Analysis of resources applied to rationalize elimination mechanisms

Abstract

This study aimed to analyze second-semester organic chemistry students’ problem-solving strategies, specifically focusing on the resources activated while solving problems on E2, E1, and E1cB elimination reactions. Using the theoretical framework by Elby and Hammer, we defined a resource as a unit of information used in the problem-solving process. The resources activated to solve elimination reaction problems were probed using a mixed-methods approach using survey assessments and think-aloud interviews. The data were analyzed quantitatively and qualitatively following a validated set of scoring criteria. The results align with existing findings that students focus on surface-level structural information and use resources that have been repetitively emphasized over multiple semesters. Resources related to acid–base chemistry were activated more often than reaction-specific resources, such as conformational analyses or carbocation rearrangements. Although acid–base resources aid students in successfully analyzing reaction mechanisms, additional resources must be activated to rationalize specific mechanisms and to explain the products formed. This calls for instructors to provide formative and summative assessments that evaluate the many resources required to elucidate elimination reaction mechanisms and product stereochemistry.

---

Introduction

Reaction mechanisms are cited as one of the most challenging concepts in organic chemistry. Because of this, there has been an influx of research studying how students reason through reaction mechanisms (Caspari et al., 2018; Dood and Watts, 2022, 2023; Watts et al., 2022; Frost et al., 2023; Kranz et al., 2023). The chemical education literature describes two primary representations of organic chemistry mechanisms: the electron-pushing formalism (EPF) (Morrison and Boyd, 1959; Bhattacharyya, 2013) and reaction coordinate diagrams (RCD) (Lamichhane et al., 2018; Popova and Bretz, 2018a, 2018b, 2018c; Atkinson et al., 2020). Watts et al. recently studied how organic chemistry students use the EPF and RCD together to determine the most probable of two reaction mechanisms (Watts et al., 2022). Together, aspects of these representations are used by students to form their mechanistic reasoning, which is the rationale for why specific bonds are broken or formed, the observed regioselectivity and stereoselectivity, and the major and minor product distribution.

Acid–base chemistry in organic chemistry problem solving

Acid–base chemistry is a key concept employed in mechanistic reasoning. Acid–base chemistry is consistently ranked by organic chemistry instructors as one of the most important concepts from general chemistry (Duis, 2011; Nedungadi and Brown, 2021) because acid–base concepts are central to developing qualitative models to explain organic reactivity. Such qualitative models include rationalizing leaving group ability, describing and comparing nucleophilicity, and predicting competition between substitution and elimination. Organic chemistry draws predominantly upon qualitative problem-solving skills, whereas quantitative problem-solving skills are typically emphasized more in general chemistry. The shift from quantitative to qualitative problem-solving skills has been attributed to the long-standing sentiment that organic chemistry is the gatekeeper course for students (Garg, 2019; Arnaud, 2020). In the past decade, the number of organic chemistry education research articles has significantly grown, with mechanistic problem-solving being a central focus, as well as the role of acid–base concepts in rationalizing organic reactions (Graulich, 2015; Schmidt-McCormack et al., 2019; Petterson et al., 2020; Salame et al., 2020; Gao et al., 2022; Yue and Gu, 2022; Boothe et al., 2023). Two studies analyzed expert responses to gauge the significance of acid–base chemistry in organic reactions. A set of twenty-eight reactions routinely taught in organic chemistry or biochemistry courses was analyzed by Stoyanovich, Gandhi, and Flynn, with 85% of the reactions requiring Brønsted–Lowry acid–base chemistry in at least one step in the mechanism (Stoyanovich et al., 2015). Boothe, Zotos, and Shultz interviewed a cohort of organic chemistry instructors who universally cited acid–base recognition as essential for predicting organic reactivity (Boothe et al., 2023). Finally, the EPF (Morrison and Boyd, 1959) is generally introduced in the context of acid–base reactions and later applied to electrophiles and nucleophiles (Bhattacharyya, 2013).

Research centered on student problem-solving in organic chemistry has provided vital insight into how students rationalize acidity and basicity in the context of organic molecules. Cruz-Ramírez de Arellano and Towns highlighted student difficulty with the identification of acids, bases, nucleophiles, and electrophiles (Cruz-Ramírez de Arellano and Towns, 2014). Other studies have found areas in which organic chemistry students commonly struggle, such as predicting pK_a values of organic molecules, incorrectly equating certain functional groups with high acidity, and misinterpreting the relationship between acidity and conjugate base stability (Bretz and McClary, 2015; Flynn and Amellal, 2016).

Although Brønsted–Lowry acid–base reactions prove to be highly important in organic chemistry and biochemistry, Lewis acid–base recognition remains a vital skill for successful mechanistic predictions because of the parallels with electrophiles and nucleophiles. Despite the ubiquitous nature of Lewis acid–base reactions in organic chemistry, students routinely gravitate towards the Brønsted–Lowry definition for acids and bases, even if this definition is less relevant to the task at hand (Schmidt-McCormack et al., 2019; Petterson et al., 2020; Watts et al., 2020). In a study by David Cartrette and Provi Mayo, students were prompted to define an acid (Cartrette and Mayo, 2011). Every study participant provided the Brønsted–Lowry definition but completely neglected the Lewis acid definition (Cartrette and Mayo, 2011). Another study found that students utilizing Lewis models to rationalize acid–base reactions were three times more likely to provide correct reaction mechanisms compared to students using the Brønsted–Lowry model or a generic description (Cooper et al., 2016; Dood et al., 2018). This finding is of particular interest to us, as reaction mechanisms tend to present large conceptual hurdles for organic chemistry students. In this research, we aimed to analyze students’ problem-solving strategies for elimination reactions – specifically, we sought to compare students’ use of acid–base concepts for mechanistic reasoning.

Student understanding of organic reaction mechanisms

While these earlier studies underscore the importance of acid–base chemistry, there have also been several detailed studies on students’ understanding of reaction mechanisms. Finkenstaedt-Quinn et al. studied student conceptions about the mechanism of addition reactions (Finkenstaedt-Quinn et al., 2020). Through their interviews, they reported students fail to apply conceptual knowledge in their explanations and choose to focus on explicit structural features, such as carbocation substitution or atomic charges (Finkenstaedt-Quinn et al., 2020). This supports a plethora of other literature that shows student reliance on surface-level molecular features in solving reaction mechanisms (Anzovino and Lowery Bretz, 2015; Anzovino and Bretz, 2016; Galloway et al., 2017; Caspari et al., 2018; Popova and Bretz, 2018b; Finkenstaedt-Quinn et al., 2020; Petterson et al., 2020; Watts et al., 2020). The Towns’ research group found that incorrect explanations often involved memorized examples or rules that were regarded as generalizable to other “similar” examples (Cruz-Ramírez de Arellano and Towns, 2014). Often, these “similar” examples were not as closely related as perceived by the student. These studies support existing literature that shows the challenges that students face in organic chemistry (Henderleiter et al., 2001; Graulich, 2015; Grove and Lowery Bretz, 2012; Salame et al., 2020, 2022).

Theoretical framework

The research was modeled using a resources framework (Hammer et al., 2005), which defines a resource as the smallest unit or bit of information utilized in the problem-solving process.(Kranz et al., 2023) In order to better conceptualize a resource, we are expanding the definition in accordance with an article by Wittmann (2018) and Parobek et al. (2021) that defines a resource as an idea students use when solving problems. Both conceptual and epistemological resources, observed in problem-solving, are activated based on how problems are “framed” by students. When students “frame” problems, they are contextualizing the scope of the problem: What is the problem asking me to do? With domain-specific problem-solving, different resources will be activated for novice students. As students gain experience with specific types of problems, a more defined set of resources are activated (Hammer et al., 2005). The activated resources provide a full overview of the ideas incorporated into the problem-solving process. This idea is counter to the conceptions framework, (Taber, 2000) which defines misconceptions as ideas that are derived from “pre-compiled” information and are incorrect. With the resources framework, answers are organized from resources that are compiled together and are neither right nor wrong. Instead, resources are productive or unproductive in accomplishing a specific objective. The resources framework more closely aligns with current paradigms that focus upon pedagogy that promotes students to recombine ideas and recalibrate strategies in lieu of overcoming misconceptions.

Elucidating elimination reaction mechanisms with acid–base chemistry resources

To build upon these findings, we have adapted one of Towns’ interview questions to specifically gauge student mechanistic understanding as it relates to acid–base chemistry. In the Towns’ interview, students were given the reactants in Fig. 1A and were asked to predict the structure and mechanism of the major product (Cruz-Ramírez de Arellano and Towns, 2014). In contrast, our study provided both the reactants and a plausible product, asking students to identify the reaction pathway, explain their reasoning, and rank their confidence in their responses. Acid–base concepts are emphasized extensively (Clayden et al., 2012; Loudon and Parise, 2021) when discussing elimination reactions. Given the repeated emphasis, we expect students to frequently draw upon concepts and resources related to acid–base chemistry.

Fig. 1 Reaction schemes and questions asked on the survey assessment for the (A) E2, (B) E1, and (C) E1cB eliminations.

While E2, E1, and E1cB are all elimination reactions, the framing of problems differs with respect to the substrate, leaving group, and base strength. Acid–base concepts link each of the three elimination reactions, and as posited by Hammer (Hammer et al., 2005), repeated activation of a set of resources in a given context is likely to lead to activation again in similar contexts, such as classifying a reaction as elimination. While acid–base chemistry shares a common thread across elimination reactions, different resources, such as conformations and competing mechanistic pathways, are required to explain why specific products are formed over others. The research outlined will focus on the cognitive resources, particularly resources related to acid–base chemistry, that students apply in three types of elimination problems and will consider what other resources are utilized to explain and rationalize the reactivity. To differentiate between acid–base concepts and acid–base resources, we are defining a resource as a specific idea whereas a concept embodies several resources or ideas.

Purpose of research

We aim to understand the cognitive resources employed by students in complex organic chemistry problems, specifically elimination reactions. To address this gap in the literature, we asked second-semester organic chemistry (OC2) students to explain their mechanistic reasoning for E2, E1, and E1cB reactions, all of which rely on the acid–base properties of their reactants. In this study, we defined mechanistic reasoning as the chemical rationale behind electron-pushing mechanisms. This study was guided by the following research questions developed using criteria by Bunce (Bunce, 2008):

1. What cognitive resources are activated while students are analyzing elimination reactions?

2. What themes arise when students connect acid–base chemistry to elimination reactions?

3. How do students gauge their own understanding of elimination reactions?

Methods

Course structure

At the institution where this study was conducted, general and organic chemistry courses have three instructional modes per course that include 2.5 hours of lecture, 50 minutes of teaching assistant (TA)-led discussion/recitation, and 3 hours of TA-led laboratory per week. The content covered in the organic chemistry sequence is shown in Scheme 1, and every topic includes some discussion of acid–base chemistry. The general chemistry course, taken before organic chemistry, spends three to four weeks discussing acid–base structure and calculations. Therefore, students call upon acid–base resources more often in the three-semester sequence than conformational resources, carbocation rearrangements, and more nuanced concepts associated with specific mechanisms.

Scheme 1 Flowchart depicting the content covered in first (top arrow) and second semester (bottom arrow) organic chemistry courses at the institution where this study was conducted. The yellow shading indicates topics that use both conformational and acid–base resources, and the orange shading indicates topics that draw predominantly on acid–base resources.

The organic chemistry courses have weekly homework, discussion activities, and in-class activities designed to provide formative assessments of specific concepts. Online homework that provides immediate feedback is also used. In-class activities are used to gauge student performance and are discussed during lectures with specific attention addressing why particular answers are incorrect. Problems are explicitly chosen that illustrate the nuanced ideas, including conformations and rearrangements. Feedback is provided to outline the significance of each resource in solving problems. The TAs outline solutions to problem sets covered during discussion sections, and answer keys are posted. The formative assessments are designed to provide students with information regarding their understanding and feedback to promote improvement. Therefore, the scope of resources needed to effectively solve problems is illustrated with the sets of problems provided by formative assessments. Two to three midterm exams are given during the semester as summative assessments, followed by a cumulative final exam. The midterm exam solutions are reviewed in discussion sections, and answer keys are posted. Problems using an array of resources are incorporated into summative assessments – therefore, some problems may only require acid–base resources, while others require several resources to solve effectively.

Student participants and data collection

All studies were performed at a private midsized research-intensive (R1) university. Data were collected using both online survey assessments and think-aloud interviews. The data were handled in accordance with the Institutional Review Board (Protocols #2021-0182, #2022-0204). Students were informed of their rights as human research subjects, and those who did not consent to their responses being used for this study were omitted from data analysis. Only students residing in the United States and over 18 years of age were included in the study. When identifiers were collected, the data were anonymized and not shown to instructors until after final grades were posted.

The online survey assessments were distributed to three cohorts of OC2 students from different instructional terms taking the course with different professors. Due to COVID-19 restrictions, the three cohorts of students experienced the course through different instruction media. Cohort A (N = 65) from the 2020 Fall semester experienced the course completely remotely, and the survey assessment was distributed as a for-credit assignment. Cohort B (N = 30) during the 2021 Summer semester was a hybrid course, with some students completely in-person, whereas others were completely remote; the survey assessment was distributed as an encouraged review assignment for the final exam. Cohort C (N = 35) during the 2022 Summer semester was in-person; the assessment was distributed as a required for-credit assignment. The response rates were 98%, 50%, and 76%, respectively. The purpose of the survey assessments was to identify the resources that students immediately employed to rationalize their explanations.

The 45 minute-long think-aloud interviews were conducted over Zoom, and the problems were presented by the interviewer via the screen-sharing function. The ten interviewees completed the interview on a voluntary basis and were not screened or filtered prior to participation. All interviewees had previously completed the survey assessment within two weeks of participation. Though the questions remained the same, interviews were used to further probe which resources students utilized or could be prompted to utilize to rationalize their answers. The interviewer filled out the student participants’ answers to multiple-choice questions, then referred to the recording and transcript for their explanations. The student participants kept their video off and the microphone on for the recordings. Further analysis of data is described under qualitative data analysis.

Development of survey and interview questions

The E2 question was adapted from a previous study and was included as a calibration tool because student reasoning about it has already been investigated extensively (Fig. 1A) (Cruz-Ramírez de Arellano and Towns, 2014). The previous study did not provide the product of the reaction; rather, it asked the students to predict the product, thus investigating students’ ability to compare S_N1, S_N2, E1, and E2 reactions. Our adaptation of the question prompted students to classify methoxide as a Brønsted acid, Brønsted base, nucleophile, or electrophile. Subsequently, students were asked to classify the reaction as E2, E1, S_N2, or S_N1 and to justify their answers. Students were broadly prompted to “justify their answer” to identify the resources activated to explain their reasoning. The E1 question was designed to complement the E2 question, utilizing the same starting material but forming a different product to probe whether students could rationalize why each product was formed (Fig. 1B). Like the E2 question, students were given an open-ended prompt to justify their reasoning. These questions were like examples emphasized in class.

The E1cB question was designed because, at the time of our survey design, no existing chemical education research had investigated the relationship between acid–base chemistry and the energetics of the aldol condensation (Fig. 1C). Since the administration of our survey, the Watts et al. published another test instrument that could be used to probe student understanding of reaction energetics in the aldol reaction, although their questions did not extend to the E1cB condensation step of the aldol reaction (Watts et al., 2022). To start, students were asked about the kinetics and thermodynamics of the aldol condensation between acetophenone and benzaldehyde. The question about the rate-determining step is framed so that students only need to rationalize experimental findings because this reaction is still an area of ongoing research (Mak et al., 2007; Perrin and Chang, 2016; Berardi et al., 2021; Coutinho et al., 2021). Specific molecules and reaction conditions were taken from Perrin and Chang's study, and the questions and premises provided are consistent with the authors’ Gibbs free energy diagram and rationalizations. Student responses to questions 2a and 2b (Fig. 1C) are not included but the questions are shown here to provide transparency to potential framing effects. We hypothesized that even if students had not explicitly covered the E1cB reaction, they could still rationalize their way through the reaction using principles of acid–base chemistry and comparisons to the other two elimination reactions in our survey assessments.

Quantitative data analysis

Survey assessment data were first tested for normality. Calculations of skewness and kurtosis for student confidence for the E2, E1, and E1cB questions suggested the data were not normally distributed. The Shapiro-Wilk test supported this (p < 0.01), so nonparametric tests were employed for all statistical comparisons.

The survey assessment data from the three OC2 sections were also compared to investigate statistically significant differences in accuracy and confidence. Two-tailed Fisher's Exact tests suggested there were no statistically significant differences in accuracy between the three cohorts of students. Mann–Whitney U tests revealed a statistically significant difference in confidence for one of the cohorts, which will be discussed in more detail later. Overall, because there were minimal statistically significant differences in accuracy and confidence between the three cohorts, the survey assessment results were aggregated.

The aggregated confidence data were then compared based on ranking accuracy and explanation accuracy. One-way Kruskal–Wallis tests (McKight and Najab, 2010) were used to make multiple group comparisons, and Mann–Whitney U tests were used to make pairwise comparisons when significant differences were found. A p-value of 0.05 or less was considered statistically significant for all tests performed. All visualizations and statistical calculations discussed above were performed in R (R Core Team, 2020).

Qualitative data analysis

To supplement our quantitative data, qualitative data was collected via think-aloud interviews. Unlike survey assessments, interviews allow for follow-up questions and clarifications, which can further characterize how students use cognitive resources (Bretz, 2008). Four researchers identified acid–base scoring criteria by consensus through discussion and used these criteria to analyze each survey assessment/interview question (Table 1). The codes were expanded for the interviews to account for the wider array of resources or ideas activated while students were reasoning through the mechanisms. The same codes were used initially to provide benchmarking comparisons. These grading criteria are reflective of the conceptual emphasis and grading schemes in OC1 classes at the research institution. Professors who routinely teach organic chemistry were consulted and provided validation for the coding scheme. Together, the same researchers classified each student's explanations as demonstrating complete, partially complete, or incomplete explanations. Furthermore, the explanations were coded both deductively and inductively; deductive codes were taken from the work of Cruz-Ramírez de Arellano and Towns for the E2 question, whereas the inductive codes came directly from the students’ survey assessment responses. Three researchers independently reviewed the transcripts and audio/visual data from the interviews. All three researchers performed an initial classification of responses as complete, partial, or incomplete, achieving over 80% agreement. The research team then reconvened to discuss their findings to reach a consensus on the final coding results. Representative interview transcripts have been included for further analysis.

Table 1 Scoring criteria for survey and interview questions

Question	Comment	Criteria
a Productive and unproductive were used to defined how the specific acid–base resources were used. Productive was used to define scenarios in which students recognized that the species will deprotonate the organic substrate.
E2 MC	Productive^a	E2 and methoxide acts as a Brønsted base
	Unproductive	Not E2 and Brønsted base
E2 explanation	Complete	Methoxide is a strong base and facilitates E2
		Proton removed is antiperiplanar to Br
	Partial	Methoxide is a strong base and facilitates E2
	Incomplete	Does not discuss either concept
		Incorrect mechanism/chemistry
E1 MC	Productive	E1 and methanol acts as a Brønsted base
	Unproductive	Not E1 and Brønsted base
E1 explanation	Complete	Methanol is a weak base and facilitates E1
		Hydride shift to form tertiary carbocation
	Partial	Methanol is a weak base/lack of a strong base
	Incomplete	Does not discuss either concept
		Incorrect mechanism/chemistry
E1cB explanation	Complete	Poor leaving group (hydroxide)
		Acidic alpha proton
	Partial	Discusses one of the two concepts
	Incomplete	Does not discuss either concept
		Incorrect mechanism/chemistry

---

For the interview data, the resources (or ideas) utilized were further categorized based upon complexity. Resources were categorized as recall, structural, or mechanistic resources. Recall resources include facts that students can memorize to compartmentalize chemical reactions. Structural resources are conclusions that students can draw from analyzing the chemical structure. Mechanistic resources are conclusions that students can only draw from manipulating the structure. We included transition states and electron pushing formalisms as mechanistic resources. These resources are arranged in order of increasing complexity as shown in Table 2.

Table 2 Categorization of resources recorded during interviews

Type of resource	Examples
Recall	A double bond is formed in an elimination reaction.
	An E2 reaction requires a strong base.
	Methoxide is a strong base.
	Methanol is a weak base.
	More substituted alkenes are more thermodynamically favored.
	Elimination reactions require a good leaving group.
Structural	The extent of alkyl substitution on the alkene.
	The type of carbocation that would form.
	There will be increased steric hinderance by the bulky group.
Mechanistic	The more substituted product does form because the needed anti-periplanar conformation is not possible for the structure.
	A hydride shift occurs resulting in a more stable tertiary carbocation.

---

Results and discussion

The goal of this study was to assess students’ explanations of elimination reactions and their relationship to acid–base chemistry at the end of OC2. E2, E1, and E1cB elimination pathways were used to characterize students’ proficiency in both acid–base and mechanistic concepts in the context of foundational organic chemistry reactions. An E2 reaction is a concerted elimination that occurs in the presence of a strong base and generally proceeds through an antiperiplanar transition state (Fig. 5) (Clayden et al., 2012). An E1 reaction is a stepwise elimination in which the leaving group is lost in a rate-determining ionization step, allowing for carbocation rearrangements before a proton is removed by a weak base solvent (Fig. 6) (Loudon and Parise, 2021). Lastly, an E1cB is a stepwise elimination that largely applies to substrates with acidic protons and poor leaving groups (Clayden et al., 2012). In general, the survey assessments were used to identify the concepts that first come to students’ minds, whereas the think-aloud interviews were used to probe deeper into their thought processes about the connections between acid–base chemistry and organic reactions.

Q1: E2 elimination

This question mirrors the types of questions utilized for in-class assignments and assessments. In providing a chemical problem like that in the students’ curriculum, we hoped to gauge the intuitive rationale that students provided for their responses. Most students (58%) answered the E2 survey assessment questions correctly. However, fewer students provided complete explanations as determined by grading criteria (Table 1) that were highly emphasized in the course. The E2 elimination is introduced in OC1 as a cornerstone reaction and is widely applied in retrosynthesis planning problems (Loudon and Parise, 2021). Since the concept was introduced in OC1 and reiterated in OC2, we hypothesized that more students would consider all necessary resources when rationalizing the mechanism. We recognize, however, that survey responses may not encompass a respondent's complete understanding of the underlying chemistry. Because we asked students to identify sodium methoxide as a base, students were primed to activate cognitive resources related to acid–base chemistry, likely skewing responses towards acid–base concepts. Acid–base ideas alone, however, cannot fully justify the reaction pathway.

We also asked students to rate their confidence in their responses to explore how students gauge their understanding of elimination reactions. The Dunning–Kruger effect has been observed in introductory chemistry students, in which low-performing students tended to overestimate their abilities while high-performing students tended to underestimate their abilities (Bell and Volckmann, 2011; Pazicni and Bauer, 2014). We observed a statistically significant difference in the self-reported confidence between students who answered the E2 multiple-choice question correctly and incorrectly, where students who answered correctly were more confident than those who answered incorrectly (Fig. 2A). The E2 explanation question paralleled this finding, with students who provided a complete response having higher self-reported confidence than those who provided partial or incomplete explanations (Fig. 2B). In contrast to previous studies that observed the Dunning–Kruger effect when comparing results from metacognitive assessments and student performance on exams, we did not observe a clear over- or underestimation of abilities in our self-reported confidence data for the E2 question. Although within each grouping of “correct,” “incorrect,” etc., there were individual students who appeared to over- or underestimate their competence, on average, the self-reported confidence reflected student competence. The difference compared to previous studies could be attributed to the metrics we used in our study. In the aforementioned studies, the authors compared metacognitive measures of competence against exam scores. Here, we grouped the students based on their performance on each question and compared the confidence between the groups. An alternative explanation could be that these differences are due to the different cohorts of students involved in the studies. Previous studies investigated first- and second-semester general chemistry and participants were largely first-year college students who may not have had prior chemistry instruction. In contrast, our study focuses on second-semester organic chemistry. Since OC2 students have prior exposure to the E2 reaction in OC1 and continue to use E2 reactions for problem-solving, students may have a more precise gauge of when their answers are only partially complete or incomplete.

Fig. 2 Students’ self-reported confidence for the E2 question of the survey assessment, compared based on their performance on the (A) multiple-choice and (B) explanation portion. Productive vs. Unproductive: Wilcoxon p = 0.002, Wilcoxon effect size r = 0.25. Complete vs. Partial: p = 0.012, r = 0.34. Complete vs. Incomplete: p = 0.002, r = 0.30. Partial vs. Incomplete: p = 0.051, r = 0.15.

Student responses to the E2 survey assessment were deductively coded using the previously established criteria (Cruz-Ramírez de Arellano and Towns, 2014). The codes were flipped to avoid deficit framing and to better align with the resources framework. The data demonstrate that acid–base concepts figure more prominently than structural and mechanistic features for the E2 reaction (Table 3). Most students recognized methoxide's role as a base in the E2 reaction, even if they did not explicitly classify it as a “strong base.” Similarly, most students (81%) correctly classified the reaction as proceeding via an E2 mechanism, even if they did not explicitly state that a strong base is typically indicative of an E2 reaction. Only 5% of students discussed the antiperiplanar proton abstraction in their explanation, despite this relationship being emphasized in the lecture, graded assignments, and textbook (Loudon and Parise, 2021). This finding in our survey results is not entirely surprising, as students were primed to consider acid–base chemistry in their responses. Additionally, acid–base chemistry is ubiquitous throughout the undergraduate chemistry curriculum (van Kesteren et al., 2018), while conformations, which are essential to understanding the antiperiplanar relationship, are only introduced in organic chemistry (Scheme 1). The chemical education literature (Hartman et al., 2022) and our theoretical framework support the idea that students are more likely to utilize cognitive resources that have been repeatedly activated over time.

Table 3 Deductive codes for E2 explanation based on Cruz-Ramírez de Arellano et al.

Code description	Percentage of students (N = 130)
a Acid-base concept. b Mechanistic feature. c Structural feature.
Recognize methoxide is a base^a	74
Classify Br as a good leaving group^a	32
Recognize methoxide is a strong base^a	30
Recognize that strong bases proceed through E2^a	31
Understand the concerted mechanism^b	8
Understand the anti-periplanar proton abstraction^c	5
Recognize that SN2 is not viable because of steric hindrance^c	1

---

To complement the deductive codes, an inductive coding strategy was developed in which we generated codes based on the trends observed in the student responses to the survey assessment questions (Table 4). These results corroborate the codes established in the literature, although there were some discrepancies between our survey assessment data and the findings of Cruz-Ramírez de Arellano & Towns. Notably, they found that 14% of students discussed both the strong base characteristic of methoxide and the antiperiplanar relationship of the proton and the leaving group. Although this figure is higher than the 5% that we saw in survey assessment responses, this discrepancy can be attributed to differences in experimental design. Our survey assessment was designed to investigate which resources students prioritize in their short-answer responses. Furthermore, our survey assessment provided the product and asked students to identify the mechanism rather than asking students to both predict the product and reaction mechanism. Taken together, our survey assessment data complement the existing literature and suggest that for the E2 reaction, recognition of a strong base and good leaving group are the resources that OC2 students utilize first, while recognition of structural features and their impact on the reaction are resources that are utilized less frequently. Previous literature suggests that student usage of “good leaving group” in their explanations may not reflect sound chemical rationale but rather the students’ ability to memorize leaving group ability (Popova and Bretz, 2018d; Galloway et al., 2019; Dood et al., 2020). This is particularly relevant for our questions. As shown in Fig. 1A, students were asked to classify each reaction as S_N1, S_N2, E1, or E2, all of which require a good leaving group. Using leaving group ability as a core justification for this assessment does not reflect student differentiation between the four reaction types. The use of other resources such as base strength suggests a more sophisticated understanding, as base strength can be a key difference between E1 and E2 reactions. Nevertheless, based on Hammer and Elby, it is not surprising that so many students addressed leaving group ability in their answers, as leaving group ability is often emphasized in organic chemistry textbooks (Clayden et al., 2012; Loudon and Parise, 2021) specifically in the context of elimination reactions. Though highly emphasized, leaving group ability is not a relevant chemical rationale to differentiate S_N1, S_N2, E1, or E2 reactions and does not advance problems that require students to select one of the four mechanisms.

Table 4 Inductive codes for E2 explanation

Inductive code description	Percentage of students (N = 130)
Deprotonation occurs	34
Br is a good leaving group	32
Loss of leaving group and formation of double bond indicate elimination	31
Methoxide is a strong base	30
The carbocation mechanism is disfavored	5
Antiperiplanar proton	5

---

To supplement the survey assessments, we conducted think-aloud interviews to further probe the resources students employed in E2 reactions. An inherent limitation of our survey assessments is the inability to interact with the students and ask them to expand on their ideas. Students also likely ended their responses prematurely once they determined they had provided a sufficient explanation. To avoid overreliance on online survey assessments, we conducted ten think-aloud interviews using the same set of questions to probe deeper and ask students to expand on their explanations. Of the ten students interviewed, four of them provided a partially complete explanation of the acid–base chemistry involved in the E2 reaction. To provide additional support in framing the questions, the interviewer prompted students: “what might be useful is to consider the transition state, for each reaction.” Problem-solving strategies could be classified into three overarching themes: (1) structural resources with sterics arguments (2) structural resources with the extent of alkyl substitution, and (3) recall resources with base strength and E1 versus E2 pathways. One or all of these themes were observed in each of the 10 interviews.

To illustrate the steric theme, consider the following dialogue. The student correctly identifies the mechanism as E2 but did not extend to more complex mechanistic resources to account for the product formed.

“So, it's definitely elimination because it's forming an alkene. You’re eliminating the bromine and hydrogen. I chose E2 over E1, mainly because there is a strong base—the negative OCH ₃ —and E1 is typically more of a solvolysis.”

This student began by noting the change in structure between the reactant and product, then correctly identified that the strong base character of methoxide favored the E2 reaction. These resources can be categorized as structural and recall with the formation of the double bond being a structural resource and the recognition of the strong base as an important feature in E2 being a recall resource. When prompted to consider the regioselectivity of the E2 reaction compared to the E1, this student discussed the steric hindrance due to the isopropyl substituent rather than discussing the antiperiplanar relationship of the proton and leaving group:

“So, I would say because E2 is concerted, it has both kinetic and thermodynamic concertability [sic]. So, when the base deprotonates there's less steric hindrance for the hydrogen to the left of bromine rather than to the right because it doesn’t have all those methyl groups.”

The steric hindrance of the isopropyl group can be a significant factor in the selectivity of the E2 reaction. The consideration of sterics is a structural resource used extensively to rationalize outcomes in chemical reactions – particularly elimination reactions when distinguishing between Hofmann and Zaitsev products. Therefore, it is not surprising that the idea of sterics is incorporated in this example given the discussions regarding the interplay of electronic and steric factors in determining the major elimination product. Although this resource is productive, we did not consider it as part of a complete explanation in our grading criteria because it does not account for the required antiperiplanar conformation of the substrate and its importance to the E2 mechanism. The full consideration of the position of the isopropyl group requires the activation of resources related to Newman projections and hydrogens in the chair conformation. The antiperiplanar conformation requires manipulation of the structure – unlike the steric argument. Given that manipulation is required, we are classifying conformational analysis as a mechanistic resource in lieu of a structural resource, which can be realized by looking at the structure without doing any manipulation. We hypothesize that the full activation of this conformational resource was not observed due to its novelty to organic chemistry students and the fact that structural manipulation is needed to draw this conclusion. As depicted in Scheme 1, conformational resources, specifically chair conformations, are only emphasized in some units of organic chemistry and are absent in most general chemistry courses. Overall, steric arguments were used in four of the ten interviews.

The position of the double bond and extent of alkyl substitution was used by three of the ten interviewees to classify the reaction as E2. Students did not provide rationale supporting this claim – even when prompted. In the discussion below, the student alluded to the transition state using the Hammond Postulate; however, the student did not completely and accurately explain the significance of the transition state in the final outcome. The Hammond Postulate is used throughout discussions in lecture to support the energies of the intermediates, which may account for why this specific resource was activated.

“Okay, for the first one (E2) I was thinking [that] by Hammond's postulate, you can have a pretty stable transition state where you have the partial leaving of the Br group and the partial formation of the 2-prime (secondary) carbocation. In the second one (E1), when there's just methanol with E1 and that transition state you’re making, it's the same idea [as] with an allyl or a benzyl carbocation. What I think we talked about is that you have more 3-prime alkene character—tertiary alkene character—so in that elimination (E1) it's more stable.”

When prompted to explain what they meant by “more tertiary alkene character,” the student said:

“Yeah, so in this one (E1), in my mind, what I’m thinking is the bond of the Br, since it's elimination and it's getting kicked out, and that hydrogen on the tertiary carbon is getting removed. In my mind, I’ve always imagined the bonds swooping in to form the alkene.”

The student classified the first reaction as E1 and the second reaction as E2 and pointed to the substitution degree of the alkene product as the key differentiator for the elimination reactions, which supports the use of structural resources without extending to more complex mechanistic resources to support the claim. This product-oriented thinking is seen throughout the chemical education literature, often at the expense of mechanistic, “process-oriented” reasoning (Caspari et al., 2018; Dood and Watts, 2023). Alkene substitution and stability are discussed throughout the organic chemistry curriculum, so it is not surprising that some students used these resources to justify their answers, even if they are not directly applicable to the problem at hand. This reflects previously described observations relating to leaving group ability. Although more substituted alkenes are more stable than less substituted alkenes, increased stability is not the driving force behind the observed regioselectivity. The extent of substitution is a structural resource commonly used to rationalize differences in stabilities of alkenes and reactivity.

Other students’ explanations of the E2 mechanism showed varying use of recall and structural resources, but the common thread with these strategies is the use of acid–base chemistry to aid in identifying the mechanistic pathway. None of the interviewees used mechanistic resources to aid in their answers. The two dialogs illustrate additional strategies. The first dialog uses base strength extensively to draw conclusion:

“So, I noticed that [in] the first one, there is [an] additional NaOCH₃…when the Br leaves, [it] forms a positive charge on the carbon where it left. That leaves a negative OCH₃which prefers to stabilize it. Therefore [it is] more likely to go through E2.”

When probed about the mechanism for the E2 reaction, the student explained:

“Because if it (methoxide) is a strong base, then it is hard to have enough time for the leaving group to leave. Therefore, everything is going to happen in one step, and that's going to be E2.”

The student above correctly identified the E2 reaction; however, the reasoning behind why the reaction proceeds was not chemically accurate. Using electrostatics, the student suggested that the negative charge on the methoxide would stabilize the carbocation intermediate that forms when Br leaves. After further discussion, the student above correctly classified methoxide as a strong base, however, they did not appear to apply the acid–base mechanism of methoxide appropriately. This is significant, as the student classified methoxide as a base, but did not attribute Brønsted or Lewis base-like reactivity to the methoxide. Recall resources that identify and relate acid and base strength to mechanistic pathways do not necessarily translate to chemically sound mechanistic reasoning. Like previous discussion of leaving group ability, students can correctly identify bases without fully differentiating between reaction types. Although the mere identification of base strength features may result in correct answers, particularly on multiple choice questions, but more sophisticated structural and mechanistic resources should be required for students to show full understanding. Six of the ten interviewees use base strength to support their conclusions.

An average number of four resources were used during the interviews. The student who used the most resources used the number of molecules in addition to base strength to aid in their discussion. This argument was unexpected:

“Oh yeah, that's elimination, but it's with one molecule I think, and E2 involves two, but that's why I’m thinking it's E2.”

This student used the heuristic of the number of molecules present in a reaction to distinguish between E2 and E1, which may stem from the recall of discussions surrounding rate laws in which E1 is unimolecular and E2 is bimolecular. While molecularity is important in mechanistic analyses, the student did not use mechanistic resources such as arrow pushing, identification of transition states, etc. to justify the claim. The student supported their claim even when the interviewer pointed out the two molecules in the reaction.

While our survey assessments aimed to probe how students prioritize their chemical knowledge in their explanations, our interviews further triangulated that, for the E2 reaction specifically, base strength is likely the first resource that students activate. As mentioned previously, acid–base strength is often discussed starting in first-semester general chemistry and then reinforced in organic chemistry, whereas conformational analysis, including the antiperiplanar relationship, is typically introduced midway through first-semester organic chemistry. Additionally, when E2 elimination reactions are introduced, the conformation resource is often introduced at the end, which may account for why it is not as readily activated. Our data further support findings from the Cruz-Ramírez de Arellano et al., which also showed that students infrequently activate conformational resources when solving elimination problems (Cruz-Ramírez de Arellano and Towns, 2014). Nevertheless, conformations and stereochemistry mechanistic resources are integral components of organic chemistry and cannot be overlooked when solving more complex organic chemistry problems.

In many undergraduate organic chemistry courses, the E2 reaction is introduced alongside the E1 reaction, thus we also designed an E1 survey assessment/interview question. The reactant molecule is the same as in the E2 question because we intended for students to compare the two reactions, particularly to explain why the regioselectivity differs.

Q2: E1 elimination

Like the previous question, explanations for the E1 question (Fig. 1B) were evaluated based on the criteria in Table 1, all of which were emphasized in the OC1 curriculum at our institution. Of all our respondents, 54% correctly identified methanol acting as a base and classified the reaction as E1. From the student sample, 11% of students explained that a hydride shift must occur in the reaction mechanism to yield the provided product, which was included in the criteria for a “complete” explanation. Carbocation rearrangements are an important caveat to E1 mechanisms and are covered extensively in our curriculum (Loudon and Parise, 2021). Many students cited methanol's weak base character in their explanations. Although this statement is correct, it does not provide mechanistic rationale for the observed reaction.

Like with the E2 survey assessment, we observed a statistically significant difference in the self-reported confidence between students who answered the E1 multiple-choice questions correctly and incorrectly, where students who answered correctly were more confident than those who answered incorrectly (Fig. 3A). Additionally, there was a statistically significant difference in confidence between students who provided partially complete and incomplete explanations (Fig. 3B). Although there was not a statistically significant difference in confidence between students providing complete and partial explanations to the E1 question, this result could be attributed to the small sample size in the complete group, with only one complete student response. However, given the similarity between the E2 and E1 questions, and the analogous trends in the self-reported confidence, it is reasonable that students on average could discern the depth of their knowledge in the E1 question. Again, our findings contrast with the literature in that we did not observe a clear Dunning–Kruger effect, and this difference could be explained by how we compared the metrics measured in our study as well as how much prior chemistry instruction our student participants have had.

Fig. 3 Students’ self-reported confidence for the E1 question of the survey assessment, compared based on their performance on the (A) multiple-choice and (B) explanation portions. Productive vs. Unproductive: Wilcoxon p = 0.023, Wilcoxon effect size r = 0.34. Complete vs. Partial: p = 0.5, r = 0.04. Complete vs. Incomplete: p = 0.17, r = 0.21. Partial vs. Incomplete: p = 0.034, r = 0.32.

We developed inductive codes based on concepts discussed by students in their responses to the E1 survey assessment (Table 5). As with the E2 survey assessment, the E1 questions helped identify resources that students utilize first, with the two concepts notably being the identification of methanol as a Brønsted base and the recognition that weak bases favor E1. Given the order and presentation of questions (Fig. 1B), these results are not surprising, as acid–base chemistry cognitive resources had been previously activated. Less than half of students (43%) classified methanol as a base, with about half of this group proceeding to explain that weak bases are characteristic of E1.

Table 5 Inductive codes for E1 reasoning

Inductive code description	Percentage of students (N = 35)
Methanol is a base	43
Weak bases are characteristic of E1	23
Carbocation rearrangements occur	11
The more substituted product forms	9
Br leaves first	6

---

Among the 10 think-aloud interviews conducted to probe deeper into students’ thought processes. Given the E1 and E2 problems were presented back-to-back, an overlap in the themes was expected regarding resources with: (1) sterics, (2) extent of substitution, and (3) base strength. Additionally, unlike with E2 discussions, four interviewees used more complex mechanistic resources for their E1 reasoning by identifying carbocation rearrangements. Students who demonstrated a complete understanding provided thorough descriptions of both the acid–base chemistry and the mechanism of the E1 reaction using recall, structural, and mechanistic resources.

The base strength was identified by six of the ten interviewees to differentiate between the E1 and E2 mechanisms. This was generally one of the first resources students used in developing their arguments. Consider the dialog below. The student identifies the base strength first, but after prompting, identifies the carbocation rearrangement to provide a complete description of the mechanism.

“There's no strong base, so methanol is a weak base. There's no charge with it.”

This student first used acid–base resources to identify the weak base nature of methanol, proceeding to use this as their rationale for choosing E1. Upon additional prompting, the student provided a complete response by discussing the stepwise mechanism and carbocation rearrangement that underpin the regioselectivity of the reaction:

“So you form a carbocation once the bromide leaves, so that’d be a secondary carbocation, and by forming the double bond between—you could have a hydride shift, where the carbocation would move to the tertiary position above it—and that's why you would form the double bond between those two carbons.”

While base strength was one of the common resources, several students also used structural resources. Students who did not identify the carbocation rearrangement used structural resources related to the extent of alkene substitution.

“For this one, I would say E1 because there's a lack of a strong base. It's just the solvent, so the solvolysis.”

[…]

“It's because E1 undergoes a carbocation intermediate. It's more thermodynamically controlled, so it forms the more thermodynamically favorable alkene, which is the more substituted one.”

Although this student acknowledged the carbocation intermediate, they used resources related to the thermodynamics of the reaction and the formation of a more substituted alkene. While these are productive resources, the student did not acknowledge that the substrate would likely undergo a hydride shift to form a more thermodynamically favorable tertiary carbocation. Our findings build on previous literature that found that students struggle to appropriately depict carbocations in reaction mechanisms (Cruz-Ramírez de Arellano and Towns, 2014; Grove et al., 2012). Our data may suggest that students use product-centered approaches to explain E1 reactions. Rather than rationalizing carbocation formation based on the stability of the carbocation, students may merely identify that a carbocation forms in a manner consistent with the given product.

Among the students who provided incomplete explanations of the E1 reaction, some activated mechanistic resources but did so with developing accuracy.:

“So, the methanol would remove the hydrogen on that tertiary carbon, and then from there that would create a carbocation. Then from there…wait, how does that get there? It's like…I don’t fully remember.”

This student correctly stated that the methanol would act as a base to deprotonate a hydrogen, but incorrectly attributed carbocation formation to a deprotonation step. While deprotonation and carbocation formation both occur in the E1 mechanism, the former does not cause the latter. This student struggled to remember the role of the leaving group in the E1 mechanism and was unable to deduce the arrow-pushing steps through mechanistic reasoning.

The results from this question complement the data from the E2 question. Although the recognition of base strength can help students answer elimination questions correctly, there are more nuanced chemical phenomena that relate to reaction regioselectivity that are equally or more important.

Q3: E1cB elimination

To capture the entire spectrum of elimination reactions discussed in undergraduate organic chemistry, we also designed a question around the E1cB reaction in the context of an aldol condensation. Even if students had not explicitly heard the term “E1cB” before, we hypothesized that students could rationalize their way through this reaction using principles of acid–base chemistry.

Like the previous two questions, inductive codes were generated based on the trends observed in the student responses to the E1cB survey assessment (Table 6). 28% of survey respondents correctly identified hydroxide as a poor leaving group. A similar proportion identified the anionic species as a justification for the E1cB mechanism. Our survey daonly introduced rly show that, like previous questions, students often did not explain mechanistic details in their answers. These data suggest that resources regarding explicit structural features were most commonly used by students to solve the problem. 13% of students stated that the acidic alpha proton was a requirement for the E1cB mechanism. This reasoning was often coupled with statements that described the formation of the conjugate base, and that deprotonation was favorable due to resonance stabilization from the carbonyl. The acidic alpha proton is a pivotal component of the E1cB reaction and the aldol condensation (Clayden et al., 2012; Loudon and Parise, 2021). The importance of the acidic alpha proton was emphasized in the instruction of aldol condensations (Loudon and Parise, 2021). Our data show that, despite this direct instruction, most students did not instinctively use this acid–base resource in this scenario. It is possible that this resource of alpha proton acidity was not commonly activated due to its novelty. Unlike leaving group ability or base identification, which are resources that are introduced earlier in OC1, the acidity of a proton being a driving force for a reaction is not introduced until late in OC2. These data may suggest that concepts that are introduced earlier are more often recalled by students, even compared to concepts that are introduced more recently and have benefited from direct instruction.

Table 6 Inductive codes for E1cB explanation

Inductive code description	Percentage of students (N = 130)
Involves a poor leaving group (hydroxide)	28
Involves a carbanion/anionic species	28
Acidic alpha proton	13
Formation of the conjugate base	10
Favored because of resonance stabilization	10
“I don’t know, haven’t learned, etc.”	18

---

The results from the survey assessments were corroborated by our 10 think-aloud interviews. Two students provided complete explanations (Table 1) regarding structural acid–base resources:

“Okay, so I mentioned E1cB, which instead of forming a carbocation, you’re forming a carbanion, which is on the left in the resonance structures. So that's the main difference. And why does this reaction go through this special pathway rather than E1 or E2? Because the alpha-hydrogen is extra acidic…because of resonance stabilization of the carbanion with the carbonyl.”

When the student continued with their explanation to discuss the leaving group character of hydroxide, they extended that concept to the mechanistic resources by analyzing the energetics of the rate-determining step, demonstrating a level of understanding that was not explicitly observed in the survey assessments:

“I guess kicking off the hydroxide, even if it won’t change the pKa (pH) much, it's still a fairly okay leaving group—it's not that great, so thermodynamically it's not very favorable, and that's why it's rate-determining. And neither is it kinetically very favorable.”

Seven students provided partially complete explanations, a much higher proportion than was seen in the survey assessments. Of the partially complete explanations, all seven students constructed arguments surrounding using structural resources relating to the poor leaving group character of hydroxide rather than the acidity of the alpha proton:

“Well, I believe since OH is a bad leaving group, that can definitely slow down the reaction…The carbanions form first, which is different from the typical E1 or E2 reaction.”

When further prompted about why E1cB is favored over E1 and E2, aside from the poor leaving group, the student utilized thermodynamics resources:

“Why does this reaction go through a special pathway? It's because it's forming that C=O double bond, which is very favorable. The negative charge actually resides on that carbon versus that oxygen so that's going to be the one that's going to be nucleophile.”

Although mechanistic resources relating to thermodynamics are important, they do not fully explain why the E1cB is the favored reaction pahway, because the carbonyl double bond would not have to be broken or reformed in an E1 or E2 reaction. Rather, the bond breaks due to the deprotonation of an acidic alpha proton that yields a resonance-stabilized enolate. Our interview data parallels our survey data and suggests that, for E1cB reactions, the identification of leaving groups and carbanion intermediates are the resources that students utilize first. From the interviews, students appeared to identify the poor leaving group character of hydroxide more often than they did during the survey assessments. The acidic alpha proton, the primary differentiating factor of the E1cB reaction, appears to be a concept discussed less frequently, even when the arrow-pushing mechanism is provided. This is not surprising given the novelty of proton acidity as the driving force for reacions, and organic chemistry students are also known to struggle with pK_a predictions (Flynn and Amellal, 2016). Even so, the identification of acidic protons is an essential skill to solve more complex organic mechanisms.

We observed a statistically significant difference in the self-reported confidence between students who explained the E1cB question partially and incompletely (Fig. 4). Interestingly, there was no statistically significant difference in the confidence of students who provided complete and partial explanations. This may be due to the question design and the scoring criteria used; a complete explanation was not much more mechanistically involved than a partially complete explanation, unlike the E2 and E1 questions. On average, however, we discovered a similar trend in the E2, E1, and E1cB questions that were used in our study. Students who performed the best also had the highest self-reported confidence, with minimal evidence of the Dunning–Kruger effect that has been observed in other studies involving undergraduate chemistry students. As with the previous two questions in our study, this difference could be explained by the specific metacognitive metrics measured in our study and the extent of chemistry background among the student participants.

Fig. 4 Students’ self-reported confidence for the E1cB explanation question of the survey assessment, compared based on their performance. Complete vs. Partial: Wilcoxon p = 0.32, Wilcoxon effect size r = 0.075. Complete vs. Incomplete: p = 0.01, r = 0.235. Partial vs. Incomplete: p = 0.001, r = 0.285.

Surprisingly, student performance on the E1cB question was comparable to that on the E2 and E1 questions. At our research institution, E1cB reactions are not covered until the end of OC2 and often are not covered thoroughly. It is important to note, however, that students were provided the arrow-pushing mechanism for the E1cB reaction while they were required to elucidate the mechanism to answer questions about E1 and E2 reactions.

Across the board, confidence in the E1cB responses was lower than confidence in E2 questions, an expected outcome based on the coverage of each reaction in the organic chemistry sequence. Additionally, the E1cB terminology could have stood as a barrier to student confidence, even if students’ chemical understanding was sufficient. This hypothesis is reflected by the number of student responses coded as “I don’t know” in their E1cB reasoning (Table 6). Students from Cohorts B and C reported a significantly higher confidence in their E1cB responses than students from Cohort A, despite no statistically significant differences in their response completeness. Given the similar presentation of the question in the assessment, this discrepancy can be attributed to individual instructor emphasis. Instructor A covered the aldol condensation without mentioning the E1cB mechanism by name, while Instructors B and C explicitly instructed students on this type of elimination reaction. This raises an interesting point in that direct instruction of E1cB reactions increased student confidence but did not improve overall performance.

Conclusion

Overall, we have designed a mixed-methods study using survey assessments and think-aloud interviews to probe student understanding of acid–base chemistry in relation to the three elimination reactions commonly covered by the end of second-semester organic chemistry. The survey assessments were used to identify the resources that students immediately employ when solving elimination reactions, and the interviews were used to probe deeper into their thought processes. The results from the self-reported confidence showed that students who provided complete explanations had the highest confidence in their answers and explanations, whereas students who provided incomplete explanations had the lowest confidence.

Our results suggest that for the E2 and E1 reactions, students primarily used acid–base resources to distinguish between the reactions. These observations were triangulated by think-aloud interviews, where further prompting rarely solicited responses based on conformation and mechanism. This is unsurprising, as students are more likely to utilize resources that have been reinforced over several semesters (Hammer et al., 2005). Base strength is introduced before conformational, steric, or thermodynamic approaches to elimination reactions in several textbooks (Smith, 2013; Karty, 2021; Klein, 2021; Loudon and Parise, 2021). Additionally, students are more likely to employ resources that have previously led to success. Our data clearly show that base strength alone can be used to correctly answer our multiple-choice questions, even if students do not proceed to describe other chemical phenomena at play.

For the E1cB reaction, students primarily used leaving group resources and the presence of a carbanion intermediate to differentiate it from E2 and E1 reactions. However, the more fundamental differentiating factor, the acidic alpha proton, was discussed less frequently in both the survey assessments and the interviews. This observation aligns with other reports in the literature that show student difficulty with identifying proton acidity and the prediction of pK_a values (Flynn and Amellal, 2016). Our results may suggest that the identification of acidic protons is a resource not as thoroughly developed as compared to the identification of more surface-level structural features, which could explain why fewer students called upon this specific resource. The results from our three questions provide insight into how to design assessments that activate all chemical resources relevant to achieving course goals.

Implications for teaching

The transition from general to organic chemistry requires a shift from quantitative to qualitative problem-solving, in which students learn to draw upon several resources to formulate models and draw conclusions. Acid–base resources are introduced in general chemistry and are reiterated in the introduction of the EPF at the start of organic chemistry. Because of the emphasis and importance of acid–base resources, it is not surprising that students readily draw upon them (Hammer et al., 2005). Although acid–base resources are important to rationalize elimination reactions, many problems require resources introduced in the specific context of each organic reaction. For elimination mechanisms, most textbooks start by either emphasizing the substrate substitution or the strength of the base, which further promotes the recall of acid–base resources. The salient points regarding mechanisms that provide a deeper understanding generally follow this introduction centered around acid–base chemistry (Smith, 2013; Karty, 2021; Klein, 2021; Loudon and Parise, 2021).

As practitioners and instructors of organic chemistry, we need to consider the learning outcomes of courses and gauge whether formative and summative assessments are calibrated to measure these outcomes. Students should have formative assessments that assess the nuanced details of reactions with opportunities for feedback and reflection. Additionally, because of the multiple layers and resources required for mechanistic reasoning, incorporating scaffolds within the classroom may promote greater recall and incorporation of the necessary resources needed for mechanistic reasoning. In a recent article, Kranz, Sween, and Graulich outlined strategies for developing scaffolds and demonstrated the gains associated with their implementation (Kranz et al., 2023). The development of scaffolds for elimination reactions for implementation during lecture and during assessments can aid in activating the appropriate recall, structural, and mechanistic resources necessary for complete mechanistic reasoning. The scaffolded assignments will require students to extend their resource use to include the more complex mechanistic resources that are generally paramount for successfully organic problems. As an example, the following are a series of questions that we are proposing to use in the upcoming semester in scaffold for analyzing general organic mechanisms:

(1) What bonds are formed and broken in the chemical reaction?

(2) What type of chemical reaction is occurring (addition, substitution, redox, or elimination)? Explain your reasoning.

(3) What is the role of each reactant? What is the solvent?

(4) What is the solvent? How does the solvent impact the stability of the starting materials? How can this impact the mechanism?

(5) Would the reaction be concerted or stepwise? Explain your reasoning.

(6) If stepwise, draw the possible intermediates and identify which one is more kinetically favored.

(7) Can rearrangements occur with the intermediates? Explain your reasoning.

(8) Write an arrow pushing mechanism for the reaction.

(9) Analyze the mechanism and rationalize the observed regioselectivity. Draw additional structures/conformations to support your reasoning.

(10) Analyze the mechanism and rationalize the observed stereochemistry. Draw additional structures/conformations to support your reasoning.

Limitations and future work

The focus of this study was on acid–base chemistry, thus the survey and interview questions were designed to prime students to think about concepts related to acid–base chemistry. This framing of the questions may have led students to identify the acid–base factors involved in each elimination reaction and end their explanation before discussing more fundamental mechanistic details, particularly in the E2 and E1, where the products were provided. Although this limitation was partially addressed by the think-aloud interviews, it would be of interest to continue the research using the same survey questions, but without the priming acid–base questions. We also acknowledge that the sample size for our interview data (N = 10) was small and could have self-selection bias since the interviews were performed voluntarily. An alternative avenue to explore would be to prompt students to extend their explanations beyond acid–base resources. The prompting could follow recommendations recently outlined in work by the Kranz et al. that investigated how to systematically build up students’ mechanistic explanations (Kranz et al., 2023).

The self-reported confidence data were intriguing because they did not reveal a clear Dunning–Kruger effect, contrary to what has been found in prior studies. These studies investigated the Dunning–Kruger effect using a course-wide lens, comparing perceived versus actual general chemistry exam scores. However, these studies did not further analyze their results based on specific questions asked on the exams. A potential avenue to further explore these metacognitive data would be to implement a research instrument analogous to an exam that focuses on elimination reactions and ask students what they think they scored. The Dunning–Kruger effect may be more pronounced for certain chemistry concepts, while other concepts may not show the effect at all.

The results from this study have also uncovered additional research questions regarding student understanding of thermodynamics and kinetics and how they relate to mechanistic reasoning. Perhaps the lack of discussion of the reaction mechanism was because students struggled to effectively deconstruct information into manageable chunks before synthesizing them again (Gobet, 2005). Additional research could be pursued to probe student understanding of thermodynamics and kinetics, as well as how students chunk information to rationalize organic reactivity.

Conflicts of interest

There are no conflicts to declare.

Appendix

 

Fig. 5 Mechanism of E2 reaction. Antiperiplanar elimination is favored, explaining the regioselectivity of the reaction (Clayden et al., 2012).

Fig. 6 Possible mechanistic pathways for E1 reaction. Most probable is the formation of a secondary carbocation, followed by a hydride shift to rearrange to a tertiary carbocation, then elimination to form an endocyclic double bond.

Acknowledgements

We would like to say thank you to all of the participants who contributed to this study. This work was kindly supported by Duke University and Duke Learning Innovation.

References

1. Anzovino M. E. and Bretz, S. L. (2016), Organic chemistry students’ fragmented ideas about the structure and function of nucleophiles and electrophiles: a concept map analysis, Chem. Educ. Res. Pract., 17(4), 1019–1029 10.1039/C6RP00111D.
2. Anzovino M. E. and Lowery Bretz S., (2015), Organic chemistry students’ ideas about nucleophiles and electrophiles: the role of charges and mechanisms, Chem. Educ. Res. Pract., 16(4), 797–810 10.1039/C5RP00113G.
3. Arnaud C. H., (2020), Weeding out inequity in undergraduate chemistry class, Chem. Eng. News, 98(34), 34–37.
4. Atkinson M. B., Popova M., Croisant M., Reed D. J. and Bretz S. L., (2020), Development of the Reaction Coordinate Diagram Inventory: Measuring Student Thinking and Confidence, J. Chem. Educ., 97(7), 1841–1851 DOI:10.1021/acs.jchemed.9b01186.
5. Bell P. and Volckmann D., (2011), Knowledge Surveys in General Chemistry: Confidence, Overconfidence, and Performance, J. Chem. Educ., 88(11), 1469–1476 DOI:10.1021/ed100328c.
6. Berardi M. D., Gentile F., Kozik I. and Gregg T. M., (2021), Aldol Condensation Reaction Rate Demonstrates Steric and Electronic Substituent Effects in the Organic Chemistry Lab, J. Chem. Educ., 98(5), 1732–1735 DOI:10.1021/acs.jchemed.0c00448.
7. Bhattacharyya G., (2013), From Source to Sink: Mechanistic Reasoning Using the Electron-Pushing Formalism, J. Chem. Educ., 90(10), 1282–1289 DOI:10.1021/ed300765k.
8. Boothe J. R., Zotos E. K. and Shultz G. V., (2023), Analysis of post-secondary instructors’ pedagogical content knowledge of organic acid–base chemistry using content representations, Chem. Educ. Res. Pract., 24, 577–598.
9. Bretz S. L., (2008), Qualitative Research Designs in Chemistry Education Research, in Nuts and Bolts of Chemical Education Research, American Chemical Society, vol. 976.
10. Bretz S. L. and McClary L., (2015), Students’ Understandings of Acid Strength: How Meaningful Is Reliability When Measuring Alternative Conceptions? J. Chem. Educ., 92(2), 212–219 DOI:10.1021/ed5005195.
11. Bunce D. M., (2008), Constructing Good and Researchable Questions, in Nuts and Bolts of Chemical Education Research, American Chemical Society, vol. 976, pp. 35–46 DOI:10.1021/bk-2008-0976.ch004.
12. Cartrette D. P. and Mayo P. M., (2011), Students’ understanding of acids/bases in organic chemistry contexts, Chem. Educ. Res. Practice, 12(1), 29–39 10.1039/C1RP90005F.
13. Caspari I., Weinrich M. L., Sevian H. and Graulich N., (2018), This mechanistic step is “productive”: Organic chemistry students’ backward-oriented reasoning, Chem. Educ. Res. Pract., 19(1), 42–59 10.1039/C7RP00124J.
14. Clayden J., Greeves N. and Warren S., (2012), Organic Chemistry, Oxford University Press.
15. Cooper M. M., Kouyoumdjian H. and Underwood S. M., (2016), Investigating Students’ Reasoning about Acid–Base Reactions, J. Chem. Educ., 93(10), 1703–1712 DOI:10.1021/acs.jchemed.6b00417.
16. Coutinho N. D., Machado H. G., Carvalho-Silva V. H. and da Silva W. A., (2021), Topography of the free energy landscape of Claisen–Schmidt condensation: solvent and temperature effects on the rate-controlling step, Phys. Chem. Chem. Phys., 23(11), 6738–6745 10.1039/D0CP05659F.
17. Cruz-Ramírez de Arellano D. and Towns M. H., (2014), Students’ understanding of alkyl halide reactions in undergraduate organic chemistry, Chem. Educ. Res. Practice, 15(4), 501–515 10.1039/C3RP00089C.
18. Dood A. J. and Watts F. M., (2022), Mechanistic Reasoning in Organic Chemistry: A Scoping Review of How Students Describe and Explain Mechanisms in the Chemistry Education Research Literature, J. Chem. Educ., 99(8), 2864–2876 DOI:10.1021/acs.jchemed.2c00313.
19. Dood A. J. and Watts F. M., (2023), Students’ Strategies, Struggles, and Successes with Mechanism Problem Solving in Organic Chemistry: A Scoping Review of the Research Literature, J. Chem. Educ., 100(1), 53–68 DOI:10.1021/acs.jchemed.2c00572.
20. Dood A. J., Fields K. B. and Raker J. R., (2018), Using Lexical Analysis To Predict Lewis Acid–Base Model Use in Responses to an Acid–Base Proton-Transfer Reaction, J. Chem. Educ., 95(8), 1267–1275 DOI:10.1021/acs.jchemed.8b00177.
21. Dood A. J., Dood J. C., Cruz-Ramírez de Arellano, D., Fields, K. B. and Raker, J. R., (2020), Analyzing explanations of substitution reactions using lexical analysis and logistic regression techniques, Chem. Educ. Res. Pract., 21(1), 267–286 10.1039/C9RP00148D.
22. Duis J. M., (2011), Organic Chemistry Educators’ Perspectives on Fundamental Concepts and Misconceptions: An Exploratory Study, J. Chem. Educ., 88(3), 346–350 DOI:10.1021/ed1007266.
23. Finkenstaedt-Quinn S. A., Watts F. M., Petterson M. N., Archer S. R., Snyder-White E. P. and Shultz G. V., (2020), Exploring Student Thinking about Addition Reactions, J. Chem. Educ., 97(7), 1852–1862 DOI:10.1021/acs.jchemed.0c00141.
24. Flynn A. B. and Amellal D. G., (2016), Chemical Information Literacy: PKa Values—Where Do Students Go Wrong? J. Chem. Educ., 93(1), 39–45 DOI:10.1021/acs.jchemed.5b00420.
25. Frost S. J. H., Yik B. J., Dood A. J., de Arellano D. C.-R., Fields K. B. and Raker J. R., (2023), Evaluating electrophile and nucleophile understanding: a large-scale study of learners’ explanations of reaction mechanisms, Chem. Educ. Res. Pract., 24(2), 706–722 10.1039/D2RP00327A.
26. Galloway K. R., Stoyanovich C. and Flynn A. B., (2017), Students’ interpretations of mechanistic language in organic chemistry before learning reactions, Chem. Educ. Res. Pract., 18(2), 353–374 10.1039/C6RP00231E.
27. Galloway K. R., Leung M. W. and Flynn A. B., (2019), Patterns of reactions: a card sort task to investigate students’ organization of organic chemistry reactions, Chem. Educ. Res. Pract., 20(1), 30–52 10.1039/C8RP00120K.
28. Gao S., Outlaw T. C., Liang-Lin J., Feng A., Roizen J., Melnick C. and Cox Jr. C. T., (2022), Students’ Identification and Application of Models to Rationalize Organic Acid-Base Trends. University Chemistry: Teaching in the 21st Centruy, University of Ljubljani, pp. 40–67. https://zalozba.pef.uni-lj.si/index.php/zalozba/catalog/view/198/456/479-1.
29. Garg N. K., (2019), How organic chemistry became one of UCLA's most popular classes, J. Biol. Chem., 294(46), 17678–17683 DOI:10.1074/jbc.AW119.008141.
30. Gobet F., (2005), Chunking models of expertise: Implications for education, Appl. Cognitive Psychol., 19(2), 183–204 DOI:10.1002/acp.1110.
31. Graulich N., (2015), The tip of the iceberg in organic chemistry classes: How do students deal with the invisible? Chem. Educ. Res. Pract., 16(1), 9–21 10.1039/C4RP00165F.
32. Grove N. P. and Lowery Bretz S., (2012), A continuum of learning: from rote memorization to meaningful learning in organic chemistry. Chem. Educ. Res. Pract., 13(3), 201–208 10.1039/C1RP90069B.
33. Grove N. P., Cooper M. M. and Rush K. M., (2012), Decorating with Arrows: Toward the Development of Representational Competence in Organic Chemistry, J. Chem. Educ., 89(7), 844–849 DOI:10.1021/ed2003934.
34. Hammer D., Elby A., Scherr R. and Redish E. F., (2005), Resources, framing, and transfer, Transfer of Learning from a Modern Multidisciplinary Perspective, Information Age Publishing, pp. 89–119.
35. Hartman J. R., Nelson E. A. and Kirschner P. A., (2022), Improving student success in chemistry through cognitive science, Found. Chem., 24(2), 239–261 DOI:10.1007/s10698-022-09427-w.
36. Henderleiter J., Smart R., Anderson J. and Elian O., (2001), How Do Organic Chemistry Students Understand and Apply Hydrogen Bonding? J. Chem. Educ., 78(8), 1126 DOI:10.1021/ed078p1126.
37. Karty J., (2021), Organic Chemistry: Principles and Mechanisms, W. W. Norton, Incorporated. https://books.google.com/books?id=RlWbzgEACAAJ.
38. Klein D. R., (2021), Organic Chemistry. Wiley. https://books.google.com/books?id=YC09EAAAQBAJ.
39. Kranz D., Schween M. and Graulich N., (2023), Patterns of reasoning – exploring the interplay of students’ work with a scaffold and their conceptual knowledge in organic chemistry, Chem. Educ. Res. Pract., 24, 453–477.
40. Lamichhane R., Reck C. and Maltese A. V., (2018), Undergraduate chemistry students’ misconceptions about reaction coordinate diagrams, Chem. Educ. Res. Pract., 19(3), 834–845 10.1039/C8RP00045J.
41. Loudon M. and Parise J., (2021), Organic Chemistry, 7th edn, MacMillan Learning.
42. Mak K. K. W., Chan W.-F., Lung K.-Y., Lam W.-Y., Ng W.-C. and Lee S.-F., (2007), Probing the Rate-Determining Step of the Claisen-Schmidt Condensation by Competition Reactions. J. Chem. Educ., 84(11), 1819 DOI:10.1021/ed084p1819.
43. McKight P. E. and Najab J., (2010), Kruskal-Wallis Test, The Corsini Encyclopedia of Psychology, pp. 1–1 DOI:10.1002/9780470479216.corpsy0491.
44. Morrison R. and Boyd R., (1959), Organic Chemistry, Allyn and Bacon Inc.
45. Nedungadi S. and Brown C. E., (2021), Thinking like an electron: Concepts pertinent to developing proficiency in organic reaction mechanisms, 3(1), 9–17 DOI:10.1515/cti-2019-0020.
46. Parobek A. P., Chaffin P. M., Towns M. H., (2021), Location-thinking, value-thinking, and graphical forms: combining analytical frameworks to analyze inferences made by students when interpreting the points and trends on a reaction coordinate diagram, Chem. Educ. Res. Pract., 22 (3), 697–714.
47. Pazicni S. and Bauer C. F., (2014), Characterizing illusions of competence in introductory chemistry students, Chem. Educ. Res. Pract., 15(1), 24–34 10.1039/C3RP00106G.
48. Perrin C. L. and Chang K.-L., (2016), The Complete Mechanism of an Aldol Condensation, J. Org. Chem., 81(13), 5631–5635 DOI:10.1021/acs.joc.6b00959.
49. Petterson M. N., Watts F. M., Snyder-White E. P., Archer S. R., Shultz G. V. and Finkenstaedt-Quinn S. A., (2020), Eliciting student thinking about acid–base reactions via app and paper–pencil based problem solving, Chem. Educ. Res. Pract., 21(3), 878–892 10.1039/C9RP00260J.
50. Popova M. and Bretz S. L., (2018a), “It's Only the Major Product That We Care About in Organic Chemistry”: An Analysis of Students’ Annotations of Reaction Coordinate Diagrams, J. Chem. Educ., 95(7), 1086–1093 DOI:10.1021/acs.jchemed.8b00153.
51. Popova M. and Bretz S. L., (2018b), Organic chemistry students’ challenges with coherence formation between reactions and reaction coordinate diagrams, Chem. Educ. Res. Pract., 19(3), 732–745 10.1039/C8RP00064F.
52. Popova M. and Bretz S. L., (2018c), Organic chemistry students’ interpretations of the surface features of reaction coordinate diagrams, Chem. Educ. Res. Pract., 19(3), 919–931 10.1039/C8RP00063H.
53. Popova M. and Bretz S. L., (2018d), Organic Chemistry Students’ Understandings of What Makes a Good Leaving Group, J. Chem. Educ., 95(7), 1094–1101 DOI:10.1021/acs.jchemed.8b00198.
54. softwareR Core Team, (2020), R: A language and environment for statistical computing. [Computer software], R Foundation for Statistical Computing. https://www.R-project.org/.
55. Salame I. I., Casino P. and Hodges N., (2020), Examining Challenges that Students Face in Learning Organic Chemistry Synthesis, Int. J. Chem. Educ. Res., 4(1), 1–9 DOI:10.20885/ijcer.vol4.iss1.art1.
56. Salame I. I., Montero A. and Eschweiler D., (2022), Examining some of the Students’ Challenges and Alternative Conceptions in Learning about Acid-base Titrations, Int. J. Chem. Educ. Res., 1–10 DOI:10.20885/ijcer.vol6.iss1.art1.
57. Schmidt-McCormack J. A., Judge J. A., Spahr K., Yang E., Pugh R., Karlin A., Sattar A., Thompson B. C., Gere A. R. and Shultz G. V., (2019), Analysis of the role of a writing-to-learn assignment in student understanding of organic acid–base concepts, Chem. Educ. Res. Pract., 20(2), 383–398 10.1039/C8RP00260F.
58. Smith J. G., (2013), Organic Chemistry, McGraw-Hill Education. https://books.google.com/books?id=7FabMgEACAAJ.
59. Stoyanovich C., Gandhi A. and Flynn A. B., (2015), Acid–Base Learning Outcomes for Students in an Introductory Organic Chemistry Course, J. Chem. Educ., 92(2), 220–229 DOI:10.1021/ed5003338.
60. Taber K. S., (2000), Chemistry lessons for universities? A review of constructivist ideas, Univ. Chem. Educ., 4(2), 26–35.
61. van Kesteren M. T. R., Krabbendam L. and Meeter M., (2018), Integrating educational knowledge: reactivation of prior knowledge during educational learning enhances memory integration, Npj Sci. Learn., 3(1), 11 DOI:10.1038/s41539-018-0027-8.
62. Watts F. M., Schmidt-McCormack J. A., Wilhelm C. A., Karlin A., Sattar A., Thompson B. C., Gere A. R. and Shultz G. V., (2020), What students write about when students write about mechanisms: analysis of features present in students’ written descriptions of an organic reaction mechanism, Chem. Educ. Res. Pract., 21(4), 1148–1172 10.1039/C9RP00185A.
63. Watts F. M., Park G. Y., Petterson M. N. and Shultz G. V., (2022), Considering alternative reaction mechanisms: students’ use of multiple representations to reason about mechanisms for a writing-to-learn assignment, Chem. Educ. Res. Pract., 23(2), 486–507 10.1039/D1RP00301A.
64. Wittmann M. C., (2018), Research in the Resources Framework: Changing environments, consistent exploration, arXiv preprint arXiv:1801.09592.
65. Yue C.-J. and Gu L.-P., (2022), Understanding and Learning of Ionic Organic Reactions in Organic Chemistry Based on Acid–Base Theory, J. Chem. Educ., 2022, 99(6), 2291–2297.

---

Footnote

† Co-First Authors.

---