Welcome
Thank you for your interest in the study.
What's involved
In this ~30-minute online study (which you can complete on your own, at a time of your choosing), you will learn about an interesting topic related to the How People Learn course, complete a brief activity, and/or answer a few questions about your understanding of the material and the experience in the study.
Who can participate
Any Ed.M. student currently enrolled in the Summer 2026 How People Learn course at HGSE.
Note
This study is best completed on a computer in a quiet setting where you can focus without interruptions. Participation in this study is completely voluntary and will have no impact on your course standing or grades.
If this all sounds good and you would like to take part in the study, press Next to continue.
Add your email address
So we can send you a short summary of the study findings once the study is complete.
Informed Consent
How This Study Works
In this study, it is very important that you read everything carefully and make sure you understand before moving on.
Reading Timer
On some pages, a reading timer will appear. You cannot click Next until the timer has finished counting down. This ensures you have enough time to read and understand the material.
The timer sets a minimum, not a maximum. Once it finishes, you are welcome to keep reading and engaging with the page for as long as you'd like before clicking Next. Take the time you need.
Timer running
Next not yet available
The blue ring empties as time passes. Read carefully while you wait — the timer cannot be skipped.
Timer complete
You may now continue
The ring turns green. A checkbox and the Next button unlock — check the box, then click Next whenever you are ready to proceed.
To Continue to the Next Page
Wait for the timer to finish
Read the page carefully while the ring counts down.
Check the confirmation box
A checkbox appears once the timer finishes. Check it to confirm you have read the page once you are done.
Click Next
The Next button activates after you check the box. Click it to move on whenever you are ready to proceed.
Please read carefully before you begin
Read the lesson carefully. We will ask you some questions at the end about what you've learned.
Thought Experiments and Children's Belief Revision
An important goal of education is to help learners recognize when their beliefs are inaccurate or incomplete and revise them accordingly. One tool that may support belief revision is the use of thought experiments. Thought experiments ask learners to imagine a possible situation and reason through what should follow if their current explanation were true. This can make hidden inconsistencies in an explanation easier to notice, because the learner has to think about what the explanation actually predicts. In this way, thought experiments are useful not because they add new information, but because they help the learner see whether their own ideas hold together.
Coherent beliefs are especially impactful in childhood. By the time children enter school, they are not blank slates. They already have intuitive explanations, referred to as naïve theories, that shape how they predict outcomes and make sense of what causes what in the world. However, these naïve theories are often incomplete, inaccurate, or internally inconsistent, and they may persist even when children encounter evidence that directly refutes what those theories predict. Incorrect naïve theories can make learning especially challenging. Some kinds of learning simply involve enrichment (adding new information to what is already known). However, when belief revision involves radically updating these intuitive theories, it is referred to as conceptual change.
Children rarely commit to a single explanation outright. Instead, a child often holds several competing explanations at the same time, believing each one to a different degree. On this view, revising a belief means adjusting those degrees of belief in each explanation as new evidence comes in: the child shifts how much "belief weight" they give to each explanation they already hold. When evidence favors one explanation, the child becomes more confident in it, while still keeping the others.
Revising a belief replaces one explanation with another (referred to as conceptual replacement). When the child revises their beliefs, the explanation that best fits the evidence overwrites the weaker explanation, so that the weaker explanation is no longer available to the child. What changes in conceptual replacement is to which explanation the child has access.
A clear case of competing explanations can be seen in children's reasoning about motion. A child may hold one explanation consistent with Newtonian mechanics, according to which a moving object continues moving at the same speed unless an outside force acts on it, and at the same time an impetus-based explanation, according to which a moving object carries an inner push that gradually fades, so the object slows down on its own. Consider a ball rolling through empty space with nothing touching it. The impetus explanation predicts the ball will slow and stop; the Newtonian explanation predicts it will keep moving. The child can hold both beliefs and move between them depending on the situation, without noticing that the two predictions disagree.
This is where thought experiments can help — they may support consistency monitoring. Asking the child to simulate the ball and trace what each explanation predicts brings both predictions into view at once and forces the contradiction into the open.
Ask any questions you have about the lesson
The lesson is on your left for reference. You may ask the chatbot on the right as many questions as you would like. When you are finished, click the button at the bottom to continue.
Lesson · For reference
Thought Experiments and Children's Belief Revision
An important goal of education is to help learners recognize when their beliefs are inaccurate or incomplete and revise them accordingly. One tool that may support belief revision is the use of thought experiments. Thought experiments ask learners to imagine a possible situation and reason through what should follow if their current explanation were true. This can make hidden inconsistencies in an explanation easier to notice, because the learner has to think about what the explanation actually predicts. In this way, thought experiments are useful not because they add new information, but because they help the learner see whether their own ideas hold together.
Coherent beliefs are especially impactful in childhood. By the time children enter school, they are not blank slates. They already have intuitive explanations, referred to as naïve theories, that shape how they predict outcomes and make sense of what causes what in the world. However, these naïve theories are often incomplete, inaccurate, or internally inconsistent, and they may persist even when children encounter evidence that directly refutes what those theories predict. Incorrect naïve theories can make learning especially challenging. Some kinds of learning simply involve enrichment (adding new information to what is already known). However, when belief revision involves radically updating these intuitive theories, it is referred to as conceptual change.
Children rarely commit to a single explanation outright. Instead, a child often holds several competing explanations at the same time, believing each one to a different degree. On this view, revising a belief means adjusting those degrees of belief in each explanation as new evidence comes in: the child shifts how much "belief weight" they give to each explanation they already hold. When evidence favors one explanation, the child becomes more confident in it, while still keeping the others.
Revising a belief replaces one explanation with another (referred to as conceptual replacement). When the child revises their beliefs, the explanation that best fits the evidence overwrites the weaker explanation, so that the weaker explanation is no longer available to the child. What changes in conceptual replacement is to which explanation the child has access.
A clear case of competing explanations can be seen in children's reasoning about motion. A child may hold one explanation consistent with Newtonian mechanics, according to which a moving object continues moving at the same speed unless an outside force acts on it, and at the same time an impetus-based explanation, according to which a moving object carries an inner push that gradually fades, so the object slows down on its own. Consider a ball rolling through empty space with nothing touching it. The impetus explanation predicts the ball will slow and stop; the Newtonian explanation predicts it will keep moving. The child can hold both beliefs and move between them depending on the situation, without noticing that the two predictions disagree.
This is where thought experiments can help — they may support consistency monitoring. Asking the child to simulate the ball and trace what each explanation predicts brings both predictions into view at once and forces the contradiction into the open.
A few questions about what you read
We are going to ask you 16 questions about the lesson you just read.
Please answer each question in your own words.
We do not expect you to remember everything, but please do your best and answer as fully as you can.
Question goes here
How was the lesson?
The following questions are about your experience with the lesson. Please indicate how much you agree or disagree with each statement.
How was the chatbot?
The following questions are about your experience with the chatbot.
One more question about the chatbot
Based on what you were told at the start of the study, who will be able to see the conversations you had with the chatbot? (Select all that apply.)
Five short questions
We are going to ask you some more questions. Try to give your best answer, or if a question seems tricky, try your best guess.
Problem text goes here
About your TF in How People Learn
The following statements are about your Teaching Fellow (TF). There are no right or wrong answers. Your responses are confidential and will not be shared with your TF or anyone on the course teaching team.
Before you finish
Your responses are confidential and will not be linked to your identity. If you prefer not to answer a question, you may select "Prefer not to say."
About the Lesson You Just Read
Thank you for participating. This lesson builds on a long-standing interest of cognitive science in how learners revise their beliefs when their existing explanations fall short. This lesson was intentionally written to be somewhat confusing (including providing potentially contradicting information about whether competing beliefs are overwritten during learning or maintained but believed to a lesser extent), in order to encourage participants to ask clarifying questions. (You may have been in a baseline condition where there was no mechanism to ask questions.)
This lesson introduced key ideas in cognitive development (though note that experts can disagree about these definitions). Here, we use belief revision to capture any kind of learning that involves updating beliefs, and conceptual change captures cases when those beliefs are a causal model of the world or an intuitive theory that gets radically revised. Consistency monitoring is when a learner holds two beliefs that lead to contradictory predictions about the same situation and notices that both beliefs cannot be true. Enrichment is a type of belief revision when a learner adds new information that fits with what they already understand, without needing to overwrite their original explanation. Conceptual replacement is the idea that a new belief completely writes over an old one (that the original revised belief is no longer accessible). Thought experiments support learning by asking the student to imagine a situation and trace what each explanation would predict, which makes hidden contradictions easier to notice.
If you would like to learn more
- Bascandziev, I., Abutto, A., Walker, C. M., & Bonawitz, E. (2025). Mind over matter: consistency monitoring and domain-specific learning. Frontiers in Developmental Psychology, 3. https://doi.org/10.3389/fdpys.2025.1496651
- Bascandziev, I. (2024). Thought Experiments as an Error Detection and Correction Tool. Cognitive Science, 48(1), e13401. https://doi.org/10.1111/cogs.13401
- Bascandziev, I., & Carey, S. (2022). Young children learn equally from real and thought experiments. In J. Culbertson, A. Perfors, H. Rabagliati, & V. Ramenzoni (Eds.), Proceedings of the 44th Annual Conference of the Cognitive Science Society (pp. 142–148). Cognitive Science Society.
- Bascandziev, I., Tardiff, N., Zaitchik, D., & Carey, S. (2018). The role of domain-general cognitive resources in children's construction of a vitalist theory of biology. Cognitive Psychology, 104, 1–28. https://doi.org/10.1016/j.cogpsych.2018.03.002
Thank you
Thank you for completing this study. In this study, we are interested in whether having access to a chatbot helps students learn challenging academic material better and whether the context in which students use that chatbot changes how they interact with it.
To test this, we randomly assigned participants to one of three conditions: (1) learning the lesson material and answering questions on their own, (2) having access to a chatbot before answering questions, or (3) having access to a chatbot before answering questions but being told that their conversations would be shared with their instructor.
While the study is related to the How People Learn course, your participation has no effect on your grades, course standing, or relationship with the HPL teaching team. No one involved in the course will know whether you participated, which condition you were assigned to, or how you performed.
The research team, and only the research team, will be able to read conversations with the chatbot, but they will not be able to tell which participant any conversation belongs to.
Because this study is still ongoing and other students may not yet have participated, we ask that you please do not discuss the details of the study design or what you learned in this debrief with other students in the course. If someone asks about the study, you are welcome to say that it involves reading a lesson and answering questions about it.
If you have any concerns about this study, you can contact Blerim Jashari at bjashari@g.harvard.edu.
You've already completed this study
Thank you for participating! Our records show that you have already submitted your responses for this study, so there's nothing more for you to do here.
If you believe this is a mistake, please reach out to the researcher, Blerim Jashari, at bjashari@g.harvard.edu.
You're all done
Thank you for participating in this study. You may now close this window.
Saving your responses…
You will be entered into a drawing for either a Gutman Café or Amazon gift card voucher, and notified via email if you are selected.
If you have any questions about the study, please reach out to the researcher, Blerim Jashari, at bjashari@g.harvard.edu.