Lisa
Shah
a,
Christian A.
Rodriguez
a,
Monica
Bartoli
b and
Gregory T.
Rushton
*ab
aDepartment of Chemistry, Stony Brook University, Stony Brook, New York 11794, USA
bInstitute for STEM Education, Stony Brook University, Stony Brook, New York 11794, USA. E-mail: gregory.rushton@stonybrook.edu
First published on 19th February 2018
Instructional strategies that support meaningful student learning of complex chemical topics are an important aspect of improving chemistry education. Adequately assessing the success of these approaches can be supported with the use of aligned instruments with established psychometrics. Here, we report the implementation and assessment of one such curriculum, Chemical Thinking, on first-year general chemistry students' conceptions of relative acidity using the recently-developed concept inventory, ACIDI. Our results reveal that, overall, students performed significantly better on ACIDI following instruction, with scores consistent with those previously reported for students who had completed one semester of organic chemistry. Students performed equally well on a delayed post-test administered ten weeks after final instruction, which suggests that instruction promoted a stable conceptual reprioritisation. Item analysis of ACIDI revealed that students generally made conceptual gains on items where inductive effects were the primary determinants of conjugate base stability and relative acidity. However, students overwhelmingly struggled on items where resonance was the primary determinant. Analysis of student–student arguments in active learning settings provided evidence for how the quality of student arguments impacted their conceptions. Overall, these findings suggest that students were able to avoid several superficial misconceptions cited in the literature about relative acidity, and that this topic, traditionally taught exclusively in organic chemistry, may be introduced earlier in the sequence of curricular topics. Implications for future studies on the role of argumentational aspects of student–student conversations and facilitation strategies in promoting or hindering meaningful learning are discussed.
Acid–base chemistry is one of the most fundamental, yet challenging, topics in the discipline. The majority of organic and bioorganic processes include an acid–base reaction at some step (Bhattacharyya, 2006; Brown et al., 2009). However, current approaches to teaching acid–base chemistry tend to leave students with a superficial understanding of related concepts, which leads to the misapplication of heuristics in making qualitative predictions (Bhattacharyya, 2006; Rushton et al., 2008; McClary and Talanquer, 2011a, 2011b; Stoyanovich et al., 2014; Cooper et al., 2016). For example, at the introductory level, students are known to rely on the presence of an ‘H’ or ‘OH’ to identify acids and bases, respectively (Furió-Más et al., 2005), and have difficulty understanding the calculations associated with acid–base equilibria (e.g., pH, buffers, titrations) (Nakhleh and Krajcik, 1994; Demerouti et al. 2004a, 2004b). Little work has been done to examine students' understanding of acids and bases in more advanced college-level courses (Bhattacharyya, 2006; Ferguson and Bodner, 2008; Cartrette and Mayo, 2011; McClary and Talanquer, 2011a, 2011b; Stoyanovich et al., 2014). Bhattacharyya (2006) reported on graduate students' fragmented mental models of Brønsted acids and Ferguson and Bodner (2008) reported on undergraduates’ superficial use of arrow formalism in the mechanistic explanation of acid–base reactions. With respect to relative acidity of organic acids, McClary and Talanquer (2011a, 2011b) have examined undergraduate students' over-reliance on heuristics (e.g., using surface-level functional groups as key indicators of acidity) as a means to compensate for their lack of deep, conceptual understanding. Specific instances of this type of reasoning are noted in the later work of McClary (McClary and Bretz, 2012), who found that students were often misguided by the presence of an explicitly-drawn hydrogen in a compound as an indication that it was more acidic than one without this representation because it could exclusively be deprotonated. These findings offer a useful perspective of the knowledge gaps of “successful” students and suggest that traditional instructional methods (i.e., didactic lecture) may be less sensitive in ways that facilitate long-term retention of core concepts.
A popular approach to improving performance and retention of content within higher education chemistry communities has been the use of cooperative learning pedagogies, in which students engage in the completion of activities designed to encourage collaboration and reasoned argumentation to promote learning (Cooper, 1995; Gupta, 2004; Moog et al., 2006). Recent meta-analyses of these approaches in chemistry have reported improved student achievement across instructional contexts (Warfa, 2015; Apugliese and Lewis, 2017). However, the more recent development and use of norm-referenced instruments to measure learning gains have revealed variations in reported results (Lewis and Lewis, 2005; Hein, 2012; Chase et al., 2013), which have led researchers to consider the means by which cooperative learning does or does not achieve desired student outcomes (Towns and Kraft, 2011; Stains and Vickrey, 2017). Despite these differences, few would argue that cooperative learning approaches as a whole are less effective than traditional didactic lecture pedagogies. Rather, the details of these implementations and whether or not they are being used alongside curricula that promote conceptual understanding over rote memorisation (Demerouti et al., 2004a) may be a significant factor in their observed impact on student performance.
Recently, Talanquer and co-authors have reported the development of a revised curriculum for teaching general chemistry called Chemical Thinking (Talanquer and Pollard, 2010; Sevian and Talanquer, 2014). In this curriculum, content is organised into units shaped by fundamental questions that guide the practice of chemistry (e.g., How do we distinguish substances? How do we determine structure?). These questions inform a discussion-oriented focus of the curriculum that goes beyond algorithmic applications of knowledge by providing opportunities for students to engage in conversation with peers or facilitators in order to make sense of concepts and ideas throughout the course. Lecture slides often include experimental data or graphs alongside thought-provoking questions meant to initiate class discussions, and similar materials are provided for small-group cooperative learning sessions. Given the highly-conceptual nature of this curriculum, we hypothesised that this type of intervention might be useful for promoting students' understanding of challenging concepts by encouraging and rewarding thoughtful discussion of course topics.
An implicit assumption made by instructors in cooperative learning chemistry courses is that students benefit from environments where small-group, student–student interactions complement the patterns in instructor–student conversations of more didactic settings. Advocates of active learning hold to the premise that peer discussions can encourage argumentation and socio-cognitive conflict to promote conceptual reprioritisation, making them at least as valuable as direct instruction from lectures (Cooper, 1995; Limón, 2001). The quality of these arguments likely dictates the extent to which students undergo conceptual reprioritisation (Cazden and Beck, 2003), as researchers associate students' use of well-constructed, scientific arguments with conceptual understanding (Kittleson and Southerland, 2004). Argumentation analysis has more recently been used by chemical education researchers to investigate student–student and student–instructor interactions in small-groups (Cole et al., 2012; Kulatunga et al., 2013, 2014; Warfa et al., 2014). An analysis of the quality of student arguments can yield insights about the specific instructional aspects that may be beneficial or counterproductive to student learning.
However, McClary and Talanquer (2011b) cite that a barrier to the uptake of new instructional approaches is their evaluation due to a lack of aligned assessment instruments, specifically those with the ability to detect student conceptions. Recently, chemical education researchers have worked to develop valid and reliable concept inventories, cognitive instruments with items designed to identify the strength with which students hold normative or alternative ideas about chemical topics. When used in conjunction with a curriculum that facilitates conceptual understanding, concept inventories are proposed as measures of the effectiveness of the approach (Bretz, 2014). In the past decade, topic-specific (e.g., chemical bonding, acid–base chemistry, equilibria, redox) concept inventories have been developed for use in general chemistry courses (Mulford and Robinson, 2002; Evans et al., 2003; Bardar et al., 2007; Brandriet et al., 2011; McClary and Bretz, 2012; Luxford and Bretz, 2014), which affords greater opportunities to compare interventions within and across institutions and topics. ACIDI is one such inventory developed by Bretz and coworkers (McClary and Bretz, 2012; Bretz and McClary, 2014), designed to investigate undergraduate students' conceptions of factors governing relative acidity based on previous studies that explore students' use of mental models in ranking and justifying trends in organic acid strength (McClary and Talanquer, 2011a). The combination of answer, reason, and confidence tiers throughout the assessment provides instructors with valuable insights about which normative or alternative conceptions students hold and how strongly they adhere to them.
The focus of this study was to evaluate a curriculum that introduces first-year general chemistry students to the topic of relative acidity using the discussion-oriented curriculum Chemical Thinking. ACIDI is well-aligned to this curriculum and was used to evaluate both specific student conceptions about acid strength and the effectiveness of instruction at helping students avoid common misconceptions cited in the literature (McClary and Talanquer, 2011a; McClary and Bretz, 2012). Students' first experience with new material was in the context of small-group, cooperative learning settings, and analysis of group conversations provided insights about how the quality of arguments may have impacted student conceptions. Our findings and conclusions speak to the effectiveness of such a curriculum at moving first-year students beyond surface-level conceptions of acid strength and help identify specific areas for improvement. From a broader perspective, implications for future studies on the role of argumentation in student–student conversations and facilitation strategies in promoting or hindering conceptual understanding are discussed.
To accommodate new ideas and beliefs, some researchers have suggested that a conceptual ‘reconstruction’ must occur (Hammann et al., 2008; Duit and Treagust, 2012). More recently, however, others have proposed that naive or preconceptions can coexist with scientifically normative ones (Clement, 2008; Potvin et al., 2015; Potvin, 2017), and as a result of instruction, students can come to prioritise one over the other. Literature regarding conceptual reprioritisation has primarily been situated within the context of traditional constructivist theories of learning, where the mental structures of individual learners and their pre-existing knowledge base are critical determinants of how new knowledge will be incorporated, and whether this integration of knowledge will result in the prioritisation of normative or alternative conceptions (Duit and Treagust, 1995; Limón, 2001). However, traditional constructivist theories operate under the assumption that individual learners come to prioritise certain conceptions over others when presented with information in isolation from their social and cultural environments. Sociocultural theories of learning, in contrast to this perspective, claim that individuals are inseparable from their respective cultures and that cultural tools (e.g., language, symbols) employed in social interactions mediate the internalisation of knowledge (Vygotsky, 1980, 1997). Our position, consistent with that of social constructivists, is that conceptual reprioritisation necessitates the coordination of ideas from both theories, with language as the mediating factor between social and individual planes (Leach and Scott, 2002). O'loughlin's (1992) perspective highlights the interplay between these two views:
“…meaning making is neither exclusively a product of the person acting, nor of the activity, nor of the setting, but a dialectical interaction between all three in a given context. This too is a form of constructivism, but it is one that emphasises the subjectivity, the sociocultural situatedness, and the intrinsically dialectical nature of the process of coming to know.”
We argue that while individual learners may ultimately come to prioritise their conceptions themselves, they do not do so passively. By engaging in social interactions, they make sense of the concepts encountered in the intermental (i.e., social) plane before being able to fully internalise knowledge and prioritise certain conceptions over others.
With respect to instruction, social constructivist views support the use of discussion-based approaches to facilitate individual conceptual reprioritisation. Discussion actively promotes efforts to make meaning of the content (Moog et al., 2006; Moon et al., 2016) and explain ideas (Inagaki and Hatano, 2008; Fonseca and Chi, 2011), and allows students to evaluate the claims of others (Greene et al., 2008). It further enables students to publicise their views, consider their peers’ alternatives, and resolve issues stemming from opposing viewpoints (Warfa et al., 2014). While discussion can involve a myriad of oral interactions, argumentation plays a critical role in the construction of knowledge within the scientific community (Erduran et al., 2004) and among students in science classrooms (Driver et al., 2000). Argumentation is the specific process by which individuals with opposing initial conceptions come to accept or refute presented claims. When these claims are directly related to specific normative or alternative conceptions, students' abilities to construct a persuasive and convincing argument is a critical determinate of the conceptions toward which they and their peers gravitate (Nystrand et al., 1997; Teasley, 1997; Wells, 2007; Rapp, 2014). A number of studies on the quality of argumentation in educational environments have applied Toulmin's model (Toulmin, 1958) as an analytic framework, in which categories of (i) claims, (ii) evidence/data (iii) warrants, and (iv) rebuttals are among those used to examine the structure and quality of student arguments (Voss et al., 1983; Kelly et al., 1998; Cho and Jonassen, 2002; Erduran et al., 2004; Kenyon and Reiser, 2006). Erduran et al. (2004), for example, used Toulmin's model to develop a rubric to assess the quality of arguments based on the presence or absence of specific argument features. In this rubric, arguments are rated Level 1 (e.g., simple claim-versus-claim) through level 5 (e.g., claims presented with data countered by several rebuttals). We employ the same rubric in our study (see Appendix Table 4) as a means of evaluating the quality of arguments constructed by students in the cooperative learning workshops. The relative quality of student arguments serves as evidence for how they may have impacted student conceptions. Therefore, instructional approaches that encourage deep discussion are likely to elicit quality arguments between students with opposing conceptions that drive internal socio-cognitive conflict, ultimately promoting student learning.
(1) What is the effect of the curriculum on first-year general chemistry students' normative and alternative conceptions about relative acidity?
(2) How does the quality of student arguments in cooperative learning environments help explain students' normative or alternative conceptions?
Content validity of ACIDI was established by an organic chemistry faculty member who teaches study participants in the following course (Organic Chemistry I). He agreed that many aspects of the instrument adequately assessed students' understanding of relative acidity, but stated that student interviews would be critical in identifying their reasons for selecting particular options. Students' cognitive interview responses to the statement “read the question and tell me what you thought it was asking” were used to further establish content validity (see Appendix). These types of interviews are valuable for assessing the nuances of particular assessment items (Carmines and Zeller, 1979). Cronbach α values (Cronbach, 1951) of 0.30, 0.28, and 0.29 (a measure of internal consistency) were determined for the pre-test, post-test, and delayed post-test respectively, consistent with low α previously reported for this concept inventory (McClary and Bretz, 2012). While the accepted standard for internal consistency is α > 0.70, the authors of this inventory have described that diagnostic tools aimed at identifying student misconceptions are unlikely to meet this criteria due to the lack of redundancy among items (Bretz and McClary, 2014). The first two authors independently coded all workshop transcripts a priori using Toulmin's framework (i.e., claim, data, warrant, rebuttal) before meeting to compare codes. After comparison, the two authors were able to achieve an inter-rater reliability of 79.7%. The two authors discussed the remaining 20% until 100% agreement was reached. These authors then coded each individual or co-constructed argument using the framework developed by Erduran et al. (2004). The authors were able to independently code and subsequently compare to reach 100% agreement.
To investigate whether students were aware of what they understood well or poorly, mean percent confidence quotients (CDQ) were calculated for each item, as previously reported (Cartrette and Mayo, 2011) (Table 2). Briefly, the CDQ quantifies the standardised difference in confidence between those who answered an item correctly versus incorrectly. A negative CDQ is an indication that students are not aware of what they do not understand while positive values indicate that students are aware of what they understand. Unlike students in the McClary and Bretz (2012) field test sample, students in this course who responded correctly were more confident on average, with positive CDQ scores for six of the nine items. Interestingly, all five items prompting the use inductive effects to gauge relative acidity yielded positive CDQ values, the largest of which occurred for item five. Consistent with post-test performance (Fig. 1), items two and three yielded negative CDQ values, which indicated that students were not aware of their misconceptions concerning the compounds in this set.
Resonance | Inductive effect | ||||||||
---|---|---|---|---|---|---|---|---|---|
Q1 | Q2 | Q3 | Q4 | Q5 | Q6 | Q7 | Q8 | Q9 | |
Total mean % confidence | 68.36 | 68.7 | 70.28 | 63.96 | 63.54 | 60.22 | 62.30 | 59.55 | 51.90 |
Mean % confidence-correct | 66.43 | 63.75 | 61.25 | 66.76 | 64.07 | 63.61 | 64.21 | 63.00 | 57.36 |
Mean % confidence-incorrect | 69.05 | 69.46 | 71.43 | 61.11 | 50.5 | 58.49 | 61.18 | 51.07 | 47.63 |
Mean % confidence quotient | −0.12 | −0.26 | −0.46 | +0.26 | +0.55 | +0.23 | +0.13 | +0.48 | +0.39 |
I think [phenol]'s definitely going to be the stronger of the two compounds as far as its acidic nature just because in [phenol] you have exactly the same compound… and then in [p-methylphenol] instead of just that you have the methyl group added to it. Again, that's going to kind of help donate and push electron density up towards the oxygen and hydrogen which are bonded together, which is going to again make it less likely to be deprotonated. (Michael)
The only difference is the methyl group. And then I think… I think carbons are supposed to be electro… electron donating groups so since there's another carbon in there it wouldn't help stabilise the conjugate base. (Sara)
Students seemed to recognise the impact of induction by neighbouring functional groups on the relative stability of the conjugate bases of organic acids, and used their knowledge of this topic to rank and justify the relative acidities of p-methylphenol and phenol (Table 1) in these items. These quotes suggest that seven of the twelve students interviewed were able to rationalise that the methyl group of p-methylphenol would destabilise the negative charge in its conjugate base relative to phenolate ion from the deprotonation of phenol, consistent with the approach presented by the Chemical Thinking curriculum.
On items 5 and 6, interviews helped identify an alternative conception held by seven of the twelve interviewed students.
… because the O is closer to the H so the hydrogen being closer to the O makes it more acidic. (Carlos)
… because I figured that the hydrogen would be closer and more likely to receive the inductive effect of oxygen. (Edwin)
When comparing the relative acidities of acetone and acetaldehyde (Table 1) these students assumed that the explicitly-drawn hydrogen of the aldehyde was the most acidic hydrogen, which impacted the way they appealed to the inductive effect to support their answer. While the approach of examining the inductive impact of neighbouring groups was employed in a manner consistent with Chemical Thinking, interviews revealed an alternative conception, where the idea that the aldehyde hydrogen would be deprotonated (instead of one of the α-hydrogens) misguided their reasoning.
When asked to justify why p-nitrophenol was the strongest acid among the substituted phenols (Table 1, Set 3), six of the twelve interviewed students echoed the sentiments of the following quotes.
I said that the presence of the NO 2 will make [p-nitrophenol] the most acidic out of the three compounds because it has that electron pulling effect that perhaps would be better at stabilising a negative charge that would be present in the conjugate base. (Ron)
… the electronegativity from the N would help draw electrons toward it so it would help stabilise the conjugate base, more or less. (Ralph)
Students made statements suggesting that the electron-withdrawing effect of the nitro group in p-nitrophenol in stabilising the charge in the conjugate base would yield the most stable conjugate of the three species. Their unfamiliarity with the Lewis structure of the nitro group (a topic that was not explicitly covered in our course) may have prevented them from identifying resonance stabilisation as an additional influencing factor.
Overall, interviews with students regarding items 5–9 revealed that at least half of the students interviewed reasoned through these items using the normative conceptions presented in Chemical Thinking.
Just saying it has two carbonyl functional groups doesn't really say anything, it just says “oh that's there” but it's not really an explanation as why is more acidic. It doesn’t have anything to do with stability or resonance, which would justify it being more acidic. (Ralph)
…I could tell that just because something's a carboxylic acid or like an oxyacid per matter doesn't necessarily justify why it's more acidic than something else… (Jim)
Half of the students interviewed clearly stated that the presence of particular functional groups in general was not a sufficient justification for increased acidity and, instead, justified this ranking as follows:
…because it has the resonance to balance the conjugate base and make it more stable, which would add to the acidity of the acid. (Carlos)
So hydrogen is just bonded to carbon in all of these so the only one with resonance… well [pentane-2,4-dione] had the most resonance structures so I figured that it was the most likely to be the strongest. (Edwin)
Carlos noted that if pentane-2,4-dione was the most acidic, it must have the most stable conjugate base due to increased stabilisation of the base from the resonance structures available. Explanations of this kind are in keeping with the scientific principles presented in the course curriculum. Edwin had a similar, but more fragmented understanding of resonance, suggesting that the acid (not the conjugate base) was stabilized by resonance.
However, students performed worse on items 2 and 3 (where students were asked to rank the relative acidities of pentane-2,4-dione and phenol) and justify their response following instruction. When asked to explain their reasoning, the following ideas were present in the comments made by five of the twelve interviewed students:
So in terms of looking at [pentane-2,4-dione] and [phenol], cause that's what you have to compare, [phenol] has a benzene ring as opposed to [pentane-2,4-dione] so it would be able to stabilise its own charge more and have more resonance structures, as opposed to what I tried drawing for [pentane-2,4-dione]. So I felt like [phenol] had more resonance structures and it was more stable and more likely to act like an acid Instead of [pentane-2,4-dione]. (Ralph)
So if you were to take off the hydrogen of the alcohol group, the negative charge of [phenol] is better distributed throughout the benzene ring than it could be throughout whatever those are- carbonyl groups. So there's more resonance structures, yeah. I was like [expletive] certain on that one. (Libby)
Students seemed to rely on the quantity of resonance structures as a determinant of acid strength to explain the significant drop in scores on two-tiered items 2 and 3. Unlike the outcome in McClary and Bretz's study, many of our students did not exclusively associate the benzene ring of phenol with resonance. Five of the twelve students recognised that resonance structures were available to both compounds. The conjugate base of pentane-2,4-dione has three contributing resonance structures, where the negative charge of the conjugate base is distributed over the alpha carbon and each of the carbonyl oxygens. Phenolate has four contributing resonance structures, where the negative charge of the phenoxide anion is distributed over the oxygen and the carbons on in the aromatic ring. Students generally recognised the contributing resonance structures for the conjugate bases of both compounds, but prioritised the quantity of available resonance structures over their stability (i.e., which atoms would bear the delocalised charge). These nuances of resonance stabilisation were not explicitly addressed in our course, but interview data revealed ways in which instruction might be improved to address this alternative conception in future iterations.
Table 3 summarises the most significant distractor responses (item choice) and alternative conceptions (italics) that were identified from interviews and quantified from concept inventory scores and confidence intervals. For several items, a significant alternative conception was identified even after instruction. Our findings indicate that students generally understood the impact of induction on the relative stability of conjugate bases better than that of resonance when determining acid strength. The most prevalent alternative conceptions held (i.e., impact of quantity vs. stability on resonance stabilisation, misidentifying the most acidic proton in acetaldehyde) were often not explained by simple heuristics unlike those reported in the previous study of this assessment (McClary and Bretz, 2012). This suggests that instruction may have played a role in helping students think more critically about these topics.
Item | Specific alternative conception | Frequency/strength | |
---|---|---|---|
Pre-test | Post-test | ||
1 | Acetic acid is more acidic than pentane-2,4-dione and phenol because it is a carboxylic acid. Functional groups primarily determine relative acidity. | 47.7% (genuine) | 61.2% (genuine) |
2 | Phenol is more acidic than pentane-2,4-dione. The quantity of available resonance structures determines relative acidity. | 58.9% (spurious) | 84.6% (genuine) |
3 | Phenol is more acidic than pentane-2,4-dione because benzene better stabilizes the conjugate base than the carbonyl groups of pentane-2,3-dione. The quantity of available resonance structures determines relative acidity. | 48.1% (spurious) | 76.6% (genuine) |
6 | Acetaldehyde is more acidic than acetone because acetaldehyde has a hydrogen atom instead of another methyl group. The aldehyde hydrogen is the acidic hydrogen and it more strongly experiences induction from the carbonyl than the methyl hydrogen of acetone. | 52.8% (spurious) | 57.5% (genuine) |
“I think the place where I really learned where to find right and wrong in terms of answers is probably [the small-group] workshop. Because I know there were a lot of times when we were doing quizzes, me and my group, we would have an answer and have the best explanation for it but then somebody else would say another an answer and give another explanation for it and it kind of showed us a different way of thinking of it.” (Carlos)
“I would say, like, through lecture and during the workshop itself. Because my two group members usually disagreed with each other. So they had different reasoning.” (Sofia)
Students seemed to appreciate the role of workshop in exposing them to different modes of reasoning that arose from discussions with their group members. We therefore sought to investigate what role the quality of student arguments in the cooperative learning section of the course may have played in forming or altering their related conceptions (Boller et al., 1990; Osborne et al., 2004; Sandoval and Millwood, 2005). Fig. 2 shows post-test item analysis for students in each of the recorded groups. While the overall concept inventory data resembles that of the entire population of students, we noticed several interesting variations between the responses across groups. First, it was worth noting that of the fewer than 10% of students (16 students out of 217) responded correctly to both items two and three of the inventory. Intriguingly, two of those students happened to be in the same workshop group (Ron and Jim, Group 2) and reported confidence levels above 50% (darker green shading). Additionally, two of the students in Group 1 (Jay and Ralph) responded incorrectly to both of these items, but were among the most confident in their responses relative to students in the other groups (darker red shading). Additionally, Amira seemed much less confident than the others in her group, as indicated by the relatively lighter shading across her inventory responses. Overall, item analysis of student responses across groups led us to more closely examine the arguments of Group 1 and Group 2 to assess whether the quality of their arguments may have impacted their conceptions on these particular inventory items.
To investigate the differences between these groups, we analysed recorded student–student conversations from the cooperative learning portion of the course. We sought to assess the quality of student arguments and whether students appealed to the course content to support their assertions. Toulmin's (1958) argumentation scheme was used to code discussions. As previously mentioned, the analytical framework identifies the basis of an argument to include claims supported by data, which can be linked through warrants. Rebuttals are made to challenge assertions with opposing claims (see Appendix Table 5). Erduran's et al. (2004) established rubric for rating the level of arguments based on aspects of Toulmin's model was then employed to assess the quality of argumentation for each item discussed. Two representative conversations coded using Toulmin's model and the prompt eliciting these discussions are shown in Fig. 3. While the prompt requires students to determine relative basicity instead of acidity, students were expected to evaluate similar structural factors (i.e., induction by neighbouring functional groups, available resonance structures, sterics) when making their predictions.
Fig. 3 Representative conversations from two recorded student groups in response the indicated prompt. Scientific arguments were coded using Toulmin's model of argumentation (red). |
In Group 1, Ralph, Jay, and Amira (who is present, but does not participate in the conversation) have a discussion led primarily by Jay around the basicity of the nitrogen atoms shown in Fig. 3. The conversation begins with Jay and Ralph co-constructing an incorrect claim about relative basicity (“to act as a base it needs to be able to accept electrons”). Despite agreeing with this incorrect claim, Jay proceeds to think out loud about the impact of induction and resonance on the basicity of each nitrogen in the molecule. He claims that N1 cannot be the most basic due to the inductive effect of the nearby carbonyl (he does not mention the additional impact of resonance here). He recognises the influence of resonance structures on basicity, but has a flawed understanding of resonance in his claim (“not really near any resonance structures”). He uses induction to decide on N2 due to the presence of “electron donating” alkyl groups. Overall, the conversation is dominated by Jay who relies primarily on claims and data to make his case. The lack of warrants and rebuttals from the other group members is indicative of a Level 2 argument (Erduran et al., 2004), and may suggest that the chain of reasoning from claims to data is faulty or only partially understood.
In Group 2, Sara, Ron, and Jim have an exchange where all three students participate in a conversation that briefly focuses on inductive effects before shifting to a discussion about resonance (Fig. 3). Sara claims that “you don’t want to have the inductive effect” in her argument for why N1 is not the most basic nitrogen, a claim that is left unchallenged by her groupmates and does not include the additional impact of resonance. In addition to making many claims, the students offer several pieces of data linked to their claims by warrants. A rebuttal is also made when Sara and Ron quickly correct Jim about the protonation state of the conjugate acid. Unlike the first group, Ron, Sara, and Jim take the time to discuss what the resonance structures of the conjugate might look like. They attempt to work through the specifics of “moving around electrons” and the consequences of doing so on the charge of the atoms. However, they ultimately ask for assistance from their TA when they are unable to rationalise the effect of resonance in their discussion. Though the students were unable to arrive at a conclusion before seeking assistance, they co-constructed a Level 5 argument, which included claims made by each student, several uses of data and warrants, and two rebuttals.
Across recorded groups, arguments about the impact of induction were minimal, though students seemed to understand inductive effects reasonably well (Fig. 1). Students spent most of their time reasoning through resonance effects, possibly because this topic is more nuanced (McClary and Talanquer, 2011b) and less familiar. While neither group was able to demonstrate a robust understanding of resonance effects on base strength, Ron and Jim (Group 2) seemed fairly confident in their correct responses to aligned inventory items two and three, perhaps owing to the more substantive arguments in their group about the related prompt. These higher quality arguments may have provided the cognitive conflict necessary for them to recognise their incomplete conceptions of resonance effects and further examine the topic outside of the workshop setting.
Students' concept inventory scores even after instruction averaged less than 50%, though their grades on course exams averaged 80–85%. These data suggest some areas of improvement of our instructional approach. Particularly, it was clear that the impact of resonance on relative acidity/basicity was not thoroughly addressed or assessed in our course. While an understanding of inductive effects on relative acidity may be achieved through more simplistic evaluations of structural features, the nuances of resonance stabilization are difficult to evaluate without an ability to explicitly identify contributing resonance structures. It is possible that students found it difficult to draw individual resonance forms that would have more accurately informed their choices. Therefore, efforts to more intentionally promote the development of this critical skill might improve student outcomes in future iterations of the course.
The workshop prompts themselves (Fig. 3) seemed to be useful for promoting student argumentation about the topic, though some groups engaged in higher quality argumentation than others. Strategies for improving student engagement and confidence in these settings (for students like Amira) may be worth considering in future implementations (Michael, 2006; Hermann, 2013). While cooperative learning settings did seem to encourage quality argumentation among some students, formal instruction on the core components of a strong argument may have improved their observed quality.
Additionally, this work has implications for the broader chemistry education community. Comparing individual item responses to the quality of student arguments suggests that these conversations may have played a role in encouraging students to move beyond heuristic reasoning and think more deeply about these topics. Future investigations of the specific aspects of student discussion that may have promoted or disrupted their conceptual understanding may be helpful toward improving the effects of cooperative learning on student performance. Examining the impact of various facilitation strategies on the quality of student conversations and performance may also yield insights about how facilitators may impact student learning. An evaluation of this discussion-oriented curriculum on other challenging topics (e.g., redox chemistry, kinetics) may help more thoroughly establish its success and areas of improvement across content areas. Finally, implementing curricula and instructional strategies that improve student reasoning and performance may reduce attrition rates in chemistry and motivate students to pursue more advanced coursework in the field.
Analytical framework used for assessing the quality of argumentation | |
---|---|
Level 1 | Level 1 argumentation consists of arguments that are simple claim versus a counter-claim or a claim versus a claim. |
Level 2 | Level 2 argumentation has arguments consisting of a claim versus a claim with either data, warrants, or backings but do not contain any rebuttals. |
Level 3 | Level 3 argumentation has arguments with a series of claims or counter-claims with either data, warrants, or backings with the occasional weak rebuttal. |
Level 4 | Level 4 argumentation shows arguments with a claim with a clearly identifiable rebuttal. Such an argument may have several claims and counter-claims. |
Level 5 | Level 5 argumentation displays and extended argument with more than one rebuttal. |
Code | Description | Example |
---|---|---|
Claim | A chemical assertion that is made by at least one student | S: Yeah so we’re looking at pKb's and we know that guanine is the weakest and cytosine is the strongest. |
Data | A statement in which a student appeals to reliable, chemical evidence (e.g., discussion worksheets, lecture notes, periodic table, statements made by facilitators). | S1: It has no oxygen to pull away and destabilize… |
S2: And it also has two methyl groups surrounding it. | ||
Warrant | A statement made by at least one student which directly links a stated claim to supporting data. | S3:…Wait which one is bigger, oxygen or sulfur? |
S2: They’re like the same group. Let's look up the periodic table…So sulfur is bigger. | ||
S3: So sulfur is larger and sulfur should be better at stabilizing. | ||
Rebuttal | A statement of opposition that is made by at least one student to a previously stated claim made by another student. | S2: Yeah so. Let's say the proton to the proximity to the hydroxyl… |
S1: (interrupts) Hydroxide. | ||
S2: No. Hydroxide is the ion. Hydroxyl is the functional group. |
(1) Read the question and explain what you thought it was asking.
(2) Explain why you chose the answer that you did.
(3) Explain why the remaining choices are incorrect.
For the inventory overall:
(4) Were there any words in the questions/choices that you did not understand?
(5) Where there any specific aspects of the course that helped you come to a different understanding from pre-test to post-test?
Additional group conversations around prompt in Fig. 3
Group 3 (Level 4 argument)
Barry: (reading the prompt) Identify the most basic…
Ankit: What's basic?
Barry: Trying to accept the hydrogen [claim]
Carlos: That means everything around it has to be really electronegative [claim]
Barry: Wait, why?
Carlos: Because if this is really electronegative that means it will attract the hydrogen [warrant]
Barry: Oh ok.
Carlos: Yeah.
Barry: Weren’t we just looking for most electronegative for acids and now we’re looking for the same thing for base…[claim]
Ankit: well I understand what you’re saying but then at the same time, the electronegative atoms would pull the electrons away [rebuttal].
Carlos: When acids like the…one side's positive and the other side…cuz [sic] the electrons moving toward the electronegative part, it becomes really negative so… [warrant] I think this one, you just have to find the most electronegative place…[claim] like for this it's here, (points to N1) here (points to N2), and there (points to N3) [data].
Group 4 (Level 4 argument)
Edwin: (reads prompt)…basic because they contain nitrogen? Wait why?
Samantha: It's because there are a lone pair of electrons [data] and that makes them partially negative [warrant].
Edwin: More partially negative? Is that correct? (shows paper)
Samantha: Yeah. Here it is [data].
George: This is the (audio overlaps). Well, they (TAs) were talking about like…uh, like…um… how the electron donating groups [data].
Samantha: So that means that this one would probably be basic [claim]and I’m gonna go with this one because that one has carbons on both sides but this one only has the methyl group on one end and also it has that thing (points to screen) [data]
George: So wouldn’t it be this one? (points to screen) [rebuttal]. Because this one has the electron withdrawing? [data]
Samantha: But we’re looking at the weakest acid. [rebuttal]
George: Oh the weakest that's right.
Samantha: So I think it's that one. [claim]
George: So yeah. Yeah yeah.
This journal is © The Royal Society of Chemistry 2018 |