Brookhart, , 2013Forsell et al., 2020Forsell et al., , 2021Forslund Frykedal & Hammar Chiriac, 2011Martinez Lirola, 2016a, 2016bvan Aalst, 2013). Group work assessment (the concept is henceforth used for both group work assessment and CL assessment) also entails additional dilemmas, compared to traditional individual assessment, such as process skills becoming salient assessment criteria, and the difficulty of disentangling each individual's knowledge and contribution from another's. In summary, when conducting group work assessment, teachers are faced with new challenges for which they are not always adequately prepared. One way of reducing this problem could be an intervention study regarding group work assessment to examine a small educational effort.

Intervention studies in connection with group work and/or CL
Most intervention studies in connection with group work and/or CL have focused on (a) implementation and effects (e.g., Ferguson-Patrick & Jolliffe, 2018;Healy et al., 2018;Martinez Lirola, 2016a, 2016band (b) identifying key factors for successful group work/CL (e.g., Gillies, 2016;Gillies & Boyle, 2010Johnson & Johnson, 2002, 2015Roseth et al., 2008), reporting several positive outcomes. Research which included comparing groups who received an intervention with control groups showed that interventions connected to group work and/or CL: (a) promote academic knowledge, skills, and understanding (e.g., Ferguson-Patrick & Jolliffe, 2018;Gillies, 2016;Johnson & Johnson, 2002, 2015Martinez Lirola, 2016a, 2016b; (b) enhance students' achievement, socialization, motivation, and personal development (e.g., Roseth et al., 2008;Slavin, 2014); and (c) increase CL achievement compared to individualistic and competitive learning (Johnson & Johnson, 1994, 2002Johnson et al., 1981). Although most studies on CL have emphasized the positive outcomes resulting from its employment, there is also research highlighting disagreements on, or criticism of, CL as an application and/or method (Davies, 2009;Healy et al., 2018;Johnson & Johnson, 2021;Opdecam & Everaert, 2018). Reduced learning, low-quality group work, and negative experiences of cooperative situations may come from students and teachers' perceptions of, and beliefs about, CL (i.e., the reduction of grading-time and negative group processes). Further, it may also be problematic if teachers believe that (a) assigning students to groups will automatically lead to learning based on cooperation, (b) peer assessment solves the group work assessment problem, or (c) CL is easy to organize and guide.
Two very relevant examples of intervention studies came to serve as inspiration for the design and implementation of this project. The first was Gillies and Ashman's (Ashman & Gillies, 1997;Gillies & Ashman, 1996) study investigating whether teaching cooperative skills to students working as groups could increase their ability to use the groups' potential. The results showed that a relatively small effort yielded positive results in both group work efficiency and productivity. The second study was Black et al.'s (2003) research on changes in teachers' attitudes regarding the introduction of formative assessment. The study focused on how researchers in collaboration with teachers developed and implemented formative assessment, by using an educational intervention. The results indicated that the teachers appropriated the benefits of using formative assessment.
Against this backdrop, we aim to contribute by increasing the knowledge concerning intervention studies on group work assessment in cooperative situations. We will accomplish this by focusing on methodological aspects, and by providing a reflective and critical account of an intervention project, and the implications of the different choices made in this process.

Previous research on group work assessment
Teachers often highlight the perceived challenges in assessing students' knowledge and abilities in group work and/or CL, and this has been correspondingly problematized in the sparse research of group work assessment (R. A. Lotan, 2014;Forsell et al., 2020;Forslund Frykedal & Hammar Chiriac, 2011R. A. Lotan, 2003;Ross & Rolheiser, 2003;van Aalst, 2013). Previous research has problematized group work assessment by emphasizing issues of concern connected to the assessment loop, what is primarily assessed (academic knowledge or process skills), and the level on which the assessment is conducted (individual or group; Brookhart, 2013;Johnson & Johnson, 2004, 2021Meijer et al., 2020;Ross & Rolheiser, 2003;Strijbos, 2016;van Aalst, 2013). Both issues seem to generate different attitudes closely connected to assessment (Rosander et al., 2020). Some researchers advocate that, regardless of whether the learning is developed in a group situation, it is still the individual student's knowledge that should be assessed and graded (Brookhart, 2013;Forsell et al., 2020Forsell et al., , 2021Johnson & Johnson, 2004). A common argument, often emphasized in research, concerns the difficulty of disentangling individual's knowledge from each other; instead, group work assessment is often based on the student's participation and contribution to the group. Thus, process skills and group work abilities, which require cooperation, constitute the most common substrate for group work assessment (Forsell et al., 2020(Forsell et al., , 2021Martinez Lirola, 2016a, 2016bvan Aalst, 2013). Skills required both in the educational context and in future working life are thereby also worth assessing, but whether or not these skills should be graded has divided opinions (Brookhart, 2013;Frey et al., 2009;Martinez Lirola, 2016a, 2016b. One suggestion proposed is to assess learning and process separately and differently. Other critical issues presented in previous research regarding group work assessment concern (a) how to assure quality and fairness (e.g., Kablan, 2014;Orr, 2010;van Aalst, 2013), thereby enhancing the probability that students will regard the assessment as fair, and (b) who is in control of the assessment loop (e.g., Forslund Frykedal & Hammar Chiriac, 2011Ross & Rolheiser, 2003), given that peer assessment is often used as the ultimate solution for problems as regards group work assessment (Johnson & Johnson, 2004) and the teacher is relatively absent in the assessment process (Forsell et al., 2020). In addition, other issues-albeit not exclusive to group work assessment only-concern (c) the purpose of the assessment-formative or summative (Black, 2013;van Aalst, 2013;Wiliam, 2009Wiliam, , 2011, and (d) finding methods for collecting empirical evidence that can provide the basis for the assessment (Dijkstra et al., 2016;Meijer et al., 2020;Messick, 1989).
Accordingly, previous research has not only identified the fact that group work assessment is problematic, but that it is also a neglected research area (e.g., Forsell et al., 2020;Johnson & Johnson, 2004;Ross & Rolheiser, 2003;van Aalst, 2013), and, consequently, there is a need for research in this area. The use of an intervention could be one way to improve teachers and students' knowledge and ability to assess.

Social Interdependence theory and group worthy tasks
As the prevailing theory in CL (Deutsch, 1949;Johnson & Johnson, 2002Lewin, 1948), the social interdependence theory (SIT) is used as the theoretical framework for this project. According to SIT, group members develop a degree of interdependence when they realize that working together on a task can increase the probability of them achieving common goals. Interdependence can be positive, and this creates opportunities for cooperation among group members. Furthermore, positive interdependence causes conditions for increased individual accountability. Interdependence might also be negative, and this can create competition among group members. There might also be an absence of interdependence, signifying that students may reach their goals independent of others in the group, and this does not provide opportunities for interaction and cooperation between group members (Johnson & Johnson, 2002. SIT emphasizes the conditions necessary to increase the cooperative potential of groups: positive interdependence (i.e., the perception of being linked to other group members and the realization that it is achieved through the pursuit of common goals and joint rewards), individual accountability (i.e., each group member is responsible for their share of the work, and has a willingness to help other group members); promotive interaction (i.e., group members encourage each other's efforts through discussions and explanations), interpersonal and small group skills (i.e., enhancing the degree of trust among group members and their ability to resolve conflicts), and group processing (i.e., group members discussing and evaluating their work). According to Johnson (2002, 2013), these conditions are crucial for promoting, affirming, and maintaining effective working relationships among members. The presence of positive interdependence and individual accountability are two key features often stressed as defining CL in comparison to conventional group work (Ahmar & Mahmood, 2010;Brookhart, 2013;Johnson & Johnson, 2015, 2021Martinez Lirola, 2016a, 2016bSlavin, 2014).
The importance of designing a group task suitable for learning and working in interaction with other students is well documented in previous research (Barkley et al., 2005;Cohen & Lotan, 2014;Forslund Frykedal, 2008;Hammar Chiriac, 2008;Johnson & Johnson, 2004). Less discussed and problematized is the fact that the design of the task also acts as a guideline for the assessment loop (i.e., what type of assessment it will be possible to conduct, and in what way). Lotan's (2003Lotan's ( , 2014) model for creating a group worthy task may support teachers in planning and implementing group work/CL, and also by extension the assessment loop. The following five characteristics are highlighted in the model: A group-worthy task (a) is open (i.e., it can be solved in more than one way) and requires problem solving; (b) provides students with the opportunity to use wide-ranging knowledge and abilities; (c) focuses on vital subject-based issues; (d) presupposes positive interdependence and individual accountability; and (e) provides explicit assessment criteria.
There are several possible theories and models suitable, and we have chosen SIT and the idea of the group worthy task to supply and increase knowledge concerning a methodological perspective on the phenomenon, which is the subject of this study.

Method
Inspired by earlier intervention studies of group work, which revealed that shorter educational sessions developed cooperative practices (Ashman & Gillies, 1997;Gillies & Ashman, 1996), we wanted to investigate whether similar shorter educational sessions with teachers and students could have comparable effects on group work assessment. We used a mixed-method longitudinal quasi-experimental design, where both qualitative and quantitative data were collected, analyzed, and compiled. Interventions in the form of shorter educational sessions were central to the project. See, Figure 1 for an overview of data collection, intervention, and group work in each classroom.

Participants
The study was conducted in compulsory schools when students in years 5 and 8 were doing group work in mathematics. Our choice of academic years was based on (a) whether the students were old enough to be able to decide for themselves whether to participate or not, (b) teachers and students were familiar with each other, and (c) the teachers were required to give an assessment in both years and also a grade in year 8. The participants came from six classes attending five different schools in Sweden. In order to promote variation, schools and participants from different social and geographical areas were included. In total, 140 students participated in the study. The classes included in the study featured both heterogeneous and homogenous groups of students (Table 1 and Figure 2).

Data collection
Three different data collections methods were used to capture various perspectives of the participants on their actions and experiences of group work assessment. To capture teachers and students' own perspectives, qualitative data were collected through interviews and observations. Quantitative data were gathered through questionnaires to capture the participants' general experiences of different parts of the assessment loop. The various data collections, with the respective analysis, will each individually but also together, contribute with knowledge concerning intervention studies of group work assessment in education. This diversity of data collection enabled the generation of different knowledge with regard to intervention studies of group work assessment in education.

Interviews
To elucidate teachers and students' experiences of group work assessment, both interviews and focus groups were used. Utilizing interview guides with open-ended questions provided the opportunity to study in-depth teachers and students' insider perspectives. Hence, the use of interviews and focus groups enabled participants to express self-chosen aspects using their own words, based on the project's issues. We also wanted to capture if and how any changes occurred (a) after participating in the project and the interventions, or (b) after participating in the project but not the intervention (Figure 2). All interviews were audio recorded and transcribed.

Observations
In order to collect data of teachers and students' actions and communications in practice, observations were carried out with video recordings during the different parts of the group work assessment project (Figure 1). Using these observations, the intention was to study how group work assessments were implemented, and to examine if any changes occurred during the group work. Unlike other data collection methods, observations enable the collection of information on actual performance, in contrast to interviews and questionaries, where the participants themselves decide what information to share.

Questionnaires
In order to capture the participants general experiences of different parts of the assessment loop, two multiple choice questionnaires were applied (Figure 1). The intention of the first questionnaire, distributed only to students, was to measure which factors influenced their attitudes towards group work and group work assessment, and whether any changes occurred after participating in (a) the group work and intervention, or (b) the group work but not the intervention (Figure 2). The second shorter assessment questionnaire was distributed to both teachers and students after each group work session (i.e., six times; Figure 1). The purpose of this questionnaire was to capture the participants' perceptions of (a) what was assessed, (b) what type of assessment was conducted, and (c) who carried out the assessment. Further, the purpose was to study whether teacher and/or students' perceptions varied during the different sessions. When responding to multiple choice questionnaires, the participants answered the same pre-assembled questions by the researchers, which entailed comparisons between individuals and/or occasions. The individual participants responded anonymously to both questionnaires, but the group membership was recorded so that comparisons between sessions were possible.

Interventions
The project consisted of two interventions in the form of short educational sessions. These were administered (a) to both teachers and students, or (b) exclusively to teachers. The interventions were randomly implemented in four classes (i.e., two in Year 5 and two in Year 8), and the two other classes (i.e., one in each year) served as the control group and did not participate in any intervention (Figure 2). The interventions comprised education on group work/CL and group work assessment.
The teachers' interventions were implemented on two separate days, with two weeks in between. The first day was theoretical, and contained theory and discussions on group work/ CL as a teaching strategy (e.g., how to create cooperation and group worthy tasks), and group work assessment (e.g., different types of assessment strategies applicable to group work/CL). The second day of the teacher intervention had an applied approach, where researchers and teachers jointly examined various ways of collecting empirical evidence that could provide the basis for the assessment, and how to document this information. In order to be able to problematize various types of assessment strategies, both proactive and reactive ways of working with the assessment, as well as giving and receiving feedback, were considered. The final task for the teachers was, in cooperation with a researcher, to produce an appropriate group worthy assignment usable for both Years 5 and 8, respectively, and also to write a pedagogical plan for the task. In addition to getting the teachers involved in the project, by drawing on researchers and teachers' complementary qualifications, the objective here was to contribute with joint competence to a feasible group work assignment customized for each year, including the assessment.
The students' intervention was accomplished over the course of two to three hours in their classroom. First, a theoretical part was conducted on how to improve group work and become a well-functioning group; this also touched on a group contract, the evaluation, and assessment. In addition, the students tried working on the group contract and the evaluation, as well as on the self-and peer assessment in the second, applied part of the intervention.

Setup and task
All classes, whether they received an intervention or belonged to the control group, used the same setup to accomplish the group work, and the groups of students were given the same task. The groups were allotted by the teachers composed of four students and mostly consisted of both girls and boys. To be able to study whether-and how-the group work assessment changed over time, the group work consisted of six lessons covering an introduction of the group work assignment, four lessons during which teachers and students accomplished the group work, and an oral presentation where each student group presented their joint product (Figure 1).
The objective for the group work task was that the students should develop their ability to select and use the appropriate mathematical methods for performing calculations and solving the (mathematical) assignment. The core content the students were working on in the task was probability and statistic; they were specifically learning about how to use tables and diagrams to describe the results of an investigation, and how to interpret data displayed in such tables and diagrams. Problem solving was included in the assignment as additional core content, and this involved learning more about strategies for mathematical problem solving in everyday situations. The students were also expected to develop several skills during the work, such as analytical, communication, conceptual, and collaborative abilities.
The task was designed to be suitable for learning and working in a group context. This setup thus enabled positive interdependence, and individual and group accountability.

Group work assessment strategies
The mathematic teachers involved in this project chose to assess the students on the following four abilities: (a) analytical ability-problematizing similarities and differences and consequences; generalizing and drawing conclusions; (b) communication ability-reasoning, discussing, justifying, and arguing for one's own opinion; presentations in both written and oral form; (c) conceptual ability-knowledge of important subject-based concepts, being able to use and relate the concepts to each other; and (d) collaborative ability-active listening, participating, contributing, and taking responsibility for both one's own individual work, as well as the group's common work (Figure 3).  As shown in Figure 3, assessments were carried out at both group and individual level; that is, both the group's common and the students' individual knowledge and abilities were assessed with regard to different parts of the group work. Both the teachers and the students implemented assessment (i.e., who was responsible for the assessments). The types of assessments applied in the project were both formative and summative.

Types of assessment
As can be seen in Figure 3, with regard to assessment strategies, the students submitted empirical evidence for assessment by way of (a) paper and pencil, (b) digital technology, (c) communication, and (d) performance. Assessment strategies also affected the teachers' methods of implementing the assessment which included (a) communication, (b) reading documents, (c) examining tests, and (d) observing abilities, performances, and the group's processes and work. The students used self-and peer assessment (e.g., evaluating their own and peers' cooperation and contribution) to support and supplement the teachers' different assessment strategies.

Ethical considerations
The ethical principles provided by the British Psychological Society guidelines (British Psychological Society, 2014) and the Swedish Research Council (2017), both emphasizing concern for participants' interests, have been applied throughout the study. All participating teachers and students gave their written informed consent to be involved in the study. All students took part in responding to the questionnaires, but the researcher omitted the nine non-participant questionnaires (as described previously), and they were not used further. The study was approved by the regional Research and Ethics Committee at Linköping University, Sweden (Dnr 2013/401-31 & Dnr 2014/134-32).

Results
A core result is that the intervention project had effects on the teachers as well as the students. The findings also highlighted the importance of implementing interventions to both groups of participants (Hammar Chiriac et al., 2019a, 2019bRosander et al., 2020). Thus, participating in the small educational sessions concerning group work assessment had a great effect on both teachers "and students" knowledge (Hammar Chiriac et al., 2019a, 2019bRosander et al., 2020).
By participating in the intervention, the teachers' developed and expanded their mode of languages use and included new modes into their repertoire in terms of terminology, concepts and models concerning group work assessment. In other words, they developed their professional language (Hammar Chiriac & Forslund Frykedal, 2019a), an important springboard for collegial use of professional language. Further, participating in the intervention had a direct influence for the teachers use of languages when providing feedback as means for group work assessment to the students (Forslund Frykedal, & Hammar Chiriac, 2018;Forslund Frykedal et al., 2021). The results displayed the importance of teachers being able to adapt their linguistic repertoire to the students' level when giving feedback. According to the students in Year 5 the teacher "use to many adult words" (Forslund Frykedal & Hammar Chiriac, 2018, p. 194), and the students wished that the teachers used a more understandable language. Additionally, the findings revealed the importance of teachers' using appropriate language when providing written feedback as group work assessment to the students . For example, the result displayed that teachers' ought to convey feedback in distinct manageable units focusing on the task. Hence, the teachers' use of language, spoken and written, and whether they "speak the same language" as the students or not, emerged as an important factor influenced by the interventions.
Regarding the students, they increased their self-and collective efficacy and positive interdependence during the project, thus the intervention gave them a more positive attitude towards group work and group work assessment (Hammar Chiriac et al., 2019b;Rosander et al., 2020). By participating in the intervention, the students' confidence in their own ability (self-efficacy) and the group's ability to collectively work together (collective efficacy) were also strengthened (Hammar Chiriac et al., 2019b). Independent of year, or school, self-efficacy and interdependence predicted collective efficacy in connection with group work assessment .
Based on the findings a theoretical model (Forslund Frykedal & Hammar Chiriac, 2011, indented to present and elucidate current knowledge of group work assessment, has been further developed. The model distinguished between different components of the assessment process; (a) the purpose of the assessment, (b) whether the assessment is implemented at group or individual level, (c) what is going to be assessed, (d) how the assessment will be carried out, and (e) who is going to assess. Thus, the model systematized and elucidated the different parts of the assessment loop integrated with and matched to the task and the working process. Group work assessment is a complex process, and the model was designed to clarify and systematise relevant key components of group work assessment.

Discussion and reflection
In this study the aim has been to contribute increased knowledge concerning intervention studies of group work assessment in education. Inspired by earlier intervention studies of group work, which revealed that a small educational effort developed the cooperative practice (Ashman & Gillies, 1997;Gillies & Ashman, 1996), we implemented similar interventions on group work assessment. The focus of this article is methodological, and its aim was to provide a reflective and critical account of an intervention project on group work assessment, and the implications of the different choices made in this process. By discussing and problematizing the methodological strategies of using educational interventions as a means to enhance teachers and students' ability to conduct group work assessment, we wanted to shed light on an under-investigated field of research. In this section, the following three issues are discussed and problematized (a) collaboration with teachers, (b) methodological choices and consequences, and (c) the effects of the interventions.

Collaboration with teachers
One of the preconditions for the study was to establish collaboration with participating teachers, as their involvement was essential for the implementation of all parts of the study. Acknowledging that collaborating researchers and teachers contributed with complementary knowledge and experiences was a cornerstone for the outcome (c.f. Black et al., 2003). Furthermore, the study was based on the prerequisite of teachers letting the researchers into their classrooms for several weeks while the teachers implemented the group work task. In addition, teachers from four of the six schools also participated in an intervention. The teachers' expectations, interest, motivation, and commitment, as well as their ability to apply theory in practice, were probably among the factors that influenced the extent to which the teachers appropriated the group work tasks as their own. On the other hand, findings from the study show whether (a) it was the teachers themselves or the school who expressed an interest in participating in the study, or (b) whether the teacher participating in the intervention or not, had any visible effect on the teachers' involvement Hammar Chiriac & Forslund Frykedal, 2019a, 2019bRosander et al., 2020). During the second day of the teacher intervention, the importance of the teacher's unique knowledge, experience, and contribution became especially significant when the teachers were given the main responsibility-based on the given framework for the study-of producing an appropriate and group-worthy task usable for both Years 5 and 8, respectively. Although the teachers had to put in a lot of work to get a suitable assignment together, the whole moment was characterized by high teacher commitment and participation. A more challenging task for some of the teachers was appropriating and implementing the group work task in their classroom. Some teachers managed this admirably, but in other classrooms, the researchers, who were in the classroom in observational roles, had to take on a more supportive role, and help the teacher with both structure and implementation. Unfortunately, we have no answer as to why this discrepancy arose, but the researchers' openness to be flexible and sensitive in terms of working methods and role in each classroom became an important means to establish beneficial collaboration with each teacher.

Methodological choices and consequences
In view of group work assessment being a relatively new research area (e.g., Forsell et al., 2020;Forslund Frykedal & Hammar Chiriac, 2011Johnson & Johnson, 2004;Martinez Lirola, 2016a, 2016bRoss & Rolheiser, 2003;van Aalst, 2013), we wanted to explore and increase scientific knowledge in relation to group work assessment. The methodological choices made for this study should be interpreted in the context of previous research which emphasizes challenges and complexities (Brookhart, 2013;Forsell et al., 2020;Johnson & Johnson, 2004;Ross & Rolheiser, 2003;van Aalst, 2013). Given that group work assessment is an under-researched area, we used a mixed-method longitudinal quasi-experimental design where both qualitative and quantitative data were collected, analyzed, and compiled. The choice of approach was based on a desire to capture various perspectives of the participants' actions, and the experiences of group work assessment with different assessment strategies, thereby increasing the likelihood of knowledge contributions to the research area. Thus, the diversity of data collection was expected to contribute with complementary scientific knowledge to the emerging discourse on group work assessment. Even though the approach was appropriate for the underlying idea, it also caused some challenges, such as teachers having difficulties accomplishing all the assessments (Figure 1), and the students (a) becoming tired of the project, and (b) filling out all questionnaires, thus negatively influencing the quality of the findings. In hindsight, we acknowledge the potential risk that the use of too many data collection methods (including too many assessment strategies) posed to the quality of the study with respect to depth and profound knowledge.

Effects of the interventions
In this last part of the discussion and reflection, we problematize which choices concerning the intervention in the form of a short educational session have influenced the outcome and the development of the model presented in the results. Unlike most previous intervention studies in connection with group work and/or CL (Ferguson-Patrick & Jolliffe, 2018;Gillies, 2016;Gillies & Boyle, 2010Martinez Lirola, 2016a, 2016bSlavin, 2014), we wanted to include both teachers and students in the project, and give both groups of participants interventions to determine whether it was possible to increase both teachers and students' knowledge and ability with regard to group work assessment. The choice was based on a desire to obtain comprehensive scientific knowledge regarding the effect of interventions concerning group work assessment per se, but also to determine whether interventions directed toward either teachers or students had diverse effects. The results confirm that the approach of including both teachers and students in the project, and giving both groups of participants interventions with complementary information, broadened the scientific knowledge concerning intervention studies in the emerging discourse on group work assessment. Even though the approach was appropriate, it might also have affected the depth and/or the width of the results negatively, by causing more superficial results, and thus compromising the quality of the findings. There is also the possibility that the effect of the intervention, to some extent, could be due to the participants' awareness of being a part of a project about group work assessment, in other words the participants' awareness of being observed and filmed resulted in them unconsciously modifying their behavior. That only participation in the project is the most likely explanation for the effect is however, disproved by the results from several studies in the project (Hammar Chiriac & Forslund Frykedal, 2019a;Hammar Chiriac et al., 2019b;Rosander et al., 2020).

Limitations
As we intended to study group work assessment, we wanted a group focus. Focusing on "the group" had implications for study units and the random sampling of groups, but also for the analyses Hammar Chiriac et al., 2019b;Rosander et al., 2020). One reflection, based on all the facts at hand, is that it would have been beneficial to be able to track individual answers in, for instance, the analysis of the questionnaires. A careful consideration of the ethical implications of this is recommended, along with its consequences for the project. In our desire to explore and increase scientific knowledge in relation to the emerging discourse of group work assessment, we included too much in the project, both in terms of methods and assessment strategies. A more limited study with well-balanced data collection methods and assessment strategies in relation to the purpose of the study and the desired knowledge contribution could, for that reason, have been more advisable.

Conclusions
In this article, we argue that group work assessment is a neglected research area and contend that there is a need to develop scientific knowledge relating to it. By presenting, discussing, and problematizing methodological perspectives on a group work assessment intervention project, and by examining the implications of the different choices made during this process, we have contributed knowledge to the emerging discourse regarding group work assessment.
Findings from the project have emphasized the importance of collaboration between researchers and teachers, and utilized complementary knowledge and experience from both parts. We found that the teachers' commitment and participation was generally high, and therefore contributed to a successful implementation of the outcome of the intervention project. Important means for beneficial collaboration highlighted in the project was the researchers' ability to be flexible and sensitive in establishing collaboration with participating teachers. Moreover, the findings confirmed that the approach of including both teachers and students in the project, and giving both groups of participants interventions, resulted in contributions of broader scientific knowledge concerning group work assessment intervention studies. Hence, the proposition that interventions have an effect on both teachers and students' knowledge of, and attitude to, group work assessment has been confirmed. Consequently, one salient conclusion from the methodological choices of the project is the importance of clearly defining the content of the project, and not including too much. A future study could focus on the most salient exclusive issues vis-à-vis group work assessment, namely, how to assess individual knowledge.