Jack Dunitz on the Nature of the Chemical Bond by Linus Pauling -
"At the time when I was reading that book I was wondering whether chemistry was really as interesting as I had hoped it was going to be. And I think I was almost ready to give it up and do something else. I didn’t very much care for this chemistry which was full of facts and recipes and very little thought in it, very little intellectual structure. And Pauling’s book gave me a glimpse of what the future of chemistry was going to be and particularly, perhaps, my future."1
I have been teaching high school chemistry for sixteen years. During the past four years I have been reading and incorporating cognitive science into my classroom and sharing it with colleagues. How the brain works is a critical component of teaching and learning. I have yet to see a more consistent framework to analyze all of education than starting with what we know about the brain. Yet the educational practices emerging from some who use cognitive science are inconsistent with my readings and use within my own classroom. As I have grown in applying cognitive science to my teaching there has been a profound positive impact on my students. I want that same growth to reach as many teachers as possible.
Teachers who utilize inquiry-based pedagogy must have avenues to explore cognitive science. Currently many teachers are stuck. They have the success of a new pedagogy that can easily expand to include cognitive science, but so many articles stem from philosophies that needlessly contradict with their teaching approaches.
Undermining phlogiston theory was challenging. Scientists developed nuanced applications of phlogiston theory that made sense individually. But when considered together, conflicts arose. Likewise I propose that we have a growing trend in education of incorrect use of cognitive load theory. Much like phlogiston or the Bohr model, we have revolutionary evidence, but have reached the wrong conclusion from it. This misapplication is revealed through conflicts with other components of learning and cognitive science. Phlogiston theory was replaced with the superior model of oxygen. Oxygen allowed better progress and refinement of chemistry. Likewise it is time to deviate from a path where cognitive load theory is reductively used in place of the many complex and intricate components of cognitive science. The field of education is ready for advanced workings within classrooms via the best teachers in the world.
How does learning occur? How do we define what learning is? Initially we are going to define learning as a permanent change within the structure of the brain either through acquisition of new knowledge or by connecting multiple ideas. These can occur in tandem. We can sequence learning into sensory memory (student observes something through sensation), short-term memory (STM) (student engages with the observation), and long-term memory (LTM) (student combines the new information with prior knowledge stored in LTM). The learning is tenuous until the information is partially forgotten, reintroduced, and the student has had practice with retrieving the information at these spaced intervals.
But should we consider all transfer of information to LTM to be equivalent learning? Of course not. You can easily learn the fact that I was born in 1983. This is especially true if you were also born or had a major life experience near that year. If I ask you later in this article when I was born you can reproduce that information. On the other hand, developing a mental model for what an electron configuration represents about an atom is a much more complex transfer of information. I hear too many teachers expressing that novice students are incapable of doing such learning. There is evidence to the contrary. And it is concerning with how readily some teachers and researchers are celebrating the move away from challenging thinking.
Many articles and books are happy to provide a definition of learning. The permanent change in brain structure. But fewer are willing to digress into the more nuanced analysis required for thinking. Recall a problem or discrepancy that you had to think about. What happened? Thinking involves an interplay between prior knowledge in LTM and the current dilemma in STM. If we observe a discrepant event, the sensory memory transfers observations to STM where we seek commonality, patterns, and information from LTM to help rationalize. The assumptions of what thinking is are where ARTC and STC diverge.
Abstract reductivist teaching and cognition (ARTC) focuses on the step where short-term memory interacts with long-term memory using a model called cognitive load theory (CLT). Our research shows definitively that our STM can be overwhelmed when tasked with remembering too many pieces of information. There are limits to how quickly new information can be taken in. This can be assisted by developing chunks. Chunks are often defined using simple examples where a chunk of information can be thought of easily as 1 thing. 1-2-3-4-5-6-7-8-9 is much easier to remember than 2-7-5-6-4-5-7-8-3 even though both are 9 digits. The first set is just 1-9 in order. I only have to remember that I start with 1, end with 9, and that the numbers go in sequence. I can chunk the digits into 3 pieces of information. As we gain experience with learning about a topic, we find ways to better chunk information both in how we recognize information, and how we structure that information in seeking solutions. In the zone of proximal development model, cognitive overload would be when the material is too advanced for a student. If I try to learn tensor calculus for general relativity models I’m likely to experience cognitive overload and shut down.
ARTC also addresses that novice students solve problems differently than experts. An expert notices key details while a novice is distracted by surface features. This is established.2 The problem with ARTC is that it definitively states that the solution to this issue is to have students solve only abstract problems with high repetition. Such a solution shows disagreement with several tenets of cognitive science. There is no research showing that students gaining experience by solving novel problems is less effective than students doing abstract reasoning. There is evidence that students who only perform abstract algorithms struggle immensely when tasked with applying these models to novel situations. This is an inherent issue within ARTC as most assessments cited involve reductive assessments. Additionally many critics define inquiry-based instruction using unstructured exploration models only. Guided instruction found in modeling instruction is ignored.
A comment on a blog post led me to an article in support of traditional teaching (lecture followed by student mimicking an abstract algorithm performed by the instructor) claims that the sequence was more effective when the direct instruction was posited first.3 However, when carefully read, it turns out that the assessment was to mimic the abstract algorithm while varying superficial details (IE the mechanical device). If one were to take the article on face value, the insufficient assessment would not only be omitted, but the article actually describes the variation as transfer when in fact no alternative representations or extensions are utilized. Contradictory claims can be found in this article which has improved assessment, although may also have similar limitations.4
Strategic teaching and cognition (STC) focuses on providing a student with a challenging problem that requires the student to work between STM and LTM. What do I know, what do I not know yet, and how can I reconcile the unknown with the tools at my disposal? This form of learning appears throughout successful cognitive science texts such as in Make it Stick.5 The common term is “desirable difficulties” but in inquiry pedagogy teachers can make these much more specific. During a circle discussion where students are presenting whiteboards of their current mental models, a master teacher can emphasize a desirable difficulty while minimizing uses of vocabulary and mathematics. Recall that cognitive load theory does not state that critical thinking is impossible, rather that there is a limit to how many different things the mind can consider at once. By off-setting algorithms and vocabulary (two abstract features) the teacher can guide students to think deeply about chemistry content. The teacher can later use the new chunks of content to introduce vocabulary as a retrieval cue or mathematical reasoning.
Evidence for this efficacy can be found in the challenging task of undermining student misconceptions. When students are told they are incorrect and have a misconception, the students are likely to dig in even deeper. Rather when students are guided through questioning, a teacher can undermine the misconception through presentation of novel questions and an opportunity for the student to find patterns, exhibit sense-making, and reflect on their model development.
Statistical evidence can be found by modeling instructors using high-quality assessments such as the concept inventories. These assessments are designed to root out scientific misconceptions within students and modeling instruction shows far greater gains then traditional instruction. Traditional science instruction led to a Hake gain of 0.23 while modeling instruction led to Hake gain of 0.48.6
When students are provided a problem that they’ve never seen before is an optimal chance for students to stress their limits of cognition. In science this is a tremendous opportunity for teachers to assess what a student truly believes if they have the tools (IE talk moves) to elicit authentic ideas from a student.7 While numerous roadblocks exist, the talk from ARTC is disheartening in that they do not even utilize such a powerful opportunity for learning on the basis (in my view) of flawed assessment.
In the book The Inner Game of Tennis Timothy Gallwey describes how he utilizes cognitive science to teach sports.8 If we consider applying the ARTC methodologies to sports or driving they would be laughably rejected. Imagine learning to drive a car without ever being put into a situation where you actually drive a car. The idea that we would have students articulate abstract ideologies about driving to the point of overlearning as a means of instruction would never happen. Likewise we would also not coach a team by having them work abstract problems about trajectories of a baseball in order for them to optimally learn how to play baseball. The lack of transfer of ARTC to other fields should give us cautious skepticism.
Evaluating teaching is incredibly difficult. The assessment used has a tremendous impact on what results will be seen. Quality assessments in education are difficult to produce and large groups of teachers have functioned without meaningful assessment. Additionally, there is variety among both teachers and learners. While it is my belief that all learning occurs in the same manner, students bring a variety of prior knowledge, motivation, and emotional engagement to the classroom. Teachers set different cultures, shape how students approach learning, and offer their own varying levels of pedagogical and content knowledge. Cognitive science should open doors for us to evaluate which approaches to teaching and learning cause the biggest shifts from middle to top, and we should be working to determine which teaching methods are most impactful on our top students and our top teachers.
Far too many teachers lack trust in educational research because any intervention can produce favorable results when compared to marginal instruction and/or assessment. Moving forward research should make an effort to distinguish instruction into groups. Methods should be compared for a various levels of students and teachers. Assessments should be scrutinized. Anecdotal evidence of student thinking should be included. I remain skeptical of those who claim that teaching and cognition are simple. We have seen for centuries how experience is critical for success in teaching. This suggests that we embrace the complexity of the dynamics within a classroom. I believe that STC is a better path forward to embracing all of what we know about cognitive science and applying that complexity to our classrooms.
Abbreviations and terms used:
STM - short-term memory
LTM - long-term memory
Sensory memory
ARTC - Abstract Reductivist Teaching and Cognition
STC - Strategic Teaching and Cognition
System 1 - A representation of automated memory processing from the book Thinking Fast and Slow
System 2 - A representation of deep thinking from the book Thinking Fast and Slow
Chunking - The representation of multiple pieces of information using a single retrieval cue.
Retrieval practice - Retrieving information directly from your brain without the use of an external source
Spaced practice - The repetition of retrieval for a particular piece of knowledge. The repetition causes the neural structure to lock in more permanently as the brain realizes this must be relevant since it has been repeated.
Emotion - Evolved traits that help us survive. Emotions for learning can include curiosity, obedience, spite, anxiety, competitiveness, and others that vary in levels of toxicity, effectiveness, and balance.
Mindfulness - A model of observation. Mindfulness research shows that using a component of an observation for a task is more effective than focusing on the object itself.
Concrete examples - Observations that can be sensed or experienced.
Abstract - Underlying concepts or ideas that are shared between multiple concrete examples.
How I use cognitive science within my own classroom:
The pedagogy that structures my classroom is modeling instruction.9 Modeling instruction advances current student mental models. Traditional instruction states the teacher’s mental model and repeats as students approach. Traditional instruction is flawed in this sequence as students bring a variety of models and thus the only way for this to progress is with narrow definitions that do not challenge misconceptions or even work to build concepts at all.
A typical sequence of events in modeling instruction would start with a phenomenon. The students would gather data and make observations. Then they are tasked with organizing their data and observations into a whiteboard. This whiteboard might include their data, particle representations of the chemicals involved, graphs, line of best fit equations, and macroscopic images. These are intentionally chosen to be a variety of concrete and abstract components. Students can then be directed to focus on 1-2 of each at a time. This offsets issues with cognitive overload while providing students a roadmap into complex thinking. Here we see the interplay between STM and LTM as they take the new sensory information (phenomena) into STM where they seek patterns/alignment with prior knowledge (LTM).
The teacher utilizes talk moves to encourage students to express their current working models of how they explain patterns, data, and observations. The teacher then pushes students into more advanced components to enhance those models. A small example of utilizing cognitive science during discussions is when a student asks a question the teacher avoids responding immediately with a solution. When a student is thinking we are best off not interrupting that thinking. The student should search their LTM for information that might apply to finding resolution with the conflict they are exploring in STM. “Wait time” is a powerful tool. Thinking takes time and we must provide opportunities within learning for that thinking to occur.
After we reach a consensus of a model, we then deploy that model. We might develop an understanding of gas pressure, and now we want to determine if we can predict how gas pressure might fluctuate as volume, amount of gas, and temperature are altered. Students are tasked with making predictions about a new pressure and explaining the changes via the particle level using collisions.
Or students might have determined that there is a pattern where chemicals react in defined proportions. The teacher may use a BCA organizational tool to help students express these recipes. Students then use this model to predict the amount of chemical formed from a given amount of reactant. They can evaluate the percent yield, or predict the limiting and excess reagents.
Often these initial deployments involve no sample calculation from the teacher. This is based on interleaving which shows that when students must select the appropriate procedures to solve a problem that their learning is enhanced. The danger here is if students experience cognitive overload. Since this is rare in my classroom, my students obtain substantial benefits over those who are mimicking and abstract algorithm that they don’t understand (IE dimensional analysis allowing students to use units to solve something that they don’t have a conceptual development of).
When there are issues of novice students having an incomplete repertoire to evaluate the relevant mechanics of a problem scaffolding is implemented. Sometimes this involves having all students work on an identical problem in groups. In others this could involve scaffolding the sequence into smaller steps. However, most introductory chemical phenomena only involves the use of proportional reasoning as far as mathematics goes.
After an initial deployment, students reconvene with their analysis to discuss how the model help up. Were there issues with confusion or issues that require alteration to the model? This cycle is repeated until we reach an assessment. The assessments in my course are standards-based with the opportunity to reassess. This allows for spaced retrieval practice, but also sets an appropriate emotional framework for students to focus on learning. We are not limited by a fixed mindset approach that overemphasizes extrinsic motivators.
Feedback focuses on learning, is delivered in a timely manner, and is directed toward the class and not individual students. Highlights of powerful student thinking are displayed for students to emulate the metacognitive processes. Multiple representations are shown to highlight abstract ideas. Color coding can reinforce those overlaps.
Feedback works in both directions. By centering the classroom around student perception, this enable the teacher to be more effective. In “Understanding How We Learn” Weinstein, Sumeracki, and Caviglioli explore a section entitled “Is Intuition the Enemy of Teaching and Learning?”10 When the teacher can get students to authentically express their mental models, they are not going to fall victim to the curse of knowledge that many teachers suffer in traditional models of instruction.11 Instead the teacher is receiving ample feedback of what students do and do not yet understand. When this can be used to guide sequencing of instruction through standards-based grading, a powerful framework for instruction is built.
At the start of the class period a powerful tool I use is to have a student do a recap of the previous day’s lesson. They do this without resources which gives me an assessment of what they remember the most. Two years ago my students were overemphasizing the general form of reaction types. This year we made substitutions to emphasize the charge model within single replacement and double replacement reactions. This year students produced better samples and better identification of the smaller structure. These recaps utilize spacing, retrieval, and emphasize the students’ models.
Vocabulary is utilized as a retrieval cue and not as an abstract association. We avoid circular reasoning. For example, a student who does not know what a cello is might have the cello described as a large violent. If the student inquires as to what a violin is, they might be told it’s like a small cello. The student leaves without knowledge of a violin or a cello. But they do maintain an abstract association between the two. This is how most scientific vocabulary is treated within traditional instruction. A full treatment of concept first, vocabulary last is described here.12
Current improvements I’m working toward:
Currently I am working on having the students articulate the models being used. This is based off of the work of Brenda Royce who is assisting me on doing lesson plans where I articulate what scientific and conceptual models I expect students to be familiar with.13 These are produced during circle discussions where we analyze our whiteboards. Students will next articulate the models they are utilizing when they construct future whiteboards. Currently we are using particle level models and conceptual models which includes features such as equations, graphs, numerical patterns, and other abstract components.
How should a teacher learn more about implementing cognitive science in their classroom?
Teachers should read books about cognitive science, try new strategies/frameworks in their classroom, and network with other teachers who are working to utilize these methods. I know that I have learned a lot from Blake Harvard who writes about several creative applications of cognitive science within his AP psychology classroom. However, you might find these not to apply successfully to a middle school science classroom and that adjustments must be made.
The Learning Scientists have a substantial amount of free resources. I also appreciate the cognitive scientists who are hesitant to assign their conclusions to classroom practices as many of them are not teachers. I am immediately skeptical of anyone claiming that teachers only need to do 1-2 simple things in order to be highly effective teachers. Teaching is a complex skill that involves expertise in content, psychology, sociology, emotional engagement, attention, motivation, culture, and much more. When someone reduces that to a single component I doubt their expertise and/or motives.
Within your classroom you should instruct students on retrieval practice, spaced practice, and more. When students can articulate the distinctions between system 1 and system 2 learning via Kahneman they will have a better range of strategies to tackle learning. I find the Veritasium video on The Science of Thinking14 to be particularly helpful for students to distinguish system 1 and system 2. The video Kahn Academy and The Effectiveness of Science Videos15 is particularly helpful to show students how discrepant the feelings of comfort and learning can be.
As you improve your instructional methods and behaviors, it is critical to work toward better assessment simultaneously. Good instruction can be spoiled by review or assessment that undermines the teaching. The best resource I’ve found for healthy grading practices is “Grading For Equity” by Joe Feldman.
Quotes from Readings:
Make it Stick - Peter Brown, Henry L. Roediger III, Mark A. McDaniel
1. How Effort Helps (pg. 82)
The more effort that is required to recall a memory or execute a skill, provided that the effort succeeds, the more the act of recalling or executing benefits the learning.
2. Mental models (pg. 83)
Mental models are forms of deeply entrenched and highly efficient skills (seeing and unloading on a curve ball) or knowledge structures (a memorized sequence of chess moves) that, like habits, can be adapted and applied in varied circumstances.
3. Priming the mind for learning (pg. 86)
When you’re asked to struggle with solving a problem before being shown how to solve it, the subsequent solution is better learned and more durably remembered.
The act of trying to answer a question or attempting to solve a problem rather than being presented with the information or the solution is known as generation. Even if you’re being quizzed on material you’re familiar with, the simple act of filling in a blank has the effect of strengthening your memory of the material and your ability to recall it later. Overcoming these mild difficulties is a form of active learning, where students engage in higher-order thinking tasks rather than passively receiving knowledge conferred by others.
Out of the Labyrinth - Robert Kaplan & Ellen Kaplan
1. "If it looks like this, do that to it. Decorated with the elevated name of algorithm, but commonly called cookbook math, it relieves the student from any need for thinking and substitutes Truth by Authority for what could be dangerous encounters with reason." (pg. 117)
2. "Students may have a strong desire for immediate comprehension, which may ultimately be debilitating. If I don't get it right away, then I never will, and I say to hell with it." (pg. 196)
Culturally Responsive Teaching & The Brain - Zaretta Hammond
1. Our ability to process, store, and use information dictates whether we are able to do more complex and complicated thinking in the future because they are the very things that stimulate brain growth. It is precisely explicit information processing that is too often left off the equity agenda for low performing independent learners. (pg. 124)
2. We learn best when we are able to talk through our cognitive routine. Talking to learn, also called dialogic talk, is deeply rooted in oral cultural tradition. This kind of talk gives us the opportunity to organize our thinking into coherent utterances, hear how our thinking sounds out loud, listen to how others respond, and often, hear others add to or expand on our thinking. Tharp and Gallimore (1991) call this instructional conversation, the kind of talk that acts like a mental blender, mixing together new material with existing knowledge in a student’s schema. (pg. 134)
***Note how this quote compares with the information provided in Chatter by Ethan Kross where he describes how internal chatter moves at such a high rate that processing errors and reasoning can lead to ineffective model development!
The Inner Game of Tennis - Timothy Gallwey
I too admit to overteaching as a new pro, but one day when I was in a relaxed mood, I began saying less and noticing more. To my surprise, errors that I saw but didn’t mention were correcting themselves without the student ever knowing he had made them. (pg. 5)
Neuroteach - Glenn Whitman and Ian Kelleher
1. The upshot is, just because the common usage of the word “attention” is happening in your class, with lots of respectful and polite behavior, it doesn’t automatically follow that the neuroscience usage of the word “attention,” which is crucial for enduring learning, is happening throughout the room. (pg. 93)
2. Denise Pope, senior lecturer at Stanford’s Graduate School of Education, highlights a problem that she calls “doing school.”
These students explain that they are busy at what they call “doing school.” They realize they are caught in a system where achievement depends more on “doing” - going through the correct motions - than on learning and engaging with the curriculum. Instead of thinking deeply about the content of their courses and delving into projects and assignments, the students focus on managing the workload and honing strategies that will help them to achieve high grades. (p. 127)
Understanding How We Learn - Yana Weinstein, Megan Sumeracki, Oliver Caviglioli
There are two major problems that arise from a reliance on intuition. The first is that our intuitions can lead us to pick the wrong learning strategies. Second, once we land on a learning strategy, we tend to seek out “evidence” that favors the strategy we have picked, while ignoring evidence that refutes our intuitions. (pg. 23)
Citations
1. https://www.youtube.com/watch?v=hpQnRwVjhDk&list=PL3F629F73640F831D&index=37 @44:38 accessed 2/20/22
2. Mestre, J. P., & Docktor, J. L. (2021). The Science of Learning Physics: Cognitive Strategies for improving instruction. World Scientific Publishing Co. Pte. Ltd.
3. Problem-solving or explicit instruction: Which should go ... (n.d.). Retrieved February 20, 2022, from https://www.researchgate.net/publication/334982114_Problem-solving_or_Explicit_Instruction_Which_Should_Go_First_When_Element_Interactivity_Is_High
4. Scientists, L. (2021, February 18). The impact of Guided Discovery vs. didactic instruction on learning. The Learning Scientists. Retrieved February 20, 2022, from https://www.learningscientists.org/blog/2020/2/14-1
5. BROWN, P. E. T. E. R. C. (2018). Make it stick: The science of successful learning. BELKNAP HARVARD.
6. Modeling instruction: An effective model for science ... - ed. (n.d.). Retrieved February 20, 2022, from https://files.eric.ed.gov/fulltext/EJ851867.pdf
7. Cartier, J. L., Smith, M. S., Stein, M. K., & Ross, D. K. (2013). 5 Practices for orchestrating productive task-based discussions in science. National Council of Teachers of Mathematics.
8. Gallwey, W. T. (1974). The inner game of tennis. Random House.
9. *Dukerich, L. (2015). Applying modeling instruction to high school chemistry to improve students' conceptual understanding20. ACS Publications. Retrieved February 20, 2022, from https://pubs.acs.org/doi/abs/10.1021/ed500909w
10. Weinstein, Y., Sumeracki, M., & Caviglioli, O. (2019). Understanding how we learn: A visual guide. Routledge.
11. Harvard, B. (2022, January 26). Psychology in the classroom #2 - Curse of knowledge. The Effortful Educator. Retrieved February 20, 2022, from https://theeffortfuleducator.com/2022/01/11/cofk/
12. Milam, S. (2022, January 1). How strategic teaching with cognition (STC) shows why you should teach concepts first and vocabulary last. How Strategic Teaching with Cognition (STC) shows why you should teach concepts first and vocabulary last. Retrieved February 20, 2022, from http://ibchemmilam.blogspot.com/2022/02/how-strategic-teaching-with-cognition.html
13. https://www.youtube.com/watch?v=e3aTDHXi97w accessed 2/20/22
14. https://www.youtube.com/watch?v=UBVV8pch1dM accessed 2/20/22
15. https://www.youtube.com/watch?v=eVtCO84MDj8& accessed 2/20/22
