Potential of Group-Based Negotiation to Promote Learner-Based Reasoning and Weighting of Arguments on Socioscientific Issues

This study analyzes the potential of group-based negotiation processes based on changes in reasoning and argument weighting for a socioscientific issue (SSI) in biology classes. In the corresponding pre- and post-study, students were encouraged to reason and weight arguments about the conservation of local biodiversity before and after a group-based negotiation. For this purpose, the students employed a target-mat structuring tool for compensatory reasoning and weighting in both the individual pre- and post-phases and the group phase. To identify changes in reasoning after group-based negotiation, the use of argumentative resources, i.e., fact-based and normative resources, was assessed. When the students added confirming reasons for arguments, they were more likely to use fact-based resources. When refuting the initial reasoning, the students tended to add normative resources. Furthermore, individual changes in weightings and their relationship to the group weighting were calculated. This analysis revealed that the students changed their weightings toward the group weighting. The results are discussed in terms of the potential of negotiations to cause students to revise and rethink their reasoning and weighting in addressing SSIs and the particular potential of the target-mat instructional tool for structured decision-making.


Introduction
Modern society is increasingly faced with challenges that require decisions based on both scientific and societal deliberations (Sadler, 2009;Zeidler et al., 2005). These socioscientific issues (SSIs) are complex and ill-structured due to multiple facts and norms that complicate decision-making (Kolstø, 2001;Ratcliffe & Grace, 2003;Sakschewski et al., 2014). Their factual complexity is attributed to the multiplicity of scientific and interdisciplinary aspects to be considered and to unknown or unpredictable consequences, such as human population dynamics or ecological disasters (Abd-El-Khalick, 2003;Zohar & Nemet, 2002). Thus, managing factual complexity in decision-making requires the comprehensive consideration of various facts and the assessment of their argumentative relevance (Jasanoff, 2004). To this, factual complexity is added the challenging normative pluralism of multiple interests and values due to different concerned individuals or groups (Bartels et al., 2015;. Furthermore, normative foundations for SSIs are highly controversial and frequently undetermined, partly because of unestablished norm foundations due to the issues' inherent novelty, such as scientific innovations such as CRISPR-Cas (Schleidgen et al., 2020).
The need to manage this factual and normative complexity in negotiation implies a challenging task for the education of future citizens (Ratcliffe & Grace, 2003). Thus, educational approaches are needed that foster decision-making strategies such as justifying arguments (Böttcher et al., 2016;Grace, 2009) or realizing compensatory weighting (Gresch et al., 2017;Seethaler & Linn, 2004). Furthermore, these approaches have to initiate simultaneously a practice of meaningful and situated learning (Sadler, 2009;Samuelsson & Bøyum, 2015). Group-based negotiation offers such a practice of experiencing controversy with dissonance in opinions (Furberg & Ludvigsen, 2008) and collaboratively advanced arguments (Mercier et al., 2017). However, the two goals of offering complexity management and simultaneously conducting open and student-centered negotiation compete with each other in terms of appropriate instructional approaches because managing complexity requires supportive instructional strategies, while student-centered and meaningful negotiation calls for discourse-based and participatory learning environments.
The instructional approach of developing, justifying, and weighting arguments is aimed at supporting strategies without harming the discourse potential (Meisert & Böttcher, 2019) by structuring the process of negotiation and decision-making in phases to decrease complexity and promoting deliberative interactions with a visualizing tool referred to as a target-mat (Meisert, 2018). The potential of the target-mat to activate argumentative resources as an essential basis for negotiation and informed decision-making was shown in a previous study (Jafari & Meisert, 2019). Based on these results, this paper draws attention to the potential of reasoning and weighting within the educational approach of developing, justifying, and weighting arguments supported by the target-mat. For this purpose, students' changes in the weightings of arguments after group negotiation during an intervention about an SSI related to biodiversity conservation are analyzed. These changes are interpreted primarily based on the corresponding changes in the students' reasoning, evaluating whether reasoning and weighting are coherently interconnected within the targetmat-guided decision-making process.

Relevance of Social Negotiation for SSIs
Argumentation is the core activity that enables students to address SSIs (Patronis et al., 1999;Ratcliffe & Grace, 2003;Zeidler et al., 2005). Several educational approaches have been developed and empirically examined to describe and analyze students' argumentation in SSIs (e.g., Anisa et al., 2019;Dawson & Venville, 2010;Erduran et al., 2004;Nielsen, 2012). Argumentation differs in various approaches in that it refers to either an end result in the form of written or oral texts or the dialogic processes of presenting opinions supported by reasoning (Martins & Macagno, 2021). By considering arguments as elements of communicative processes in argumentation, arguments are described as functional instruments constructed by the participants in a particular debate to convince an opponent (Evagorou et al., 2012;Walton, 1990). Approaches that refer to an end result mostly build on structural description of arguments (Andrews, 2005), such as Toulmin's prominent argumentative pattern (Toulmin, 2003). Although corresponding approaches to SSIs in the classroom have been described (Erduran et al., 2004;Evagorou et al., 2012;Jiménez-Aleixandre, 2007), whether the explicit teaching of sound argumentation structures as the intended product of the learning process truly promotes the core competences in an SSI can be justifiably doubted. Building upon the argumentation theory from cognitive science (Mercier & Sperber, 2011) and the learning psychological perspective of Kuhn (2018), argumentation is understood as a social practice that can be fostered "by taking the everyday social practice of argumentation as a starting point and pathway for development of individual argumentative thinking and writing" (Kuhn, 2018, p.123). Kuhn's approach of "extended engagement in dialogic argumentation as a pathway to development of individual argumentative skill" (Kuhn, 2018, p. 121) is anchored in "the sociocultural tradition of Vygotsky" (Kuhn, 2018, p. 123) and consequently uses the potential of learning prerequisites and the possibilities to foster skills by experiencing real-life learning contexts. The critique that structure specification of arguments does not properly account for typical conversational dynamics is shared by various other authors (e.g., Andrews, 2005;Böttcher & Meisert, 2013;Nielsen, 2012).
The corresponding science education discourse on the promotion of students' argumentation skills therefore features two central distinctions (formal vs. nonformal; individual vs. social), which can be characterized as follows with regard to the approach favored here: First, argumentation (and reasoning) is aligned with its nonformal expression to ensure its suitability for everyday use and with its connectivity to strategies based on prior knowledge (Sadler, 2004). Second, argumentation is mainly understood as a social practice rather than an individual activity (Kuhn, 2018), as individuals are required to provide reasons with in-depth justifications due to controversies between participants. Accordingly, Mercier and Sperber (2011) explain that the main function of reasoning is argumentative in its inherent cooperative character (Mercier & Sperber, 2011).
The correspondingly assumed link between argumentation and social interaction is based on the controversial potential of many individual positions, which in turn demands negotiation (Mercier & Sperber, 2011). In such negotiations, individual participants are invited to present perspectives, exchange ideas, and confirm or refute arguments to challenge each other, including both factual aspects and normative aspects (Furberg & Ludvigsen, 2008;Jiménez-Aleixandre, 2007). Thus, despite an apparent controversial and competitive interaction, argumentation is cooperative at its core, as each individual is responsible for a success with the contributions that he or she makes (Mercier & Sperber, 2011). Hence, argumentation offers participants the opportunity to evolve themselves by rethinking their reasoning, as it is tested for validity and viability through cooperative competition by feedback and comparison. This offers opportunities to strengthen one's own position or to get to know critical perspectives on it and, if necessary, to adopt new points of view. However, any change or revision of one's own initial reasoning is described as particularly difficult (Mercier, 2016). Regardless of intelligence or the degree of openmindedness, people tend to confirm or even reinforce their own positions (Kunda, 1990;Nickerson, 1998). This so-called confirmation bias is a particular hurdle for participants in dealing with SSIs, as it narrows their argumentation and hides other perspectives. As socially situated negotiation requires dealing with other arguments, a broader awareness of other or new points of view and incorporation of them into one's own argumentation can be initiated (Mercier & Sperber, 2011) with potential to overcome confirmation bias and other cognitive limitations (Mercier, 2016). Negotiation as a group process is therefore a means to not only improve the situational quality of argumentation but also enhance the individual skills of each participant (Kuhn et al., 2016).
This potential of group-based negotiation is confirmed in the context of SSIs in biology classes by Jafari and Meisert (2019), who analyze the initiation and revision of normative and fact-based argumentative resources as an important factor in the development of argumentation competences (Hostenbach et al., 2011). The study demonstrates a distinction between the explicit mentioning of norms (such as biodiversity protection) and the naming of individual preferences (a wish not to expel animals), a decrease in preferences and an increase in explicit norms (often presented as moral principles), and fact-based resources after negotiation. Given the broadly problematized lack of both facts and norms in SSI argumentation and decision-making (Abd-El-Khalick, 2003;Kolstø, 2001), these changings demonstrate the negotiations' potential, as "students themselves can recognize the relevance of explicit moral principles, if the setting offers opportunities to support studentdirected discourse" (Jafari & Meisert, 2019, p. 17).
This potential becomes clearer when a person clarifies that norms are conventions or rules that are shared in a particular social or cultural group and that define and thus organize what is acceptable in a society (Bartels et al., 2015;Bicchieri et al., 2018). Correspondingly, norms are the result of long-term social negotiation processes about what is undesirable (Deitelhoff & Zimmermann, 2020), while preferences express only individual desires that are not necessarily shared by others (Miller, 1992). Group negotiations to identify consensual goals or norms on the basis of individual preferences are one of the core challenges in addressing socially controversial SSIs to establish a democratically stabilizing consensus in the face of societal pluralism (Deitelhoff & Zimmermann, 2020;Krook & True, 2012;Ratner, 2004). Pedagogical approaches therefore have a particular responsibility to practice these deliberative negotiations because students have difficulty engaging in discourses without practice and support (Kuhn, 2015). This responsibility presents a particular requirement for teachers to "manage such activities within an appropriately secure learning environment, […] enabling pupils to think seriously about how they might respond if they were directly involved" (Grace & Byrne, 2010, p. 79).
The classroom environment provides a promising platform for practicing and thus enhancing negotiation skills through exposure to diverse opinions (Samuelsson & Bøyum, 2015), as "argumentation requires both a supportive community and the strengthening of individual skills and understanding" (Kuhn, 2018, p.8).
Differing controversial opinions within a discourse have the potential to test the viability of justifications (Goodwin, 2001). Here, it is not so much externally prescribed structural specifications that have a role but rather the learners' perspectives as "naturally born arguers" and an appropriate and situated learning environment in which students' own arguments actually compete within a meaningful controversial discussion with each other to question their own perspectives and evolve their individual abilities (Kuhn, 2015;Mercier et al., 2017).

Compensatory Weighting as the Preferred Strategy for Decisions in SSIs
Compensatory strategies (or trade-off strategies) are decision-making strategies in which advantages and disadvantages concerning different criteria are weighted, compared, and related (Jungermann et al., 2010). In contrast, noncompensatory strategies (or cut-off strategies) describe decision-making strategies in which unacceptable options are simply excluded as only one criterion is considered (Gresch et al., 2017;Jungermann et al., 2010). Thus, managing the multitude of arguments in SSI decision-making requires demanding compensatory weighting (Fang et al., 2019;Seethaler & Linn, 2004). Especially in scenarios of sustainability issues, compensatory weighting has been described as appropriate (Gresch et al., 2017). Accordingly, it is assumed that the ability to make such informed, appropriate, and compensatory decisions is often also closely linked to the ability to engage in reasoning (Fang et al., 2019). Because compensatory weighting is cognitively challenging, decision makers tend to avoid these strategies and focus on cut-off strategies (Betsch & Haberstroh, 2014;Jungermann et al., 2010).
Thus, instructional approaches need to provide strategies that help students manage this complexity and facilitate weighting activities that are included comprehensibly in decisionmaking (Grace, 2009). Since students might reject decisions if they contradict their intuitive and original choices and they do not understand the strategy instructed, transparency is one of the main criteria for developing instructional approaches (Böttcher & Meisert, 2013;Fang et al., 2019).
However, it is important to consider that providing instruction on how to apply compensatory weighting strategies alone is not sufficient to achieve trade-offs. It has also been shown that students consciously avoid the cognitive effort of compensatory strategies when making decisions on SSIs (Gresch et al., 2017). Narrow instruction of complex but appropriate decision strategies can decrease the willingness to use weighting activities, for example.

The Target-Mat as a Tool for Coping with Complexity
Ensuring student-centered and collaborative discourse and nonformal but deep reasoning, including compensatory decision-making, creates high demands for educational settings. Students need support for different challenges simultaneously; therefore, corresponding approaches have to support complex mental processing integrated in demanding controversial social interactions without reducing the characteristic complexity of the issues (Gresch et al., 2017) and without threatening the self-determination of argumentation (Grace, 2009). Thus, students must be encouraged to cope with a diversity of multidimensional information (Abd-El-Khalick, 2003) and develop effective reasoning against a background of controversial interactions (Kuhn, 2015) focusing on identifiable facts and norms as well as on intuitive preferences (Fang et al., 2019). A central interest in fostering with tools is therefore individual participation with the parallel support of group processes as shaping human action (Evagorou et al., 2012;Vygotsky, 1978;Wertsch, 1985).
To this end, there is already a wide range of tools that focus on various educational outcomes. Different examples of scaffolds target different aspects of the argumentation or decision process. Ratcliffe's argumentative framework (Ratcliffe, 1997) attempts to provide support for students' discussion by providing task-driven steps to focus on, since a well-managed discussion leads students to make good or elaborate decisions (Grace, 2009). Other supportive learning tools use concept maps as pre-established scaffolds whose facilitation of complex knowledge has been examined in different contexts (e.g., Brandstädter et al. 2012;Kalyuga et al., 2003). Such scaffolds are used here for knowledge structuring as a supportive basis for reasoning and decision-making in SSIs (Eggert et al., 2017). Other supportive tools offer scaffolds to support argumentation (Belland et al., 2015), communication (Lin et al., 2012), or evaluating the credibility of evidence in SSI-based questions (Nicolaidou et al., 2011). These few examples illustrate well that the various scaffolds often address single aspects, while neglecting other important steps, such as knowledge building or making reasoned decisions, and do not always make the process of decision-making sufficiently transparent (Böttcher & Meisert, 2013;Ke et al., 2020).
Against this background, the approach of developing, justifying, and weighting arguments (Böttcher et al., 2016) aims to achieve these aforementioned goals through the steps of analyzing the issue, developing arguments, and justifying and weighting them. In this approach, a group-based negotiation process is used to decide on the respective weighting of the previously developed arguments. This group-based negotiation as the core process of controversial and deep reasoning is preceded by individual written reasoning on every argument. This individual reasoning for each argument is shared via a worksheet, the socalled target-mat ( Fig. 1) (Meisert, 2018), which visually supports the connection between oral and written reasoning as well as argumentation and weighting (Meisert & Böttcher, 2019). The preparatory step of individual writing has been shown to be effective because "representing ideas in concrete, visible form enhances students' ability to reflect on and work with them" (Kuhn, 2018, p. 3). In addition, the target-mat visually separates the individual arguments from each other as core elements relevant to the decision and offers a close and thus orientation-facilitating arrangement of both single-argument reasoning and weighting (in relation to multimedia learning in the form of the segmenting principle and the principle of spatial contiguity (Mayer, 2020), as well as a clearly arranged and grouping of pro and con arguments. On this basis, each reasoned argument is negotiated regarding its weight for the decision visualized by marks along a scale on the target-mat. Building upon these discussed weightings, a final group decision is made (Böttcher et al., 2016). A previous pre-and post-analysis provides initial insights into the potential of the target-mat and shows the use of students' argumentative resources (Jafari & Meisert, 2019). However, there are key aspects that still await analysis. First, the way in which reasoning is changed needs to be considered. Second, the functionality of the weighting step within the approach with regard to the target-mat needs to be analyzed. Although compensatory weighting can be considered unquestionably relevant in its own right, this study aims to determine its potential for group-based negotiation in terms of its coherent link to reasoning.

Research Focus
Based on the aforementioned research interests, the goal of this study is to analyze the potential of group-based negotiation to initiate revised reasoning and weighting in SSIrelated decision-making. An intervention is conducted on an exemplary SSI to analyze changes in students' individual reasoning and weighting as well as their relationship to the weightings negotiated in the group. These analyses are discussed in terms of the functionality of the target-mat to support students in coping with decision complexity and to initiate socially situated and thus meaningful discourse. The analysis aims to tackle the following concretized research questions: (RQ1) What types of changes occur in students' individual reasoning after participating in a group-based negotiation process? (RQ2) What changes do students make to their individual weightings after group-based negotiation, and what is the relationship between the group process and individual revision of weightings? (RQ3) What kinds of connections exist between changed reasoning and changed weightings?

Intervention
The intervention was conducted according to the approach of developing, justifying, and weighting arguments (Böttcher et al., 2016;Meisert, 2018) in an integrated middle school in Lower Saxony in Germany with 146 eighth-grade students from six different classes. The intervention lasted a total of two school hours in each class, corresponding to a total length of 180 min in each class (Fig. 2). The students were in class 8 (i.e., 13 or 14 years old) and had experience in basic ecology in science classes. The students in this study were described by the teachers as good performing. They were familiar with assignment work in groups but were less familiar with negotiation in science lessons. Moreover, they were not familiar with the target-mat for structuring argumentative negotiation but were certainly familiar with its basic components, such as a target and placemat. Finally, the topic of decision-making with respect to ecological problems had not been taught in any classes during the current school year.
According to the approach of developing, justifying, and weighting arguments, the students became familiar with the case and received background information on a fictional SSI about the construction of an amusement park in a nearby forest (step 1) during the first 90 min of the intervention. For this purpose, the students were presented with different controversial opinions about the project by different stakeholders. In a second step (step 2), the students acquired information from worksheets about particular subjects: (a) red list endangered species (or saving the wildcat, an endangered species that lives in this forest), (b) noise pollution and its effects for animals and people, and (c) fragmentation and its consequences, particularly for a population of animals in their habitat (see the translated worksheets in the appendix). The corresponding task invited the students to explain the content of each worksheet to each other within a limited time frame to enable them to understand the relevant factual issues. The students were involved in solving the task so that they could not have any further discussions, for example, about the assessment of the case. Next (step 3), the students developed arguments based on the given information and anticipated prior knowledge (refer to Table 1). Although the arguments were developed by the students, they were predetermined in the research group for all learning groups before the study was conducted, and the materials were developed accordingly to ensure the comparability of the data. If the students thought of further arguments, they could name them.
These first argumentative ideas served as a heuristic anchor of aspects to be considered (referred to in the following text as "arguments" regardless of their diverse and underdetermined character).
The intervention materials were tested in two pilot phases regarding their feasibility for the students. First, the comprehensibility of the worksheets and tasks as well as their processing effort were assessed through process observation and analysis of the learning products. After the textual information was shortened and the tasks were partially reformulated, the intervention materials were tested on another group and validated as suitable in terms of accessibility and time required.
Steps (4) to (6) were performed with the target-mat (Fig. 1). The students (step 4) reasoned and weighted each argument individually (pre-phase). The weighting of arguments was performed by placing a mark in the central part of the target-mat. The further the weighting mark was placed inside the target-mat, the more important the argument was rated. After this step, the students formed groups (step 5) to transfer their reasonings to a group target-mat (treatment). By exchanging different reasonings, the students were supposed to reach a consensus on the group weightings and reasoning for each argument. As a final step (6), the students reweighted and reasoned voluntarily if they felt that something had changed individually on their own target-mat (post-phase).

Data Collection and Analysis
During the intervention, both individual target-mats and group-based target-mats were completed and collected for analyses (Fig. 3). This step resulted in two types of data presented here: (a) individually written reasonings in the pre-and post-phases and (b) individual and group-based weightings on a four-tier rating scale. A total of 146 students could weight and reason eight arguments in the pre-and post-phase.
Hence, a total of 1168 arguments could possibly be weighted and reasoned individually in the pre-phase and modified in the post-phase. From these two types of data collected, three forms of data analysis related to the research questions were conducted: 1) Analysis of the comparison of the type and direction of changes in individual reasoning from the pre-to post-phase (related to RQ1). 2) Analysis of the individual weightings with comparison to the group weightings (related to RQ2) 3) Analysis of the relationship between the changed reasonings with and without a change in weighting (related to RQ3).

Comparing Individual Reasonings from the Pre-to Post-Phase
Focusing on RQ1, the before-and-after comparison of the individual written reasonings using MAXQDA 2020 (VERBI Software, 2019) served to inductively identify changes. These changes in reasoning were related to individual reasoning in the student's target-mat. For example, a change in reasoning for the argument Wildcat is shown in Fig. 3 (Example 1), whereas the reasoning for the argument Tourism is not changed. The number of reasonings The amusement park would make the city more attractive for potential tourists Infrastructure Because the park would be located in the forest, the infrastructure (buses, trains, etc.) would need to be improved Con arguments Explanations Wildcat The wildcat is a threatened species that lives in this forest, and it would be more endangered by the amusement park Traffic Because of visitors, traffic volumes would increase Fragmentation If the forest were divided by the amusement park and by roads, animals could not pass and enter other parts of the forest Noise pollution There would be a high level of noise during construction work and while the business is running that were changed on each argument was determined. The quantitative occurrence of change was tested statistically against the null hypothesis to determine the extent to which this number of changes differed from the expected change (which is no change in reasoning). The corresponding binominal test was carried out with SPSS 27 software (IBM Corp. Released, 2020). Then, through the analysis of 1026 completed pre-and post-phase reasonings, different types of changes, such as reinforcement and confirmation (= stabilizing types of change) as well as weakening and refutation (= revisional types of change), could be distinguished (see detailed category description and examples in Table 2). A total of 142 reasonings with post-phase but no pre-phase reasonings were categorized as new argumentation. Additionally, the changed elements of reasoning were assigned to the categories of fact-based resources or normative resources according to Jafari and Meisert (2019). The applicability of the coding system was assessed in several evaluations by the research group. After all data were coded, a second coder analyzed a part of 10% of all data using the coding agenda after an intensive introduction by the first coder. Intercoder reliability was calculated using Cohen's kappa coefficient and was 0.73. This is considered substantial according to Fleiss et al. (1981).

Comparing the Individual Weightings to the Results of Group Weighting
Focusing on RQ2, the students' weightings of arguments with the four-tier rating scale of the target-mat were used to record weighting changes from the pre-to post-phase.  The boxes were assigned values so that the innermost box of the target-mat (very important) was given a value of 4 and the outer box (unimportant) was given a value of 1. The higher the score, the higher the respective argument was rated. A change in weighting can be seen in Example 2 as shown in Fig. 3. Here, the student changes the weighting of the argument Infrastructure from very unimportant (outermost box) to rather unimportant (second outermost box). These changes were further analyzed in relation to the respective group weighting (Example 2, with a group weighting of rather important (second innermost box); the student changed his/her weighting in this direction). To calculate the correlation between the change from the pre-to group-phase weighting and the change from the pre-to post-phase weighting, the correlation was analyzed according to Spearman's rank correlation due to its lack of a normal standard distribution. These results were interpreted using Cohen's effect size (Cohen, 1992). These quantitative data analyses were conducted using SPSS 27 (IBM Corp. Released, 2020).

Analyses of the Reasonings and Weightings of the Students who Changed Both
Focusing on RQ3, whether the type of reasoning change is consistent with the change in weighting related to the same argument was analyzed. The selected cases with changed reasonings were divided into two selected groups: those with changed weightings (such as Example 3 in Fig. 3) and those with no change in weightings (such as Example 4 in Fig. 3). Building upon this deviation, the occurrence of types of reasoning change was comparatively analyzed in both groups.

Reasoning Results (RQ1)
From the 1026 reasonings for all arguments of all individual students in the pre-phase, 345 reasonings were changed in the post-phase (i.e., approximately 33.7%), with p = 0.000 for every argument (Table 3). This is a highly significant change in reasoning against the backdrop of an assumed null hypothesis referring to the common 0.05 level of the binomial test (Georgii, 2009). Among the total of 142 arguments without any reasoning in the pre-phase, 94 of them had reasonings in the post-phase (i.e., approximately 66.2%). Relative to an assumed null hypothesis (see above), students' reasonings changed significantly for each argument (Table 4).
In the analysis of the frequency of change types in reasoning (Fig. 4), refutation was the least frequent (12.5%) change type, weakening was the most frequent (36.5%) change type, and confirmation was the second most frequent, with 30.4%. The category reinforcement was represented by 20.4%. Comparing the sum of refutation and weakening (= revision type) as well as confirmation and reinforcement (= stabilization type), both type pairs were almost equally represented. The category new argumentation occurred for approximately 66% of the blank reasonings in the pre-phase (Fig. 5).
With regard to the occurrence of argumentative resources in changed reasoning, the following distribution can be observed. Considering the category new argumentation, factbased resources occurred in approximately 60% of the arguments, and normative resources  occurred in approximately 40% of the arguments. In reinforcement and weakening, almost a similar number of normative and fact-based resources were identified. In cases of confirmation, 72% of the resources were fact-based resources, and 28% of the resources were normative resources. When considering refutation, normative resources (approximately 63%) rather than fact-based resources (37%) were identified. Normative and fact-based resources mostly occurred equally for weakening and reinforcement. A higher occurrence of fact-based resources was observed for confirmation and new argumentation, while normative resources were observed for refutation.

Weighting Results (RQ2)
A total of 1137 of all possible arguments were weighted in the pre-phase, and 1145 were weighted in the post-phase. Furthermore, 31 student groups completed a total of 248 group weightings. The number of weightings the students changed within their own target-mats was determined. Changes to one, two, and three arguments occurred to a similar extent (approximately 15% of the students for each). Very few students changed six, seven, or eight arguments (1 to 2% each). Only one-third of the students did not change any of the eight possible weightings. Focusing on all arguments, 310 changes (27%) from pre-to post-weighting could be analyzed. These changes were evenly distributed, varying between 8.9 and 33.1% (Table 5).
The median of the rating of the weightings of each argument ranged between 2 and 3, and the only median that changed from the pre-phase to post-phase was for the Traffic argument (Table 6). However, all possible weightings (from 1 to 4) occurred in all arguments in both the pre-and post-phase. When comparing the mode, it often corresponded to the median, except for the contra arguments Wildcat, Traffic, and Fragmentation.
When comparing the change in individual weighting from pre-to the group-phase with the difference between individual pre-and post-phase weighting, similar tendencies were obvious (Fig. 6). Calculating a corresponding Spearman correlation, a strong significant effect of the difference between change from individual to and group weightings and the individual weighting change (pre to post) was observed for almost all arguments according to Cohen (1992), except for the Traffic and Fragmentation arguments.

Weighting and Reasoning Results (RQ3)
A total of 236 arguments showed both a change in reasoning and a change in weighting, while 182 arguments were identified in which changed reasoning was combined with unchanged weightings. In both groups, all possible types of changes in reasoning occurred but with different frequencies (Fig. 7). Within the change-both group, weakening was most frequent (approximately 38%), while confirmation (30%) dominated within the change-reasoning group. A comparison of the sum of weakening and refutation (= revision type) with the sum of reinforcement and confirmation (= stabilize type) revealed that the stabilization type was higher in the change-reasoning-group, while the revision type occurred mostly in the change-both group. When examining the category new argumentation, the weightings were mostly unchanged.

Discussion
Understanding argumentation as a social practice demands the fostering of students' argumentation in a socially situated learning context in school (Kuhn, 2015;Meisert & Böttcher, 2019;Mercier et al., 2017). Using argument weighting and group negotiation to initiate deeper negotiation and reasoning, the target-mat aims at promoting this social practice by visualizing argument-related reasoning and weighting and structuring the task steps, such as written reasoning and negotiated weighting. Using a pre-post comparison of students' individual reasoning and weighting before and after the corresponding group negotiation reveals two important discourse aspects from the analyses of the results: (1) A broad rethinking of students' initial assessment showed a significant influence of group negotiation on the tendency of this individual to rethink (see "Potential of Structurally Supported Group-Based Negotiation to Expand and Revise Students' Argumentation"). (2) The instrument can be used as a scaffold to support these negotiation processes (see "Potential of the Target-Mat for Compensatory Weighting").  Pro  More jobs  3  3  3  3   Leisure activity  2  2  2  2  Tourism  2  2  2  2  Infrastructure  2  2  2  2  Contra  Wildcat  3  3  4  3  Traffic  2  3  1  3  Fragmentation  3  3  4  3  Noise  3  3 3 3

Fig. 6
Change in weighting difference from pre-weighting to group weighting and pre-to post-weighting

Potential of Structurally Supported Group-Based Negotiation to Expand and Revise Students' Argumentation
One remarkable finding in this study is the large amount of modified (approximately 34% of all reasonings that had a reasoning in the pre-phase) or new reasoning (approximately 66% of all reasonings that had no reasoning in the pre-phase at all) after group-based negotiation, even though any change in this phase was explicitly voluntary. Both the extent of reasoning changes and their broad distribution across arguments demonstrate the potential of the intervention to initiate rethinking and partial revision of the initial reasoning. Described as cognitively and motivationally demanding (Mercier, 2016), this broad reconsideration might be due to the ability of the target-mat to both (1) partition and visually structure the reasoning content (Kuhn, 2018), which provides cognitive relief by lowering the hurdle to mentally processing and rethinking the whole discussion, and (2) enable participative group negotiation, which offers additional aspects and perspectives (Mercier et al., 2017). Based on these findings, this approach of structurally supported negotiation promises to be a student-centered way to overcome the highlighted challenge of students having such discussions independently, even without support or previous training (Kuhn, 2015). In more detail, the pre-and post-comparison shows a broad spectrum of change types ranging from stabilization and revisions to new thinking, i.e., new argumentation. A revision example of the argument More jobs shows that the student obviously questioned the quality of jobs in the pre-reasoning ("Those aren't the best jobs") but qualified this by mentioning the general advantage of being unemployed ("These aren't the best jobs, but jobs are always good, so the unemployed at least have something to do."). The same argument is reinforced by another student, who also questions the quality of the jobs in Fig. 7 Distribution of categories of changes in reasoning in the change-both group and in the change-reasoning group this amusement park in the pre-phase reasoning ("The jobs are bad") and concretizes this opinion in the post-phase reasoning regarding the pay ("The jobs there won't help the unemployed because they're probably going to be low-paying work."). First, these examples demonstrate the neutrality of the target-mat by facilitating different directions of change, which initiates rethinking without prescribing its direction. This is considered an important feature with regard to the nonmanipulative or nontendentious character of scaffolds such as the target-mat (Meisert & Böttcher, 2019). Second, the high occurrence of revisional changes in more than one-third of all changes is particularly remarkable. It is broadly problematized that individuals have difficulties revising their own opinions or prior knowledge and that they tend to filter information to fit and sometimes even reinforce their existing attitudes or values (Fischer et al., 2005;Jonas et al., 2001). Although the students' revised reasoning in this pre-/post-comparison cannot be clearly attributed to the influence of negotiation, the literature supports the claim that negotiation in general (Mercier et al., 2017) and particularly combined with a meaningful learning setting (Kuhn et al., 2016) helps to overcome the hurdles of limited argumentation skills. Accordingly, this revision activity can be interpreted as an indicator of an intensive or meaningful negotiation process that enables students to "make the best of […] [their] reasoning abilities" (Mercier et al., 2017, p. 1). In this context, two central limitations of the analysis presented here must be discussed. Firstly, the research findings are based upon a pre-post comparison; thus, further evidence about influential factors has to be produced with an experimental-control design. It is important to consider that any control design will vary in multiple factors from the complex setting realized here, with resulting problems in identifying single relevant factors. Secondly, a further content analysis of the students' reasoning (comparing preand post-reasoning) could offer deeper insight into the process quality. Although there are sophisticated tools related to their structural completeness (Toulmin, 2003) or their degree of justification and functionality (Kuhn et al., 1997), the keyword character of the students' written reasoning does not allow the application of the complex analysis tools. This limitation is mainly due to the intervention that aims to apply only the kind of task typical for the target-mat-supported process and no further research tools with side effects by initiating additional considerations. Despite these limitations, selected examples of the students' written reasoning provide exemplary insight into the content quality of reasoning changes. The following reasoning shows an example in which the student elaborates his or her reasoning after the group-based negotiation with two functional and justifying facts (habitat loss and fragmentation), while his or her initial reasoning refers to a consequence in an undifferentiated and thus unjustified way (animal suffers): The pre-reasoning: "Important, as otherwise many animals will suffer." The post-reasoning: "Damages habitats of animals and fragments the forest." The reasoning of another student regarding the argument Wildcat shows the addition of a normative resource after negotiation. While the pre-reasoning is still unobjective and conceivably dysfunctional, the revised reasoning is well reasoned with a functional and justifying norm reference (wildcats' protected status and threat of extinction), which strengthens the reasoning: The pre-reasoning: "I think cats are stupid." The post-reasoning: "We must protect animals, and the wildcat must not become extinct." In addition to the abovementioned limitations to applying quality standards on the students' written reasoning, the audiotaped negotiation will be analyzed in a further step of the research project. Additionally, this further analysis of discourse quality corresponds more narrowly with the educational approach presented here, which is aimed at meaningful communicative interaction (Mercier et al., 2017).
Further analysis of the argumentative resources accompanying the reasoning changes offers deeper insights into the type of argumentative changes: Students are more likely to use fact-based resources to confirm their initial reasoning, while normative resources dominate in refutation. An example of confirmation by fact-based resources (in this case, by the added comment on leisure possibilities) can be found in the argument Leisure activities: pre-phase reasoning: "Don't need it." post-phase reasoning: "There are enough leisure activities, like swimming pools. I don't need an amusement park." An example of a refutation is given in another example, in which a student starts in the pre-phase with a general relevance rejection of the argument Noise by referring to the student's own lifestyle habit, while the student switches to the counter-position with reference to the normative issue of health, which furthermore entails a remarkable expansion of the entities under consideration (from "I" to "people and animals").
pre-phase reasoning: "I don't go into the woods." post-phase reasoning: "The forest must be protected because it is important for the health of humans and animals." Thus, the previous example shows another potential of group-based negotiation in the form of taking into account other people (or animals) concerned as a broadening of perspectives, which is not systematically analyzed here but is discussed as a core objective in dealing with SSIs (Herman et al., 2020).
Different interpretive assumptions can be used in assessing the use of facts in confirmations, as in the previous example, and the use of norms in changes, as in the latter example: These results have the potential to problematize the role of facts in SSI decision-making, which tends to be used to stabilize previous reasoning (Nielsen, 2012). From another point of view, however, it seems reasonable that facts are employed primarily for stabilization. Preferences or explicit norms are directly connected with the preferred world to come that is reasoned for, while factual aspects mostly function as a frame that clarifies its feasibility or probable consequences (Bartels et al., 2015). It is therefore plausible that revising changes mostly requires changes in the normative resources of the argument. This interpretation reconstructs the process of assessment change as a result of perceived fact-based or normative resources. However, this process can also be interpreted in the opposite way. Attitudes and beliefs, which form the intuitive basis of normative orientations (see Bicchieri et al., 2018), strongly influence the perception of information. Ajzen et al. (2011) show that information only has an impact on behavior if it is coherent with attitudes and beliefs. Thus, the low frequency of changes associated with an added factual aspect could additionally be due to psychological filters that prevent the perception of facts that do not fit the existing assessment. These processes during negotiation that potentially initiate revisional considerations will be analyzed on the basis of the discussion transcript in a further research step. Irrespective of this consideration, the very important role of normative aspects and their negotiation in SSIs is confirmed due to their correlative affiliation with revising considerations.
Despite this limited importance of facts for revisions, fact-based resources are often perceived in the process. The data show that confirmation occurs repeatedly and primarily through fact-based resources. These corresponding changes or supplements may mostly be due to factual misunderstandings or incompleteness. This result is consistent with the findings of Simon and Amos (2011), who show that students in SSI discussions could not consistently and accurately state scientific content. This is confirmed in particular by the finding that the argument Fragmentation has the highest lack of reasons in the pre-phase and at the same time the highest complementation frequency in the post-phase (73%). This illustrates that complex scientific facts, even with potentially high relevance, are at risk of being ignored or not fully considered. Structured and visually supported group-based negotiation thus has the potential to make those aspects more accessible.
Conversely, fact-based resources with high relevance already have a good chance of being identified in the pre-phase in the setting. This finding supports the findings of Nielsen (2012), who investigated the use of scientific facts in group discussion on SSIs and discovered that they are mentioned partly to provide argumentative support for previous argumentative tracks. In both cases, the results here, supported by other findings (Lewis & Leach, 2006;Nielsen, 2012), show the importance of factual grounding to cope with the process of discourse on SSIs.

Potential of the Target-Mat for Compensatory Weighting
The structure of the target-mat, in addition to initiating and supporting group-based negotiation, is aimed at combining nonformal reasoning with corresponding weightings to promote compensatory weighting strategies as an educational goal in addressing SSIs (Seethaler & Linn, 2004). The structure of the target-mat aims at compensatory weighting, as argument-related weightings must be accomplished by marks in a four-tier scale. Crucial operations, such as the analyses of strengths of every argument built upon its reasoning and its group-based negotiation, are thus carried out by the target-mat. In addition, the structure further allows students to have an overview of the weighting decisions to better compare them and to realize trade-offs for the final decision (Böttcher & Meisert, 2013). Although the influence of the different setting elements (argument-related reasoning and weighting, group-based negotiation, alternation between written and oral communication, etc.; see previous remarks on limitations) cannot be determined with this pre-post comparison, the multifaceted potential of the entire setting, including the target-mat, is evident in the data.
The application of compensatory weighting as a central goal of fostering decision-making strategies in SSI-related topics requires that weighting does not take place as an additional task but is meaningfully linked with corresponding reasoning (Fang et al., 2019). An initial hint of this meaningful embeddedness is the high frequency of weighting changes (a quarter of all weightings) after the group negotiation across all arguments. This high occurrence demonstrates the students' awareness of the weighting relevance -particularly given the voluntary nature of all pre-and post-phase changes. This is particularly noteworthy because compensatory weighting is generally considered to be challenging or even demanding (Gresch et al., 2017;Kool et al., 2010), although students are able to apply compensatory decision-making with minimal structuring support (Seethaler & Linn, 2004). Furthermore, the results demonstrate a meaningful linkage between reasoning and weighting: The students tended to change their weightings according to the group weighting, demonstrating a correspondence between group negotiation and post-phase changes. This correspondence can be confirmed by the combined analysis of changed reasonings with and without changed weightings. While a stabilization of reasoning is mostly connected with the same weight, a revisional type of change in reasoning tends to require a changed weighting. This coherence of changes in reasoning and weighting conforms to the meaningful linkage between reasoning and weighting through the target-mat-structured decision process.
This coherence could be interpreted in different ways. First, these changes could be substantially initiated by the group negotiation, convincing some students to change their initial assessments concerning both reasonings and weightings. This interpretation is elementarily based on deliberative processes predicted by argumentation theory (Mercier et al., 2017). Second, the change in weighting in accordance with the group weighting could be interpreted simply by the influence of social norms through which students follow the group voting without being convinced of it (Schultz et al., 2007). Even if this interpretation is not the essential aim of the process, it would at least be a sign that the negotiation process has led students to become aware of the group vote. This finding can also be classified as a basal level of goals that perhaps serves as a precondition for future negotiations with more active participation.
Based on the thesis that students need support in decision-making (Arvai et al., 2004;Bohnenblust & Slovic, 1998;Grace, 2009), the findings offer evidence that the approach of developing, justifying, and weighting arguments combined with the target-mat is suitable for fulfilling the demand for both student-centered social discourse and structuring aids for appropriate decision-making strategies to manage factual and normative complexity in SSIs. Following the assumption that "deliberative skills, knowledge, and values are learned through practice" (Samuelsson & Bøyum, 2015, p. 77), further analysis of the detailed interaction during the negotiation processes is required to identify which type of discursive interaction initiates deeper reasoning and corresponding changes. Accordingly, the results and patterns discussed here will serve as a basis for discourse analysis in search of suitable methods (Hennessy et al., 2020) to reveal the degree of dialogical interaction in relation to the achieved depth of reasoning supported by the target-mat.