This semester I am taking a new approach to teaching my Classical Mechanics I course. In the past this course has been entirely lecture-based, although I have a lot of one-on-one interaction with my students outside of class because it is a class taken only by physics majors (and there aren't that many of them). I have also used computational problems to get them some hands-on computing experience while they learn classical mechanics (hopefully a paper describing this aspect of the class will soon be published in the American Journal of Physics). But this semester I am mostly abandoning lecture. Only about one fourth of our class meetings will be lectures, and these will be devoted almost exclusively to computational demonstrations (where I show them how to use computation to analyze the dynamics of classical systems). At least part of these "lectures", which are given in the computer lab, are given over to the students running some computations themselves.
For another fourth of our class meetings I am using a few of the Intermediate Mechanics Tutorials developed by Brad Ambrose and Michael Wittman. Tutorials are worksheets that lead groups of students (my students work in pairs) through a series of questions that challenge misconceptions and elicit insights into important physical principles. These tutorials are then followed up with homework assignments that are designed to test student understanding of the tutorial material and build on the insights gained in the tutorials. Typically a pretest is given before the tutorial. I have found that we often need the full 75 minutes of the class period to complete the tutorials, even though they are allegedly designed for a 50 minute class.
The remaining half of the classes are devoted to student presentation of homework problems. This is a take-off on the Moore Method, which was a method for teaching mathematics developed by R. L. Moore at the University of Texas. The strict Moore Method forbids students from working together or using textbooks - they must prove theorems on their own and present their proofs in class. I deviate heavily from this strict version, as my students definitely use their textbook and can discuss the problems with each other (but not copy). Still, the focus of the class is the students presenting their solutions. Other students are expected to critique the solutions that are presented. The presentations are evaluated on content as well as on quality of presentation. Each assigned problem is presented by a student and all students must turn in written solutions for every problem. I've been guided by all of this by my fellow physicist Chuck Lane who pioneered this approach in our department last year.
I'm only halfway through the semester, but so far I would report both great success and some disheartening results. The disheartening thing was that I started with 11 students enrolled in the class but 5 of these withdrew because they thought the workload would be too great. However, the remaining 6 are doing quite well. I was very impressed with the performance of the class on the first exam. The homework presentations have improved steadily, both in terms of correctness and style. This has also led to similar improvements in the written homework (which has not generally been the case when I've taught this course in the past). The tutorials have also been great - I plan to use them more systematically the next time I teach the course. And I'm still doing the computational problems, which seem to be going pretty well.
The mixture of tutorials and homework presentations seems to give a nice balance to the class. There isn't an adversarial environment, because a lot of the time they are working together on things. One problem is that it is hard to get the students to really critique each other - they think they are going to hurt someone's feelings. And I am probably a bit to willing to jump in and give my own critique if the other students aren't willing. But they (and I) are getting better about this. I try to prod them when I can tell they want to say something. I periodically remind them that they aren't helping their friends by failing to point our errors in their solution. I know its wrong, so their presentation grade will still suffer, and if the errors aren't pointed out then their written solution will likely get a low grade as well. Pointing out the error will help them fix it and avoid a lower grade. They aren't totally convinced, but I'll keep working on them.
In summary, it's been a very experimental semester but it seems to be going very well. I will tinker with the balance of presentations, tutorials, and lectures but I will probably not get rid of any of the three (although I feel I definitely won't increase the lectures).
Friday, October 26, 2007
Sunday, October 7, 2007
PHY 111/112: General Physics I and II with Algebra
In my General Physics courses I use a mixture of Peer Instruction (a teaching method developed by Eric Mazur) and Just-In-Time Teaching (developed by Novak, Patterson, Gavrin, and Christian) with Physlets (Java applets with physics content developed by Wolfgang Christian and Mario Belloni). Prior to each class meeting students are expected to read the section of the text that will be discussed in the upcoming class. In addition, they must complete an online quiz, the deadline for which is about 2 hours prior to the class meeting time. The online quiz usually consists of 5 questions. The first two questions are generally somewhat challenging questions that cover material from the previous class meeting. Often these are Physlet Problems (problems in which students must interact with a Java applet to obtain the information necessary to solve the problem). Using Physlets helps students to connect their physics knowledge to real-world situations (as simulated by the Physlet) and also helps them to develop an operational understanding of physics terms (since they must frequently make "measurements" with the Physlet to obtain the needed information). The third question on the online quiz is usually taken directly from the reading assignment from the textbook. This question is designed to determine whether or not students have completed the assigned reading. The final two questions deal with material that has not yet been covered in class, but is slated to be covered in the upcoming class. These questions typically involve Physlet Illustrations or Explorations, which allow students to interact with a Java applet to develop their knowledge of a new concept. To answer the questions correctly the student must make use of the Physlet and pay close attention to what they see, but they do not generally have to already possess a solid understanding of the relevant concepts. These last two questions provide a lead into the class discussions, as well as an illustration of what the student has read in the text.
During the hour or so prior to each class I review the student responses to the online quiz questions. This allows me to determine which questions the class as a whole seemed to understand, and which questions gave the class trouble. This helps me to target the discussion in the upcoming class to focus on those concepts that are causing the greatest difficulty for the most students. I typically start class by discussing one of the Physlet Problems from the online quiz, and then I give a short lecture in which I present the key concepts that will be the focus of the day's class. I often use a Physlet (shown to the whole class using a computer projection system) to illustrate key points. Sometimes the Physlet will be one that students used to answer one of the last two online quiz questions, sometimes it will be one they have not seen before.
This part of the class usually last about 20 minutes, but sometimes takes longer.
The next 20-25 minutes is spent doing Peer Instruction. Students are asked a series of conceptual multiple-choice questions. They respond to these questions using an electronic response system. I encourage students to discuss the questions with their neighbors, and I also walk around the room and discuss the questions with students. Once the entire class has responded to a particular questions I display a histogram of the responses on the screen. If a significant majority chose the correct answer we usually move on after a brief statement from me about why that answer is correct. If a significant fraction of the class has answered incorrectly, I ask students to try again and usually provide a few hints about how to approach the question. We usually cover about 4-5 questions in this way.
If there is time remaining in the class I usually demonstrate the solution of a calculation-oriented problem. Most of the time I show them how to set up the required calculations, but I don't work out the solutions all the way. However, I do post a fully-worked solution to this demo problem immediately after class so that students can review this demo before attempting their own homework.
Homework assignments consist of calculation-oriented problems of my own devising, with a few conceptual questions mixed in (but not designated as conceptual - they are not readily distinguishable from calculation-oriented problems at a glance). A typical assignment consists of two questions, each of which has multiple parts. The idea is to have students examine a single scenario and, on the basis of a little bit of information, answer a series of questions related to that scenario. These homework assignments are due at the beginning of the next class period. The homework is graded during the class meeting by a student grader (almost always a physics major). The grading is essentially binary for each part of a question, but I have developed a very lenient scoring system that is designed to encourage students to attempt every part of each question. If a student tries each part of each question they can score no lower than 7 out of 10 (with scores increasing from there on the basis of correctness). Students can then pick up their graded homework as they leave class, which provides them with immediate feedback on their work. I post homework solutions online once the homework has been graded, so students have the opportunity to look at my solutions to try to understand any questions they missed.
Labs for the class focus on improving student understanding of physics concepts. Very few of the labs are about verifying the physical principles presented in class. Instead they provide an opportunity to for students to work extensively with the principles in order to develop a deeper understanding of what they mean and how they apply in various situations. The goal is to help students internalize these concepts.
Tests are given during the lab period so that students have two hours to complete the test. The tests include both conceptual and calculation problems (typically about 40% conceptual, 60% calculation - although I emphasize to students that calculation problems are ALSO conceptual). The goal of the tests is to assess students' conceptual understanding of physics as well as their ability to apply the concepts in solving problems. Almost all questions have a specific correct answer, although occasionally students are asked to estimate.
I typically have 40 students in the General Physics I course, and about 25 in the General Physics II course. Lab sections have a maximum of 20 students. Both classes typically meet three times a week for 50 minutes each, with one two hour lab period. Most of the students in these classes are biology or animal science majors and a large number of these students plan to go on to professional schools (medicine, veterinary medicine, physical therapy, dentistry, pharmacy, etc.).
I use the Force Concept Inventory (FCI) to assess the effectiveness of my teaching in General Physics I. I was initially seeing pre-test to post-test normalized gains of about 35-40% (well above what is typical for traditional lecture courses, which tend to show normalized gains around 25%) using the method described above, but without the Physlets. Last year I used Physlets for the first time and my FCI normalized gains increased to 47% (basically double the gains seen in traditional lecture courses). I also receive reports from many students that they do very well on the physical science section of the MCAT. I have also heard reports from students who have taken MCAT review courses and say that they already know all the physics, while students who have taken physics at other institutions don't know any of it. So while there is still room for improvement, I have reason to believe that the teaching methods described above are effective in this case.
During the hour or so prior to each class I review the student responses to the online quiz questions. This allows me to determine which questions the class as a whole seemed to understand, and which questions gave the class trouble. This helps me to target the discussion in the upcoming class to focus on those concepts that are causing the greatest difficulty for the most students. I typically start class by discussing one of the Physlet Problems from the online quiz, and then I give a short lecture in which I present the key concepts that will be the focus of the day's class. I often use a Physlet (shown to the whole class using a computer projection system) to illustrate key points. Sometimes the Physlet will be one that students used to answer one of the last two online quiz questions, sometimes it will be one they have not seen before.
This part of the class usually last about 20 minutes, but sometimes takes longer.
The next 20-25 minutes is spent doing Peer Instruction. Students are asked a series of conceptual multiple-choice questions. They respond to these questions using an electronic response system. I encourage students to discuss the questions with their neighbors, and I also walk around the room and discuss the questions with students. Once the entire class has responded to a particular questions I display a histogram of the responses on the screen. If a significant majority chose the correct answer we usually move on after a brief statement from me about why that answer is correct. If a significant fraction of the class has answered incorrectly, I ask students to try again and usually provide a few hints about how to approach the question. We usually cover about 4-5 questions in this way.
If there is time remaining in the class I usually demonstrate the solution of a calculation-oriented problem. Most of the time I show them how to set up the required calculations, but I don't work out the solutions all the way. However, I do post a fully-worked solution to this demo problem immediately after class so that students can review this demo before attempting their own homework.
Homework assignments consist of calculation-oriented problems of my own devising, with a few conceptual questions mixed in (but not designated as conceptual - they are not readily distinguishable from calculation-oriented problems at a glance). A typical assignment consists of two questions, each of which has multiple parts. The idea is to have students examine a single scenario and, on the basis of a little bit of information, answer a series of questions related to that scenario. These homework assignments are due at the beginning of the next class period. The homework is graded during the class meeting by a student grader (almost always a physics major). The grading is essentially binary for each part of a question, but I have developed a very lenient scoring system that is designed to encourage students to attempt every part of each question. If a student tries each part of each question they can score no lower than 7 out of 10 (with scores increasing from there on the basis of correctness). Students can then pick up their graded homework as they leave class, which provides them with immediate feedback on their work. I post homework solutions online once the homework has been graded, so students have the opportunity to look at my solutions to try to understand any questions they missed.
Labs for the class focus on improving student understanding of physics concepts. Very few of the labs are about verifying the physical principles presented in class. Instead they provide an opportunity to for students to work extensively with the principles in order to develop a deeper understanding of what they mean and how they apply in various situations. The goal is to help students internalize these concepts.
Tests are given during the lab period so that students have two hours to complete the test. The tests include both conceptual and calculation problems (typically about 40% conceptual, 60% calculation - although I emphasize to students that calculation problems are ALSO conceptual). The goal of the tests is to assess students' conceptual understanding of physics as well as their ability to apply the concepts in solving problems. Almost all questions have a specific correct answer, although occasionally students are asked to estimate.
I typically have 40 students in the General Physics I course, and about 25 in the General Physics II course. Lab sections have a maximum of 20 students. Both classes typically meet three times a week for 50 minutes each, with one two hour lab period. Most of the students in these classes are biology or animal science majors and a large number of these students plan to go on to professional schools (medicine, veterinary medicine, physical therapy, dentistry, pharmacy, etc.).
I use the Force Concept Inventory (FCI) to assess the effectiveness of my teaching in General Physics I. I was initially seeing pre-test to post-test normalized gains of about 35-40% (well above what is typical for traditional lecture courses, which tend to show normalized gains around 25%) using the method described above, but without the Physlets. Last year I used Physlets for the first time and my FCI normalized gains increased to 47% (basically double the gains seen in traditional lecture courses). I also receive reports from many students that they do very well on the physical science section of the MCAT. I have also heard reports from students who have taken MCAT review courses and say that they already know all the physics, while students who have taken physics at other institutions don't know any of it. So while there is still room for improvement, I have reason to believe that the teaching methods described above are effective in this case.
Subscribe to:
Posts (Atom)