Buggy lab mid-teaching analysis
After we completed the Marshmallow Challenge, we jumped right into the Buggy Lab, which introduces students to small battery powered buggies. I’m going to try to give the play by play here and raise a few questions about my own practice.
I started by turning on the buggy and letting it roll across the table, and asked “What do you notice?” (Thanks Brian for teaching me about the power of this question).
Students quickly notices all sort fo things, from “it moves” to “it lights up” to the “the wheels are rotating at a constant rate.”
I then push my students to tell me what they could measure, and the most common response is “speed”, and so I say how do you measure speed, which they usually unpack as measuring a distance and the time it takes to travel that. After we talk about how to measure the distance and time, I usually set them loose.
Almost all of my students glom onto the following basic experiment—measure a fixed distance (say a meter) then release the cart and measure the time it takes the cart to travel that distance, and repeat this 3 times, then average, and here a few might call it a day.
So I go around and ask them if discovering the speed of the buggy is really all that exciting—would they want to tell their parents that’s what they learned on the very first day of school? They say no, and we talk about digging deeper to find a relationship.
I ask my students what sort of things have relationships, and we work our way through people (eg father-daughter), and then quantities, like circumference and diameter). I ask them how you know those things have a relationship, and they all said that if you graph circumference vs diameter, you’d see some sort of line. So we then decided our ultimate goal should be to determine the relationship between the distance traveled by the cart and time.
We talk about why they repeat trials—most of them seem to have think that is just something you do to avoid “human error.” So in the spirit of Brian Frank, I resist my temptation to simply ban the word human error completely and ask them what they mean by that—that gets them to tell me something about how it’s hard to press the stopwatch at exactly the same time as when you drop the car on the ground, so the measured time will always be different form the actual time, and averaging seems to account for sometimes you might be a little slow, and other times a little fast in pressing the buttons.
And so I ask them to visualize their 3 trial experiment would look like at this moment, and they all describe a graph that looks like this:
My students see pretty quickly that this doesn’t really allow one to determine the relationship between distance and time, and that what they really need to do is have the car travel a number of different distances and record the times it take so that you have data for a wide range of distances.
To get to this point, it’s taken most groups about 30 minutes. And we’ve still got lots of good points left for discussion for when groups come together, like whether we should be plotting distance or time on the horizontal, and what to do when one group measures in inches, one group measures in floor tiles, and one group measure in cm. When I ask the question “which cart is fastest?”, I am hoping they will suddenly see the need to standardization in terms of how we graph and measure in a more visceral way than if I had just told them what to do.
Now here’s where I have some questions.
Last night at the Global Physics Department, Kelly O’Shea and I got to chatting about the lab, and she was telling me that while everyone is together, she pushes them a bit further toward agreeing on how they will proceed. She draws up a graph with labeled axes and a data table on the board, and then everyone leaves to make measurements knowing they will be plotting distance in cm vs time in seconds. This is a huge time saver, for sure, and in many ways it saves kids a lot of frustration of having to go back and re-measure/re-graph things once they realize they did something wrong.
Ultimately, this is the age old question of just how much guidance we should provide—more and more I’m feeling I should give more guidance than I did yesterday, if for no other reason it will save me from having to have 6 individual conversations with each lab group and feel like I’m trying to push them to mold their experiment toward my thinking, rather than having them set up some sort of common foundation and then exploring on their own from there.
I would love any thoughts you may have on my approach, and suggestions for finding that balance between open-endedness that allows students to go down blind alleys that can lead to frustration, and guidance that keeps them on the path, but may prevent them from fully seeing the nuance of what they are doing.
Quantum Progress 
I think how to graph the data (x vs t, or t vs x) is too important to do before the students have “done” the lab. From my experience this summer, and thinking about it from a student’s perspective, they’re probably dying to get their hands on the buggy, and not really paying attention to you that much. I like what our leaders did, lead the students as you suggested up to the purpose (what do you notice, what can you measure, what can you manipulate, then purpose: to determine the graphical and mathematical relationship between position and elapsed time). After playing with the car for a few minutes, they’ll be ready to actually start the lab. They’ll look up at the purpose on the board and have an idea of what to do. During the board meeting, you can now have meaningful discussions about what the graph should look like (which variable was your groups independent/dependent?, which one should go on the x/y axes? What does the slope of graph represent {probably the most important question according to Arons!}), and the students will be engaged in the discussion (not dreaming about the car!).
Scott, I don’t think you’re giving the kids enough credit about being engaged before the lab. At least for my sophomores and juniors, most of them can function on a level above “Oooh, shiny”, even on the first day of class! 😉
More importantly, I don’t think it’s fair to think of it as getting to the “fun” part right away, then getting them to pay attention once it’s “boring” and the carts are gone. At the beginning of the year, when they have no idea what I’m expecting them to do yet, I think they want as much structure as possible before starting with the cart. They are used to pretty cookbook labs from chemistry (for my juniors, at least), and are expecting to know all of the “directions” (and honestly, also the results) before they make any measurements. I think one of the powerful things about the modeling approach to labs is that they have designed the whole experiment themselves (curiously, it ends up basically the same every year and in every section…!), so they know what to do and what to graph. They are the ones who came up with it about 5 minutes ago, after all!
OK, So I got two nods in this post, so I should give this a try. This is what I hear you saying:
First, you seem to want your students to learn about why (and when) standardization might be important–and you believe that can’t be done well if you tell them exactly what units to measure and how to graph before hand. You also seem to want the students to have some autonomy in their work to decide how they will measure and how many times. In your version of the lab, “purpose” with “autonomy” comes first, and the need for conventions across multiple autonomous groups comes during discussion.
Second, you also seem to want your students to measure multiple distance/durations, so they can look at relationships or trends, not just a single ratio. You need them to do this because you want to talk about interpreting slope as speed, which is a big deal, especially in modelling.
Both of these seem like good things. But I think, because the task doesn’t need students to do the second, it sets up a weird dynamic. Let me elaborate on that:
The public purpose of the lab–the one you developed with your students–was to take some measurements to figure out which buggy was fastest or to measure speed. BUT secretly you want them to do something else–you want your students to look at relationships using multiple data points. So there’s the “public task” and then there’s the “secret task”. Because of this, you might be having to run around to each of the six groups helping them to learn what the “secret task” is–the one the teacher really had in mind.
It could be that you need to provide more structure so that the secret purpose is more public, or it could be that the task itself doesn’t require students to do what you secretly want them to do. I’m sitting here thinking, what would lead a student to care to consider multiple data points? One kind of question would be, “Does the car have the same speed during the first half as it does during the second half?” This might not be something you want to do until later, but at least the task would press students to consider multiple points without it feeling like pulling teeth.
Most likely to me, as long as some groups were taking multiple data points, I wouldn’t change anything. If you are going to send them back after discussion because of standardization of units and graph type, then why not also discuss the difference between those groups who took multiple data points and those who only multiple trials at one point. Let the groups make mistake, do different things, and leverage that during discussion. You say in your post, “feeling like I’m trying to push them to mold their experiment toward my thinking”, when you might be better off helping them to better express their thinking and do work that will convey that thinking back to the whole class.
Aha… in my class, the “which cart is fastest?” question never comes up until the post-lab because they only see one cart moving during the pre-lab. They don’t know right away that the carts have different speeds. Our purpose is always “To find the relationship between position and time for the cart.” So, in sum, I think I’m agreeing with your secret vs public goal problem.
Then again, maybe my conversation is pretty different with juniors than it would be with freshmen. They almost always suggest that after we take some data, they should make a graph (this is during the discussion, before we split up). It’s not a stretch to have them decide then on using the same axes (though actually, we agree to use the metric system, but not on whether our measurements should be in cm or m).
Maybe another difference between what I do and what John does (not sure… John?) is that because a lot of our discussion is centered on how will we compare what we do when we’re done, we also discuss an origin. I have a taped down origin line that runs across the classroom, and I have 6 defined start positions with starting directions. We notice that no one is going to have their first point be at 0s, 0m because none of the start lines are on the origin. Since we’ve agreed to measure position rather than distance, that means that most groups are kept out of the measure a distance and divide by time thing (usually one group during the day will still do that, but then realize that they aren’t measuring what we said we would measure).
So this results in the groups having graphs that all have different slopes (not all with the same sign) and different intercepts. And then we’re ready to discuss what’s the same and what’s not the same about the results we got. So what is the defining characteristic of this motion and what changes depending on which cart you use. And then in that discussion we start to question who had the fastest cart.
Also, for the record, I like “What did you observe (about the motion, in this case)?” better than “What did you notice?” Observe connotes intent while notice connotes happenstance. Before they answer, I tell them that I’m not looking for any particular right or wrong answers, just that I want them to get their brains going and to start talking before we really get into it. That seems to work just fine.
Keep the modeling posts coming John. It’s really great to hear about what is actually happening in the moment.
Double down on that!…
Weighing in on the origin question. First, origin is a terribly confusing word for many students because they have seen it in the context of a graph before, and we’re drawing a graph for the experiment as well. But second, I have them do the lab twice. The first time, they tend just to measure distance and time because that’s what occurs to them. The slope of the graph is speed. Then I introduce two buggies of different speeds going in opposite directions, asking for observations (particularly of differences in the two motions) and then for how to take measurements so that when graphed all the differences in the motions show up. The idea is for the task to motivate the need for the concept of position vs. distance, and to have a clear comparison. After the second lab, I ask what the distance vs. time graph for the two buggies would look like, and then what the change in position vs. time graph would look like. This is a great task that really reveals whether they’ve understood the definitions and their implications.
I haven’t thought about this before, but I start out with a bouncing ball lab (drop height vs. bounce height), and its not so obvious that the trend will be linear. In fact for large drop heights (such as from our balcony for the lab practicum), some types of ball start to deviate from the linear trend. So there’s a reason to take multiple data points and check the trend. This is also a good initial lesson about models and their limited range of validity. This is followed up with a sliding puck lab (launched by a rubber band) that produces curved graphs – squared and inverse. Now it’s really apparent that the trend needs to be checked with multiple points.
You guys are freaking amazing. So many excellent comments and ideas hours after I throw out these questions. I’m going to be able to put these ideas into practice tomorrow when we finish up the lab, and my class will be immeasurably better thanks to you. (Kelly, my floor is now all taped up with masking tape).
I will definitely keep blogging about modeling, and would consider myself extremely fortunate to get more feedback like this.
Another idea…
Instead of letting student take their time initially, which presumably is what leads them to do three trials and average their results, you could impose an initial time limit for their measuring – say 5 minutes – or, alternatively, ask them to only do one “quick” measurement. Try to ensure that various groups use differing distances for their trial.
Once every group has a first measure, ask them to each post their results on the board. The attempt to compare results will hopefully lead to a student-led discussion as to what units are going to be best for all to use, and what variable to graph on what axis, and how the results can/will be compared.
Then send groups back to work, this time to come up with their “best” measurement, using repeated trials if necessary.
By having different groups measure time over differing distances – that can help determine if the speed of the buggy is changing over time (assuming differing groups use the same buggy). If each group has their own buggy, then you could ask each group to conduct their trials using two different distances.
By having a quick “first round”, you start the get students used to the process of using a first pass to get the lay of the land, figure out what works and what does not, think about how the results will be used, and think about sources of error – all without spending too much time on the actually gathering the initial data. Then, on the second round, the students will “know where they are trying to go” with the data, can work to deliberately reduce possible sources of error, and can get the bulk of their measurements and calculations “right” the first time.
This same first-pass followed by one or more run-throughs happens when modeling, coming up with proofs, writing/editing an English paper, creating a work of art in a new medium, etc. – so it is a valuable process to familiarize students with. A satisfactory result is almost never achieved on the first pass!
So, keep the time “wasted” on the first pass short – to maximize the lessons learned while minimizing student frustration at not being able to use initial results in the final presentation. Then provide plenty of time for students to discuss and compare first-pass results, gain insights, then go back and, armed with their new-found insights, have the bulk of their time spent on the project be productive.
Dorrie,
My students did comment on how the video is nothing but the guy succeeding, and how much footage of him failing must be on the cutting room floor. I plan on weaving a metacognative thread through the entire year—and I try to revisit some of these topics raised in this discussion specifically, but I still think there is much I can do to improve this thread in my class.
I have loved this post…particularly the comments. Thanks for all you folks do to help me learn – I don’t teach physics, per se, but I do facilitate learning for young and older learners. This is such a rich meta-discussion, too. Doug Fisher and Nancy Frey write a lot about “gradual release of responsibility” in the teacher-student relationship. Their work may be interesting to you, John, but you may already know of it.
Bill,
Thanks for the thoughtful and detailed response. I’d be inclined to agree with most of what you said, but students are on facebook now—I’d say 90%+ of my students use it every day. If adults who care about students don’t begin to guide students in how to properly use these tools, who will? The peer culture? Plus facebook has many tools and settings for preserving privacy of both students and teachers; they just need to be instructed and reminded to use them. I definitely think parents should be a part of the conversation as well.
John,
Thank you so much for your post…last year I was like you trying to go to 8 different groups and get them to where I “secretly” wanted them to be. Thanks to the comments, I figure I just need to relax and let my kids do their thing…. I do love the extension part that Matt brought to light…Usually I give each group a different scenario for the buggy’s (go in the other direction, start behind the origin and go through it, run into the wall and flip over and go the other way, etc)
The origin conversation can easily stretch out for a period or two before consensus. It’s very confusing, especially for my regents students.
I know that this is a post that is over a year old, but I was wondering if anyone has performed this lab using Vernier probeware? Is there some reason not to? Is value placed on students arduously collecting and graphing the data before they can start to analyze their graphs?
I’m going to try this lab for the first time tomorrow. I’m new to modeling – a member but have not had the time to make it to a workshop though I desperately want to!
Using the Vernier motion detector, students can collect and graph data real time! Maybe 15 – 20 minutes depending on their dexterity/focus. They can whiteboard their data quickly and the class can shift to the boardroom process.
To keep things interesting, I like for the groups to all be running different “experiments.” I set up the four following scenarios:
1) Car in front of motion sensor moving away
2) Car 1.5 m away from sensor, moving towards sensor
3) Car 0.25 m away from sensor, moving away
4) Car 0.5 m away from sensor, moving away
I wish I could do a second trial moving towards the sensor, but the sensor isn’t very sensitive to small objects at a range of more than 1.5 m. This is the kind of situation where collecting/graphing by hand would be “handy.”
This is generally the first lab of the year in my physics class, and I prefer to start with measurement tools that are easier to understand—working with a motion sensor can be pretty black-box for a first lab. Later, when we study acceleration, and need more precise measurements we use the motion sensor with dynamics carts and tracks.