Reasoning in chemistry teacher education

1 Introduction

Reasoning in science is characterized by cognitive processes that can contribute to, or allow for, the formulation and testing of hypotheses, two important aspects of scientific endeavours (Giere et al., 2006). These cognitive processes, among others, involve formal logic, including probabilistic logic, and non-formal processes, including model-based reasoning (Khan, 2007) and analogical reasoning (Bruce et al., 2022). Several of the more non-formal cognitive processes have been associated with scientific activity at the revolutionary end of the continuum on advancement in science (Kuhn, 2012; Simon, 1996).

While science involves both formal and non-formal reasoning and problem-solving, the terms have been used interchangeably. While solving a problem, a problem-solver might engage in both forms of reasoning. As Zimmerman (2000) points out, one of the main generalizations about problem-solving is the use of heuristic searches. A problem solver constructs or develops a representation of the problem situation (i.e., the problem space) in order to generate a solution within some set of constraints. A problem space includes the initial state, a set of operators that allow movement from one state to another and a goal state or solution. It is proposed that goal-oriented behavior in general involves a search through problem spaces. In the sciences, these spaces have been thought to include a hypothesis space, experiment space, and model spaces, among other types of problem spaces (Schunn & Klahr, 2019). Reasoning, on the other hand, appears to distinguish itself from problem-solving and its use of heuristic searches in problem spaces in that direct retrieval of a solution from memory is not possible for reasoning (Zimmerman, 2000).

Magnani (2009) contends that much of scientific inquiry involves reasoning where models are generated and used in experimental and theoretical work, rather than problem-solving using heuristics. Even contrary to a more standard view of scientific reasoning as hypothetico-deductive or logic-based, abductive or model-based reasoning is arguably more descriptive of research in the sciences (Nersessian, 2008) and in chemistry (Wackerly, 2021). Modeling as an epistemic practice can yield explanatory mechanisms and manifest predictions beyond the scope of the original target domain or phenomenon (Rost & Knuuttila, 2022). Thus, models are useful for solving problems which cannot be solved by heuristics alone. Lehrer & Schauble (2005) further contend that other practices associated with science (such as argumentation) develop more as a contest about the adequacy of one model over another. This article describes several studies on model-based reasoning to generate hypotheses about teacher education and respond to the overarching research question, how can preservice chemistry teachers reason with models in teacher education?

The purpose of investigating science teacher reasoning is because teachers support students who actively reason about the world. Furthermore, advancing the UN sustainable development goals (SDGs) can occur through science teacher education. SDG 4c states that by 2030, the goal of the UN is to substantially increase the supply of qualified teachers. This goal is supported by educational research that suggests that teacher education is one of the strongest correlates of student achievement in math and science (Darling-Hammond, 2000; Monk & King, 1994). Reasoning is one of the areas that is important to investigate in teacher education as teachers may influence their own students’ reasoning competencies in classrooms. Promoting reasoning in chemistry classrooms is ever more necessary to foster greater understanding of systems in flux that models of chemistry influence, such as climate systems. Studies on reasoning continue are prevalent in science education research (DeBoer, 2019). Pertinent to science teacher education; however, the research ranges from teacher general pedagogical reasoning (Loughran et al., 2016) among others, to teacher research on reasoning competencies (Göhner & Krell, 2022; Krell et al., 2020). This article aims to contribute to the latter arena.

2 Review of literature

2.2 Turning the lens to science and chemistry teacher education

Valuable research on scientific reasoning in laboratory and educational contexts has occurred (Hogan, 1999; Lawson, 1978; Schauble, 1996), yet fewer studies have been done on the state and advancement of scientific reasoning among university students who choose science or chemistry teaching as a career (Lawson et al., 2000). Kind and Osborne suggest that scientific reasoning is highly dependent on content and procedural and epistemic knowledge (Kind & Osborne, 2017). It is also arguably dependent on the context, such as undergraduate contexts. Much of the research on reasoning in undergraduate contexts has students working on artificial problems (Fischer & Bidell, 1998); however, classroom contexts with authentic teachers and students can yield special insights (Krell et al., 2020). Ding’s (2018) suggestion is that the kind of instruction in undergraduate education might result in limited advancement in scientific reasoning. Less is known about what happens to reasoning in chemistry teacher education contexts.

In terms of reasoning and chemistry teachers’ backgrounds, in 2023, my colleagues and I (Krell et al., 2023) examined the contribution of three factors to the development of preservice science teachers’ scientific reasoning competencies: the amount of science education classes, the amount of science classes and the preservice science teachers’ ages as they began their teacher education programs and progressed through completion of this university program. This investigation of teachers’ backgrounds involved six universities in Germany, Chile and Canada. Preservice science teachers anonymously responded to an established multiple-choice instrument and statistical method (Krell et al., 2021) that assessed scientific reasoning over the course of their teacher education programs, taking into account the number of credit hours of science courses and science education courses of various programs in these universities and countries.

Multiple linear regression showed in this study that amount of science education classes, the amount of science classes, and age explained about 9 % of preservice science teachers’ scientific reasoning. These preservice science teachers consisted of chemistry, biology, and physics teachers. Notably, the factor: amount of science classes, was the only significant predictor of scientific reasoning among preservice science teachers. This empirical finding on content knowledge is aligned with previous research that shows a positive relationship between the two (Fischer et al., 2014), but it contrasts with other studies that point to a positive correlation between the amount of science education classes and SRCs too (Bruckermann et al., 2018). It appears that science content knowledge may be requisite for developing preservice science teachers’ scientific reasoning.

The Ding (2018) study, on the other hand, is different from the research we did in two ways: first, a different test of reasoning was employed and second, the Ding study was only deployed within a single country. The aforementioned 2023 study; however, went beyond a single country. Furthermore, and in addition to reasoning involving scientific investigations, in the 2023 study, model-based reasoning was also examined. Model-based reasoning entailed judging the purpose of models, testing models, and changing models. Analysis of the 2023 questionnaire results revealed substantial differences among the three countries, as reported in Table 1 and published in Krell et al. (2023).

Table 1:

Sample models utilized by preservice teachers from their science curricula.

Grade	10	11	12
Curricular topics	Chemical reactions, life sciences	Atomic theory	Nature of acids and bases, dynamic equilibrium, electrochemical cells
Models	Ice	Atomic (Bohr) model	Arrhenius and Bronsted–Lowry
	Chemical respiration		Chemical equilibrium
			Lead-acid battery

These findings showed that Canadian preservice science teachers possess greater competencies related to model-based reasoning compared to either their German or Chilean preservice teacher counterparts. Regarding the Canadian sample, possible explanations for the significantly higher results could be that, (1) Canadian preservice teachers had more of a science background by about 1–2 years of study and that (2) teacher education or the kind of instruction had an impact on fostering model-based reasoning above and beyond their greater science backgrounds.

Model-based reasoning is a scientific practice that has been described as relational (Alexander, 2019; Sevian & Talanquer, 2014), mechanistic (Moreira et al., 2019), and analogic (Lehrer & Schauble, 2015). It has been proposed that model-based reasoning requires mental models, or internal representations of the way the world works (Pietarinen & Bellucci, 2014; Williams, 2018). These mental models are invoked when solving problems or reasoning about phenomena. One can have a mental model of climate change, Pandemics, rumours, ownership, locks and keys, atoms, or the planets, to name only a few.

Mental models can be expressed externally and socially negotiated (Campbell et al., 2015; Halloun, 2007; Khan, 2007; Khan & Chan, 2011). In teaching contexts, dialogic processes have been shown to be powerful for students engaged in model-based reasoning, especially when students themselves can point to initial activities involving modeling and author the stages (Windschitl et al., 2018). Reasoning with a model can also surface its limitations and new criteria for evaluating models (Lehrer & Schauble, 2005; Pluta et al., 2011). Epistemic messages, such as those on the purposes of adjudicating models, support greater understanding of the nature of models (Ke & Schwarz, 2021). Reasoning with models is not only done among scientists but also among teachers and students (Nersessian, 2008).

In a 2007 study on model-based reasoning in chemistry teaching contexts, Khan compared undergraduate first year chemistry classrooms for their pedagogical approaches. The approach that characterized the vast majority of chemistry teaching involved beginning the undergraduate class with defining concepts or terms in chemistry, the instructor showing how to solve a problem, and time for students to receive and practice similar problems. This approach could be termed a traditional approach to teaching chemistry in the undergraduate department, as represented by Figure 1 (Khan, 2001).

Figure 1:

Traditional approach to teaching chemistry (Khan, 2001).

One course; however, outperformed others on a general scientific reasoning test given to multiple institutions science classes (Khan, 2001). The approach to teaching by this chemistry instructor could be characterized as model-based or the generate-evaluate-modify (GEM) approach, a first case of model-based instruction inside a classroom (Khan, 2007).^[1] Figure 2 depicts the GEM approach to instruction. This approach to instruction involved the teacher promoting the generation of relationships internal to a model, evaluation of those relationships (often in light of new information), and modification of their models. Evaluation and modification occurred repeatedly, until a more accurate consensus model was co-constructed with teacher guidance in the chemistry undergraduate class.

Figure 2:

GEM instructional cycle (Khan, 2007).

The GEM approach that was observed in this chemistry undergraduate class involved asking students to reason about unobservable phenomena in science. In the first phase, students generate a relationship between two or more variables in a dialogic environment using information or data sets. In one example in the undergraduate chemistry classroom of an investigation of molecular models, the undergraduate teacher encouraged chemistry students to explore the relationship between vapor pressure and boiling points. Students also generated relationships between molecular weight and the mass of a compound in this phase of instruction. In the next phases of instruction observed, chemistry students continued to reason about their models of molecular compounds. In this phase, they grappled with information that led them to re-evaluate their originally formed relationships between molecular weight and boiling point. This information came from a simulation. The chemistry students noted that the higher the molecular weight, the higher the boiling point.

The chemistry students saw; however, that methanol and methyl amine fell out of the expected trend using a simulation. This was the beginning of a discussion on what kind of bond could exist between two hydroxyl groups as exclaimed by a student question, “What kind of bond would there between two hydroxyls?” While the idea of a bond is not correct, the undergraduate students went on later to invent a hidden causal factor of what would later be termed by the teacher as an intermolecular force. Their newly revised models of a chemical compound better aligned with the anomalous data points as it later took into account intermolecular forces (Khan, 2007). The GEM pedagogical approach was elucidated through several years of classroom analysis, and 34 guidance strategies to scaffold model-based reasoning were identified from observations of this approach (Khan, 2011). This approach to teaching chemistry did not require reading in advance, and the undergraduate chemistry teacher did not correct students right away. Instead, reasoning was fostered and sustained during GEM discussion. Recent research in science education settings further reports on the benefits of model-based instruction on student reasoning in undergraduate chemistry settings (Cooper et al., 2017), and even students elementary settings are using chemical models (Baumfalk et al., 2019) with similar instructional techniques.

Given that future teachers often come from this pool of undergraduate science, it is fruitful to explore whether teacher education itself could have an impact preservice teachers’ scientific reasoning competencies. In a third study, Khan and Krell (2019) used the same instrument as the international comparison (a validated pre- and post-questionnaire) to investigate a single teacher education class more closely. That is, the single teacher methods course was assessed for whether preservice science teachers’ scientific reasoning was the same before a science teacher methods as afterwards. The teacher education course contained common science methods topics, such as unit and lesson planning and assessment techniques in science.

The preservice science teachers were asked to reason about two types of problems in a validated pre- and post-questionnaire (Krell et al., 2020): investigatory-process problems and modeling problems. Statistical analyses of the data revealed that teacher education significantly contributed to the development of preservice science teachers’ competencies for those who had two previous degrees compared with those with one. Furthermore, when grouped together, a greater proportion of preservice teachers were better at planning investigations and analyzing data-moreso than any dimension associated with model-based reasoning in the questionnaire, and formulating questions and generating hypotheses. In a related study which included teacher specializations and the same questionnaire (Khan & Krell, 2021), chemistry preservice teachers had the lowest probability of answering items on the scientific reasoning questionnaire correctly compared to preservice biology teachers. Preservice chemistry teachers were found to be significantly worse at planning investigations, analyzing data, drawing conclusions, testing models, and changing their models than their biology counterparts. The questionnaire items were more procedural and epistemic rather than content-based, leading us to renew calls to examine chemistry preservice teachers.

The former findings (Krell et al., 2020) reify the notion that content knowledge may be a potential mediator of science teachers’ reasoning. The former findings are also aligned with the aforementioned cross-country comparison in 2023 of preservice teacher education (Krell et al., 2023), that affirmed that the amount of science classes is a significant predictor of SRCs in comparison to age and amount of teacher education courses. The latter findings (Khan & Krell, 2021) illuminate that chemistry preservice teachers fare significantly worse than biology teachers at scientific reasoning-including model-based- reasoning, even though both have a degree in science. Questions remain as to how teacher education might amplify the model-based reasoning of chemistry teachers.

3 Case study of a science teacher education course

While there is established quantitative and case evidence that science content courses are associated with gains in model-based reasoning skills (Khan, 2007; Krell et al., 2023), there is less evidence of similar gains in our research from participation in science teacher education courses (Khan & Krell, 2019). Investigating how teacher education programs can be designed to better promote the development of model-based reasoning skills among preservice chemistry teachers would be valuable. A current program of research is exploring this very question (Faikhamta et al., 2024; Khan, 2018; Khuyen et al., 2024). From Khan and Krell (2019)’s study in the same context as the present study, it had already been found that a science teacher education methods course had the potential to foster scientific reasoning competencies of some preservice teachers. In this study, preservice teacher competency in SRC significantly improved (for those that had 2° in science) after a teacher education course where no changes had been made to the course syllabus. This finding stands in contrast to our larger study where the amount of teacher education did not make a significant difference to SRC (Krell et al., 2023), leaving to question how intentionally designed teacher education activities might impact preservice teacher SRCs. Also, it was considered that there might be other ways to detect teacher competency at scientific and model-based reasoning, including through an analysis of preservice science teachers own actions to design and promote scientific reasoning to their students. It is thus theorized that opportunities to engage in teacher education activities to discuss model-based reasoning and foster them through the design and teaching of lesson plans will support preservice teacher’s own competency development (Khan & Krell, 2019). In this way, the kind of instruction in teacher education is hypothesized to impact model-based reasoning above and beyond preservice chemistry teacher’s previous undergraduate degree in science.

The following case is an example of a wider program of research in science teacher education to investigate high impact activities for preservice chemistry teachers. Below are excerpts of this study of a science teacher education methods course that was entirely reformed to support model-based reasoning. The revisions to the course included: (1) an expanded focus on how science proceeds using readings, depictions of science, and black box activities, (2) additional opportunities to teach and reflect upon the reasoning skills associated with modeling, including the use of a teacher design guide and class debriefs, and (3) opportunities to test teaching to promote model-based reasoning across three teaching contexts. Teacher education activities were operationalized as being successful if preservice teacher’s could plan for and facilitate their students’ model-based reasoning across multiple teaching contexts.

The design of this case study of teacher education activities followed a longitudinal pre-post design. Preservice teachers’ model-based reasoning was observed across three contexts: the science methods course itself, the practica experience, and a required assignment where preservice teachers taught in after-school organizations. Data sources included: a new pre- and post-questionnaire on the nature of modeling and model-based reasoning (different from the aforementioned SRC questionnaires), 2 sets of preservice teacher lesson plan assignments at earlier and later times in the refreshed course, and highschool classroom observations and debriefs, coded in a rubric. Data was collected during micro-teaching, on practicum, and in out-of-school settings, such as homework clubs. To illustrate preservice teacher reasoning with their students, only several excerpts from the data are presented.

The chemistry preservice teachers were advised to focus on a model from their chemistry curriculum to design and enact lessons. The models were from the secondary chemistry curriculum (constituting grades 8–12). The models reflect conceptual/symbolic (Coll & Lajium, 2011) or concrete process models, containing some abstract and concrete elements, according to the typology offered by Harrison and Treagust (2000). For example, models of electrochemical cells contained conventional elements of chemical batteries technology and wires with electron flow to and from the battery. Given that the models were evident within the secondary curriculum, they could also be referred to as “expressed consensus teaching” or “curricular models” (Chamizo, 2013) in chemistry education.

4 Results

To generate hypotheses about how science teacher education activities can be designed to better promote the development of model-based reasoning, analyses of three excerpts from the data on preservice teachers reasoning with models in their teaching is reported. The specific models selected by the preservice teachers were represented in Table 2.

Study questionnaire results pinpointed activities that preservice teachers reported that may have had an impact on their reasoning. Initial hypotheses were generated on the types of activities that could promote the development of model-based reasoning among preservice chemistry teachers.

For the first excerpt, a preservice science teacher designed a lesson plan in the teacher education course, engaging two students to reason about a model of an electrochemical cell. In this instance of microteaching, other preservice teachers served in the role of ‘students’. To begin, the preservice teacher in this excerpt asked her ‘students’ to share their ideas about a battery. After encountering data using a battery simulation, ‘the students’ identified themselves that their expressed model of a battery works is wrong. The preservice teacher prompted ‘students’ to revise their models and arrive at a consensus. The teacher reframed the discussion around modeling as the “act of building on our knowledge”.

She then was observed posing a question after asking the students to hook up the battery in a simulated circuit:

And if I can get you to label the positive and negative terminals, and then what you think the electron distribution will look like inside of this battery. I see both of you the electron flow from the positive terminal to the negative terminal….If we are saying the bulk of the electrons are being stored at the negative terminal, does that make sense? … Why would it, wouldn’t it be easier for the electrons just to flow through the battery back to the positive terminal instead of going through a circuit? Do you still think that the electrons in the new circuit will follow that pattern?

The preservice teacher, pointing to the battery, prompted her students to reason about the battery and why it would not have electrons circulate in a different pattern. This episode was coded as an example of teacher reasoning with students, because she asked students to consider a pattern (underlined) and an anomaly in the pattern (does that make sense). The student said their original models of a battery and circuit “made no sense”, and the preservice teacher responded let’s discuss why we think that.

An analogy was later brought up by the preservice teacher to help her students grapple with apparent anomaly in their electron flow models in this microteaching episode. This spontaneous teacher analogy was to water in a pipe system, and it was coded as an example of analogical reasoning using a model. The preservice chemistry teacher reasoned with her students at the start of the dialogue, “[I]f you have a lot of water at the top, in a basin at the top, and a pipe connecting the basin at the top, and then one that’s empty, just slightly lower down, where is the water going to flow as the pressure builds up? Is it going to flow from where there’s lots of water to more, or more water?” As an illustrative excerpt, the preservice chemistry teacher could be said to be engaging in reasoning with her students and their models.

In similar ways, reasoning was apparent in high school teaching contexts too. For a second illustrative example of preservice teacher reasoning about models in a high school context, two preservice teachers working in an after school club experimented with melting ice with their students, as they wrote for their debriefs:

…performed a science experiment which dealt with melting points. We encouraged the [high school] students to hypothesize which type of household seasoning (sugar, salt, cinnamon, pepper) could be used to lift an ice cube out of a container of water using only a thread … We began by asking if any of the students had an idea as to how we could accomplish the task of lifting the ice cube out of the water without touching it. The students then generated an idea about how the string could be used to pull the ice cube up. We then informed them that they would be able to also utilize one (or more) of the household seasonings to accomplish this task. The students proceeded to come up with different scenarios using the different seasonings until the desired outcome was achieved. After they had selected the correct seasoning (salt) we discussed what they though was happening, and why salt was the only seasoning to be successful.

The preservice teachers appear to be working with their high school students to develop an explanatory model of water. In a third excerpt, the preservice chemistry teachers debriefed what had happend and their reasoning processes as they asked key questions of the high school students about their models of water.

Interestingly one of the students thought that sugar would be the successful lifting agent, as he hypothesized that sugar is ‘sticky’ and should therefore result in the thread adhering to the ice cube surface. The sugar actually did result in a bit of stickiness however it was not sufficient to support the weight of the ice cube. He then decided to try a combination of salt and sugar together, however this did not work either, likely due to the ratio which was predominately sugar. The student appeared to be very frustrated when the sugar was not successful, and became visibly upset by this. Clearly this idea that sugar is sticky, and should therefore cause the thread to attach to the ice cube, was quite engrained in his mind, and as a result it was difficult [for him] to explain why salt was actually the correct ingredient. We also asked them to explain their answers as to why something did not work or did work. The students needed to understand that salt can melt ice, and the ice would refreeze the water over the thread, keeping it firm in place as you raised the ice.

The preservice teachers promoted reasoning by helping the student eliminate possibilities as underlined (e.g., salt and sugar), postulate hidden causal factors (stickiness), and better attempt to coordinate their models of water with the evidence from the melting experiment (teacher asks why something [sugar] does it not work). In terms of success, all of the preservice teachers were able to facilitate model-based reasoning in these ways over multiple teaching contexts. Compared to the pre-questionnaire which asked preservice teachers to reason with a model about chemical equilibrium, their post-questionnaire showed evidence of improvement in teacher reasoning with models, leading one to believe that activities in the course may have contributed in some way to their own competencies at model-based reasoning.

Regarding the kind of instruction in teacher education that may have contributed to an impact on the preservice chemistry teachers, they reported in the questionnaire that learning that their students’ might have alternative conceptions had the highest impact course activity. One can hypothesize that preservice science teachers learning that chemistry students have alternative conceptual models would support the idea that reasoning involves changing one’s mental models of the way the world works. Ranking close to the top were designing GEM lessons. The preservice teachers were provided a design guide to support their efforts at creating GEM lessons using this approach to foster model-based reasoning. Finally, reflection on teaching was ranked lowest by the preservice teachers, perhaps because it was a program-wide activity. The teacher education course and the community site were reported by the preservice teachers as the contexts with which preservice teachers were most able to try methods to promote model-based reasoning, according to the questionnaire. The excerpts provide a window into activities with students that might enhance their own scientific reasoning competencies.

5 Discussion

It is hypothesized that teacher education course activities may foster chemistry teachers own scientific reasoning competencies. A preservice teacher education course was intentionally designed to support preservice teacher’s model-based reasoning. Preservice teachers acknowledged that the highest impact activities in the course that supported their own understanding of model-based reasoning included learning about students’ alternative conceptions, GEM lesson planning, and enactments in the teacher education course and in community sites. The first two activities were supported by readings and a design guide, and the latter, teacher enactments, were supported by reflection activities including debriefs. Lesson observations and a post-questionnaire revealed that not just their pedagogy to facilitate high school students’ model-based reasoning improved, but there was evidence from the post-questionnaire that so did their personal reasoning competencies. A hypothesized mechanism for doing so might be that the above course activities required preservice teachers to plan for and engage in prolonged facilitation of reasoning with models. They could not anticipate fully what their students might say in the different, requiring the preservice teachers to actively reason with their students about their models. The illustrative excerpts from the data suggest preservice teachers’ reasoning with a [student] model happened when they helped high school students compare evidence with their personal models, develop explanations, eliminate less probable explanations for experimental findings, resolve anomalies, and postulate hidden causal factors. Preservice chemistry teachers were also observed spontaneously engaging in analogical reasoning to support students’ understandings of models.

Taken together, it is plausible that teacher education courses, intentionally designed, can create opportunites that foster teacher reasoning and elevate their own competencies at reasoning, as suggested by Khan and Krell (2019) in their case study of a different methods course. As Ding (2018)’s study alludes; however, SRCs may not be developed in undergraduate settings, the first degree for teachers in a consecutive program. Thus, explicit teaching of SRCs may be necessary in post-secondary education first degrees where the pool of future teachers often comes from. Despite science content courses being a significant predictor of SRCs (Krell et al., 2023), this case study suggests, in a preliminary way, that chemistry preservice teachers have the potentail to advance their SRCs from intentionally designed teacher education activities.

6 Recommendations

The research program investigating reasoning in science teacher education is especially important for chemistry education. Research on prior degrees (Krell et al., 2023) suggests further that content knowledge is a potential mediator of chemistry teachers’ capacities to reason. Therefore, the first recommendation for chemistry teacher education is that science content knowledge is considered when developing chemistry teacher education methods courses. Additionally, preservice science teachers who may teach chemistry in the future but do not have a chemistry major could be supported in teacher education with courses for non-chemistry majors. Consecutive teacher education programs appear to be the beter way to design a program that fosters teachers SRCs compared to concurrent programs (Krell et al., 2023), as content courses in an undergraduate degree are significant predictors of future SRCs.

Secondly, preservice chemistry teachers significantly underperform on questions involving scientific reasoning compared to biology preservice teachers (Khan & Krell, 2021). Based on the aforementioned research, explicit instruction of scientific reasoning would be especially important within the first chemistry degree. Teacher education interventions are also recommended. In Engelmann et al.’s (2016) meta-analysis of 15 studies to promote scientific reasoning, it was found that the interventions had a significant positive effect on scientific reasoning. Concerning interventions, this present research program in teacher education has developed targeted, high impact activities to promote model-based reasoning.

These teacher education activities include approximating practice across contexts and engaging in GEM activites that promote model-based reasoning. Science preservice teachers, despite working with a plethora of models in their first degrees, fare worse with regards to model-based reasoning than other questionnaire items (Krell et al., 2020). Course-based activities, such as those that promote model-based reasoning in various contexts, may also help preservice teachers activate more spontaneous forms of reasoning with students. It is recommended that similar high impact activities such as GEM (Khan, 2007) be integrated in teacher education and that multiple opportunities to teach in various contexts be sought.

In addition to the design of consecutive teacher education programs and intentional and explicit activities to promote reasoning in chemistry and chemistry education, additional research is necessary. First, evidence for an undergraduate class producing higher achievement than other chemistry classes was provided (Khan, 2007), and in another study, evidence for being able to teach using model-based reasoning successfully in teacher was provided (Khan & Krell, 2019). Based on the present study in teacher education that also incorporated a model-based reasoning approach, GEM, the use of this instructional approach may serve as one avenue to sustain and amplify teacher reasoning. More research; however, is needed that shows the relationship between course activities and preservice chemistry teachers outcomes in terms of their own SRC to begin to make clear claims about the impact of teacher education. It is plausible that the act of teaching itself promotes reasoning and over time and across contexts, fosters reasoning competencies as a result. The nature of the relationship is not yet fully understood or documented. Thus future research could examine this relationship in teacher education. Also, research on how to best support reasoning competencies in undergraduate education would complement ongoing projects in teacher education. Finally, greater specificity on the classroom activities that have the best possible impacts on future chemistry teachers is needed. Bolstering the reasoning competencies of teachers is one proposed way to students capacities to reason, a fruitful avenue to explore in chemistry education research.