A Handbook for Mathematics Teaching Assistants

(Preliminary Edition)

Tom Rishel, Cornell University & The MAA

  Contents

  1. Introduction

  2. Types of TA Assignments: Recitation, Lecture, Grading

  3. Before You Teach: A Checklist

  4. Day One

  5. What Goes On in Recitation?

  6. What Should be on a Syllabus?

  7. Lesson Planning: Survivalist Tactics

  8. Grading Issues

  9. Cooperative Learning

  10. Technology

  11. Writing Assignments

  12. Making Up Exams and Quizzes

  13. Using Cognitive Models to Make Appropriate Problems(with Mary Ann Malinchak Rishel)

  14. The Active Classroom

  15. "What Was That Question Again?"

  16. Motivating Students

  17. How to Solve It

  18. Course Evaluations

  19. Get Along with Colleagues

  20. What is a Professional?

  21. Teaching Methods for Various Types of Classrooms

  22. Problems of and with Students

  23. Student Types: Who is the Audience?

  24. How to Get Fired

  25. Advice to International TAs

  26. Silly Stuff...

  27. ...And Not So Silly Stuff

  28. The Semester in Five Minutes

  29. Jobs, Jobs, Jobs

  30. Letters of Recommendation

  31. Mathematical Talks

  32. Becoming a Faculty Member

  33. University and College Governance

  34. What Does an Evaluator Evaluate?

  35. The Essence of Good Teaching

  36. Case Studies

 

Introduction

This is a text about teaching college mathematics.

My view is personal, informed by over forty years in higher education, over thirty of them teaching in some form or another, and almost twenty of those involved with training and evaluating teaching assistants and junior faculty. If I seem to emphasize first-and second-person narrative in my writing, it is because much of this text has come, literally, from discussion with you, the TA or junior faculty member, about the real world situations we are encountering daily in our classrooms.

At most every juncture in the text, I emphasize nuts and bolts considerations over theory. This is not because I believe that theory does not exist or is not important, but because I think that good teaching starts with seeming trivialities--"talk loudly, write large, prepare carefully, explain a lot, be friendly." Only after we are familiar with such simplicities do we begin to feel comfortable moving into theories of learning. This last is not to say that such theories are never useful or important--otherwise, Mary Ann Malinchak Rishel and I would not have written the long section on how using cognitive methods can lead to better examinations, for instance. However, I do think that you, as a graduate TA or a young faculty member, will profit more and improve faster from short, simple, clear suggestions that have immediate and obvious impact in your day-to-day classroom. If this improvement leads you to decide that you want to think more deeply about your current and future teaching, so much the better. Send me e-mail so we can talk trishel@maa.org.

Finally, let me address a very common view about the discipline of teaching; namely, as I was told again just last week, "Teaching can't be taught." Well, maybe, just maybe, great teaching is lightning in a bottle and can't be explained, but I claim emphatically that good teaching can be taught. Of course, I am biased in my view, if only because I have spent the last twenty years (roughly) trying to achieve this aim. But, in fact, I believe not only that teaching can be taught, but that if mathematics is to progress, it must be taught -- to the bright young people who will carry it on after us. I hope that, by the end of this volume, you will agree with me.

So, let's stop talking and get to work...

Return to Top

Types of TA Assignments: Recitation, Lecture, Grading

Most teaching assignments for graduate students fall into one of three categories listed in the title of this section.

Probably the most common TA assignment in mathematics, and the one with which the majority of the faculty began their careers, is that of recitation instructor. Those of you who have received an undergraduate degree from a large university will be familiar with the lecture-recitation format: a faculty member lectures to a large class of students two or three times a week on an assigned topic from a textbook, after which a graduate student answers questions about the lecture and discusses assigned homework problems. In this format, the lecturer decides which homework to assign, and often determines the structure of the recitation. By this I mean he or she may say: "Don't do all the problems; just the ones that are designated not to be turned in for grading." Alternatively, the lecturer may suggest that you begin each recitation with a couple of "example problems." Generally, however, most instructors will give you little or no advice, except to say something like: "Just do a standard recitation." (For a sample "standard recitation," whatever that may be, see the later section, What Goes On in Recitation.)

Another common assignment for TAs is to be asked to lecture. Schools vary as to when in a graduate student's career this is to be done; at some institutions you are handed an algebra and trigonometry text and told, "Go teach this. Don't mess up!" Other schools wait for a year or two until you have had some less demanding assignments before they ask you to plan lessons, make up your own exams, determine grading policy, and generally deal with the problems of teaching undermotivated freshmen (or worse, undermotivated seniors!) the joys of precalculus.

It is probably worth pointing out here that at some point in your graduate career you should pursue a lecturing assignment, for two basic reasons:

1. A graduate student who has lectured has a real advantage in the job market (see the section, Jobs, Jobs, Jobs).
2. By lecturing before you take a first faculty position, you remove some of the stress over teaching that goes into the tenure-pressure.

A third common TA assignment is that of grading, sometimes in an elementary course, more often in an advanced undergraduate or even a graduate course. Many TAs describe such assignments as "easy" or "boring." While the assignments can be either or both, grading jobs, however, can teach you how far you have come since the days when this coursematerial was a real effort. These assignments can also show you how hard it is to teach others to write clear, concise answers and proofs. A third benefit to a grading job is that you can use it to review the material that may be asked on a graduate comprehensive examination. I will say more about the questions involved in grading papers later on in the section titled Grading Issues.

For now, think about:

Which type of TA assignment appeals to you most now? Is there one that you might never want to do? Do you think that your opinions might change later on in your career, or are they set in stone?

Return to Top

Before You Teach: A Checklist

Return to Top

Day One

It is fitting that I begin writing this section now, for today is the first day of the second semester. I have just walked past a large lecture hall; the instructor is animated; students are listening intently, wanting to know what is coming. For me, the irony here is that I have passed this room during past semesters, often observing several students sleeping or reading the campus news.

Day one of the semester is too important to throw away. If all we do is call the roll and dismiss the class, what message are we sending? Yet, many instructors do just that. "I didn't really think about this class until now," maybe, or, "You don't need to be any more serious about the material than I have been just now."

On the first day of class, students want to know how the course will be run: what are the major topics, why is the material relevant, and, of course, "How will we be graded?" In light of these student interests, what can be done with day one? Here are some suggestions:

Now let's discuss some aspects of each of the categories above.

By the seemingly simple act of calling the roll, you signal that you want to know the students. You will get to know some names, and that will make the course more personalized. This can lead to better attendance, fewer problems like cheating (since the students feel more invested in the class, and since they know that you know who they are), and better course evaluations for you at the end of the course.

Handing out a syllabus is another common first day activity. If you are new to teaching, you will have many questions as to how to construct such a syllabus, some of which can be answered in a later section, What Should be on a Syllabus.

Many instructors assume that students will read what is handed to them; I think this is incorrect. Every time I hand out a document, whether it be a syllabus or a homework assignment, I read it to the students. By reading through the syllabus, I allow students to ask questions that I may not have answered clearly in my text, and I also ensure that, within reason, students know what is required of them. First-time graduate students are often teaching first-time undergraduates. The undergraduates need to know how college operates: "Should I bring my textbook to each class?" "Will you collect homework every day?" "Do you answer questions during class, or do we wait until later?" "Do you grade on attendance?"

More advanced students will have questions, too. Maybe they have never had a mathematics course in college, or more likely, they just want to know what the rules are: "I have lots of job interviews this semester. Do you require attendance?" "Will you have answer sheets in the library, the way they did last semester?" By the way, there is nothing wrong with your answering, "I don't know; I'll check it out let you know next class." Just make sure that you carry out your part of this bargain and give them a definite answer at the next class. As to more specific comments about how class is to be handled, we will return to this topic in the section, Types of Assignments. Students want to know whether and how often homework are going to be collected. Will you grade each problem, or only some? How will they know which? Do you have an idea of how you'll assign grades to the homework? For instance, will you use a numerical system where each problem is worth, say, from zero to five points? If you know what system you or the course leader is using, now is a time you can tell the students.

Similarly, you can describe when you will give exams, and whether they will occur in class or in the evening. You can also describe where the exams will be given, for instance, in a large lecture hall with 400 students, or in the classroom. You can also tell your class that "You will have ninety minute exams, and I will show you some old exams for review."

Then you can explain what you know of the final exam and grading policies. Is the final cumulative? Does it have the same length as the other exams? Does it count for more points than the earlier exams?

There are other bits of information you also should give: The names of the texts for the course, your office hours, and any supplemental texts or materials you will use.

Now that you have spent about twenty minutes on the nuts and bolts of the course, it is time to turn your attention to content. What are the topics your students will be learning? How do those topics relate to other subjects they may be studying? In what ways will the material be useful in the real world?

Let's be more specific about details; many of you will start teaching with a first semester calculus course. You may want to say something like this:

Calculus is usually split into two types: differential and integral. Differential calculus deals with instantaneous rates of change: how things change right now, not over six years or ten miles (those are average rates of change), not over six seconds or six one-hundredth of a second, but right now, this instant. We will be learning about this instantaneous change this so-called derivative, how to find it, how to manipulate it, and how to use it in problems from physics and chemistry to business and economics. For instance, if the instantaneous change takes place over time, then this derivative is the velocity of the object that is moving, and this concept is of special interest to physicists and engineers; it is one of their tools for explaining the physical world. When Isaac Newton wrote F = ma, for instance, he was saying that forces are related to acceleration, and acceleration is a derivative, a rate of change.

Scientists are not the only people interested in calculus. Economists and business people also use the subject; for instance, the cost of doing business changes essentially instantaneously over time; this change of cost is called marginal cost. Monitoring marginal cost allows businesses to track their changes today, not over the last twenty weeks or twenty months.

Then you might go on to explain how taking a derivative requires having a function to work with; thus you will begin with a review of some continuous and not-so-continuous functions. After that, you can say that you will go on to talk about various methods of taking derivatives of more and more involved functions, and then you will discuss some applications of derivatives, such as how to maximize and minimize profits, say, or maybe velocities, or areas of land.

At this point, I will leave as an exercise for you can decide what you might want to say about integral and/or differential equations. Meanwhile, let's shut the door on this first day calculus class, and move down the hall to the precalculus class, where a more "activist" discussion has begun:

Instructor (I):"... and we'll also talk about functions. Maybe some of you have seen some functions, like, say, polynomials. Can you name some functions that are polynomials?"

Two students together (S1and S2): S1: "Sure. y = axn + bxn-1 +..."
S2: "Unh maybe x2?"

I: "O.K. y = x2 works. It's a polynomial. Any others?"

S1: "x3?"

S2: "How about y = x2+ x + 1?"

I:"Yes." [Writing both polynomials on the board.] "Anything harder?"

S3: "How about the square root of x?"

I: [Writing y = x = x1/2 on the board.] "That one doesn't work. Does anyone know why? "

[Silence. Then]

S1:" Cuz one-half is wrong."

I: "Good. One-half doesn't work as a power, right? I mean, y = (1/2)x2 is a polynomial, right? [Pause] So, this 1/2 points to the power in x1/2 doesn't work--I mean, it's not 'legal' for being a polynomial, although it is 'legal' for being some kind of function, yes? (This [points] is called a power, by the way, and the other is a coefficient of the polynomial. We'll define these terms pretty carefully during the course..."

[A couple of minutes later.]

I: "How about some other kinds of functions? Have any of you heard of trig functions? Can you name some?"

S1: "Sure. y= sin x."

I: "Yep, sine works. We'll study it, and the others, like cosine and tangent and why they're all different from polynomials. 'Sine's' picture, by the way,is, sin (x) right? And, it comes up in spring and pulley mechanisms, and electrical stuff, and things like that, and..."

Let's tiptoe away now, we get the idea.

This last instructor can teach us a lot about managing the classroom. Notice how she accepted the answer she needed to her first question, rather than going with the seemingly more complete response from Student 1, who obviously knows a good deal of the material she may be spending the semester teaching to the others in the class. She also did a good job adapting to the incorrect answer y = x1/2 suggested by Student 3. She did so without emphasizing the student's wrong answer; in fact, she turned a common mistake into a learning experience for the entire class.

There are many good points to the classroom discussion we have just witnessed, but in the interests of keeping the discussion short, let's just say the following: Most people say that teaching precalculus is boring, boring, boring, but this particular instructor doesn't make it seem so.

Exercises:

Which of the two methodologies described above for a first-day discussion of course material would you be more comfortable with? Fill in the details of what you would say to a first semester calculus class about the topics of integration and differentiation. (Your answers may be nothing, of course, but you should then have an explanation based on the syllabus.)

Return to Top

What Goes On in Recitation?

One typical format for a recitation is this: The TA begins by asking if there are any questions on the assigned homework problems. A student then asks to see "section 6.2, number 17." Other students chime in with "I couldn't do number 29," and "How about number 5?" Others ask for some problems from section 6.3. One fairly quiet student says, "I wonder if you could do an old problem from section 6.1?" Then, for good measure, another student asks you to try one of the questions from section 6.4, the next assignment, so we can see how they are done.

You, as the person in charge, can field questions in the order in which they occur, taking section 6.2, number 17 before number 5 from the same section, say. Or, you can ask for a list of all the problems at the start of class, collect them on the board, and do them in the order in which they occur in the textbook. The advantage of the first method is that you answer questions in the order in which they arrive. The disadvantage is that the student who couldn't do one of the easy problems may be totally at a loss as to what you are talking about when you start off with the hardest problem in the section. The second method solves the latter problem, but only at the risk of "falling behind in the material." This is a point you may not consider too important, but students always do.

A via media for making the best of both methodologies is to collect the questions as above. Then tell the students you will do the section 6.4 question "if there is time at the end of class." Starting with current material, do two or three of the problems from section 6.2, one or two from section 6.3, and then go back to the one from 6.1. Finally, if there is time, you can "suggest a hint to get people started" on the 6.4 exercise, which, after all, essentially no one has looked at but the one student who asked. In this way, you emphasize current material of most interest to the majority of the class, while at the same time showing that you are willing to deal with "old and new business" as time permits. And, by giving just a hint as to how to do the new problem, you allow the entire class the opportunity to puzzle out the secrets of that particular problem.

It should be clear by now that, since recitation consists mainly of discussing homework problems, you should show up on time and be prepared to discuss past and current assigned problems. A shocking number of TAs and instructors try to "wing it" often with unpleasant consequences for themselves, their students, and for their end of term evaluations. So I will say this again, with emphasis:

A recitation instructor will show up on time prepared to discuss past and current homework problems. No excuses are acceptable; this is part of your job.

This means that you will read through all the problems the night before recitation, you will perform the required computations (Yes, the chain rule is dull, and you have used it so often before, but, just when you don't prepare a set of problems because they're too easy, that's when you'll get stuck in front of your class on the day before the exam.), and you will get "the answer in the back of the book," because that's the one the students prize so highly.

Why do you want to prepare meticulously when you know this stuff so well? Because:

People never learn course material as well as when they have to explain it to others. Even though you took and passed this course some years ago, that doesn't mean you can't learn from a refresher. After all, it was six years ago in high school that you took AP calculus, right? Textbook authors love to put little tricks into the exercises to keep students on their toes; these tricks can trip up unsuspecting instructors, too. You are getting paid to do these exercises. Even a TA who has done this course three times already needs to recall where the pitfalls are placed.

You can probably add one or two more well placed reasons to this list. Remember those reasons when you decide to take a day off from preparing.

One final thought on this topic: Course evaluations bear out the importance of instructor preparation in students estimations of teaching. Even those faculty who are described as "boring" and "unmotivating" usually receive overall evaluations in the B-minus to B-natural range if students say that they "can do the coursework" as shown by their being well prepared.

In this section, I have emphasized the importance of being prepared in teaching recitations. Preparation is important, but it isn't the only thing. For more advanced advice, see the sections on The Active Classroom and Motivating Students.

Exercise:

Name some of the topics you think I have slighted or ignored in the above discussion. How essential do you think they are to good recitations?

Return to Top

What Should be on a Syllabus?

Some departments keep syllabus files, which provide a major impetus for institutional, not to mention personal, memory. Even if such a file is not readily available, you can still find out who taught your course last time, what books they used, which chapters they covered, how pleased they were with the outcome, and what they thought of the students. In the unlikely possibility that the previous instructors all have retired or left town, you can get some hints and advice from members of the curriculum committee, or maybe even from the staff member who deals with the campus bookstore. If all else fails, you can look at the appendix of this text for some sample syllabi of randomly chosen undergraduate courses.

Enough said about how to find old syllabi; now, what should yours describe?

First, give the name, number and section of the course. Especially if multiple sections are taught, you want to identify yours as specifically as possible. Also write days, times and room numbers on the syllabus; e.g., MWF 10-10:50, 304 White Hall. Put your name on the syllabus (some prefer Professor A. B. C. Jones others like D. Smith), your office number and hours if you know them. If you haven't decided your office hours yet, promise to write them on the board as soon as you do know them, and do so often over the next few weeks.

It is worth saying here that I am always amazed but probably should not be at how little verbal information is processed, and I am reminded of this fact over and over again throughout my career. ("I told the students that that topic would be covered on the exam, but they didn't listen.").

Name the required and recommended texts and readings, including edition numbers, specifying which texts are required and which are recommended. Then explain which chapters will be covered (Thomas-Finney, Chapters 1-7). If you are new to teaching, you may not be sure as to which chapters are required. It is very important that you find this information, for if students go on to the next course without having seen some of the requirements, a lot of people will be annoyed and upset, and people don't often forget what and who caused their problems.

Also on the syllabus, discuss homework, exams and grading in general; if you try to get too specific about requirements, students will come back to tell you how you have "changed your syllabus"--"unfairly," of course. (For more details about grading schemes themselves, see the section entitled Grading Issues. For now, we will stick with what goes into the syllabus.) Will you be assigning homework by the class? By the week? The month? The entire semester? Will you collect and grade all the homework, or just some? If just some, will you be announcing in advance which ones you will grade? When will you collect these problems, e.g., "Right at the start of class each Monday"? Do you want the homework written out in any particular format?

Then there are exams. Do you know when they are to occur? If so, put that information into the syllabus, along with any other details you may have, such as how long the exams will be and where they will take place.

As far as grading is concerned, offer a general statement like, "...three equal exams, along with a comprehensive final exam counting double [alternatively, one-and-one-half?] the value of the exams. Homework and class participation will also count about ten percent of the total grade." In this way, you offer the students a framework, while at the same time allowing yourself some leeway-- "what is class participation" for instance, and how do you propose to measure it? Yet, it's logical to suggest that such participation is worth something, and you do want to have a mechanism for rewarding students who make an extra effort.

At this point, syllabi often diverge, depending on course, material and style. For instance, I have seen a few-- very few, actually -- faculty put a short description of their academic credentials in the syllabus. Others, especially those who are teaching in a fairly nontraditional way, will feel the need to describe the classroom situation as they see it happening. For instance, they might describe how their "project-oriented" calculus sections will work, what kinds of writing assignments they will offer in their geometry class, or how they will handle group work in their precalculus class.

Regardless of what you put into your syllabus, it would be well to remember that this document takes on the character of a contract with the students; you are telling them what you plan to do, and in turn what you expect them from them. Thus it behooves you to take a little care with what you write. You might consider passing it by an older, wiser faculty member for approval.

Courses often require unplanned or unexpected changes in midstream. Most of these are acceptable to students. On occasion, however, some adjustments you understand to be minimal or benign will elicit an unexpected outburst "Why are you canceling exam three? I was counting on that one to boost my grade! You can't do this unless the entire class agrees, you know!" And then, heaven forbid, if you decide to "take a vote" on the question, those students with higher grades plus those who just don't want to take an exam along with those who just want to get the course over with will simply outvote the three really angry ones who want the exam. The ultimate outcome is that you end up giving an exam you hoped not to give, while you have lost the respect of, and authority over, your students.

No one can avoid all difficulties or see all the possible problems about to appear. However, you need to think carefully about your syllabus before you start the semester. That and getting input from colleagues is a strategy that will make for a more coordinated course. The outcome of such planning will then be better for you and for the students, and will make your course less work in the long run.

Exercises:

What goes into your course? What would you add to the above syllabus? Is there anything that you think should be subtracted from the syllabus, and if so, why? How would you resolve the problem discussed in this section of the student who wants to take the third exam? Is he being unfair? Are you wrong for suggesting that the exam be dropped?

Lesson Planning: Survivalist Tactics

Suppose you came to town on Thursday; it was your first time at the college. Suppose further that classes begin on Monday, that you have to move into your new apartment, register for classes, wait for the cable to be connected, and oh yes you have to start teaching your very first class on Monday morning. For what to do on Day One, of course, you can look back to that section in these notes. Of course, that doesn't let you off the hook that easily, because you are stuck trying to build a syllabus. And then, once you have given the students a general introduction to the course, you will have to start making up lesson plans. Further, you won't just have lesson plans for Tuesday (or, if you're lucky, Wednesday) you have to plan an entire semester's worth.

In this section, let's consider the most basic aspects of lecturing. Later, in sections called The Active Classroom, and Motivating Students, we will look at more refined aspects of making such plans.

Once, some years ago when I was a graduate student teaching a night course in third-semester calculus, I got the twenty-four hour flu about an hour before class started. My office mate, being a very kind person, offered to substitute for me. "Just tell me which section you were supposed to do," he said.

The next day, after I had sufficiently recovered, I asked my officemate how things had gone. "Fine," he replied, and went on to tell me how far he had gotten in the material. "But how did you prepare?" I asked. "Easy. I went in to class, announced that I was substituting for you, asked to borrow a copy of the text, and gave the students a five minute break while I looked over the author's approach to the material. Then I made up three examples of varying difficulty, and I went with it. By the way, how are you feeling?"

So there it is; a basic plan for lesson plans ("a plan for plans"), courtesy of my officemate.

Start by finding out what today's topic is supposed to be.

Peruse the text to see how the author approaches the topic; this helps you preserve the same notation as the text, among other things.

Prepare an intuitive explanation (a "heuristic argument") as to why the topic is important, useful, and relevant.

Next, prepare a few homework-style problems of increasing difficulty to illustrate to the students the main concepts of the section of the text.

Allot remaining class time to answering questions or doing old homework problems.

Of course, this methodology doesn't solve all problems. If it did, teaching would be trivial. So, let's discuss some of the issues raised in the above outline more fully.

One complaint often voiced is "But I don't like the way the author does this section. Why should I encourage bad mathematics?"

Fair enough. Even though we may not have had a choice in the textbook, the students will still be using it for explanations, exercises and homework. We can offer alternative proofs or better methods, but if the students are getting their homework from the text, they would rather not have to keep "translating" from our language and symbolism to the author's. Thus, we owe it to the students to at least say, "Here's how the author approaches... An easier [more common, better, more useful, more sophisticated] way is as follows. On the homework and tests, use whichever method you like best. I don't care as long as you get the right answer and can explain your method.

Another common objection is that we should not use "heuristic argument" rather than an "honest, direct, complete proof."

This suggestion may simply be a function of audience level (discussed more fully in the section Student Types). Clearly, if you are teaching the intermediate value theorem in real analysis or topology, you will want to consider the roles compactness ands connectedness play in the discussion. But, for a freshman English or biology major, some pictures of continuous and discontinuous functions that have positive y-values at x = 1 and negative y-values when x = 3 will be much more convincing than an unintelligible, unmotivated "formal proof."

This last is also not to say that you can't be lucky and draw a class of eager students in an enriched calculus program for potential mathematics majors but now we are back to the Student Types question.

A third objection is, "Why do examples? They're right there in the book."

You're right; there are worked out examples in the textbook. But, first of all, many, if not most, students donÕt read the book. Second, not every detail of the examples is spelled out in the author's exposition. Further, it isn't always necessary to choose the examples in the text; many instructors I know don't choose the author's exercises. Instead, they opt for a few problems "near" the assigned homework problems, telling the students, "If you understand how to do these examples I'm showing you, you'll have a great start on tonight's assignment." The underlying message is the "great motivator": "It's worth watching me do these problems, because they're like the ones you'll be trying soon."

One more objection to the proposed lesson plan is often brought up: Is the suggested allotment of time for a lecture correct? That is, how can you leave so much time for questions and homework? Don't you need all that class time to explain the details of the current topic?

This, too, is a reasonable objection. Different instructors find that they take different amounts of time to explain details of a lesson. Still, I try to find ways to leave time for student questions; otherwise, how do I know whether they are absorbing the material I claim to be teaching them? The best way to find out if my lecture is being received is to give the students a chance to tell me what is still bothering them. I will return to this topic again in the Active Classroomsection.

Return to Top

Grading Issues

It goes without saying that grading can bring on problems. Many students seem to feel as if they "start out with 100%, and we [faculty] must justify the removal of each individual point." At the same time, faculty sometimes take the exact opposite approach.

Grading is best treated as a learning situation for all concerned. The instructor learns how well he or she has taught the material and designed the exam, while the student learns how well he or she has absorbed the course information and studied for the exam.

In mathematics, you will usually be grading homework, quizzes or examinations. Other possibilities are that you may grade writing assignments or class projects and, of course, you will be involved in assigning final grades.

Homework is generally the easiest to grade; the assigned problems are usually well written out in the text, and the solution method is fairly clear.

A common, but not universal, technique for grading homework is to assign each problem a fixed number of points. Some graders use a two-point system, "0" for a wrong answer, "1" for OK but not complete, "2" for fully correct. After using this methodology once or twice, most graders find that it doesn't have enough points to properly distinguish among the variety of possible errors that a group of students can make. Students also tend to sense the same problem. Their complaint about the grading is usually to say something like, "I only got one number wrong, and all I got was a 1".

A zero to five scale is probably better:

"0"-- didn't even try the problem,

"1"-- tried, but not even close, "2" and "3"-- various levels of somewhat valid but mistaken attempts,
"4"-- correct answer but with some minor errors,
"5"-- the correct answer with details spelled out.

Note the last comment: Only the correct answer with details merits full credit. There will be points early in the semester when students will ask you to reconsider grades because they "got the right answer" without showing any supporting evidence as to how they did so. You can use this as an opportunity to instill good habits into the students. Explain to the questioner that he has lost one point on this particular assignment for not clearly describing the way he went about solving the problem. This is much easier than trying to convince the same student that he should have lost twelve points out of twenty for the same approach to a problem on the second exam.

This last paragraph points out a good general principle, both for students and assistants. Homework time should be used to instill good habits. For the student, this means writing correct, clear, complete solutions. For the instructor, look to make uniform, defensible grading with useful comments.

New TAs often ask how long comments on papers should be. My response is usually "Not long at all." I say this because it is fairly common for newer TAs to continue the solution to the problem in the margin of each student's paper exactly from the point at which the first error occurred. Students often don't read these comments; sometimes they do read them, but still don't understand what they did wrong.

There are a least two ways to reduce the amount of commenting you need to do on homework. One way is to simply put an "X" mark at the place where the first error occurs, and then after all the papers have been graded, write up solution sets of the most commonly misunderstood problems for all the students. A second way is to start or finish the next class with a "couple of homework problems lots of people seemed to have difficulty with."

Quiz grading is not dissimilar to that of homework. You can use a similar point scheme, and again you can save some grading time by putting the answers on the board when you hand back the quiz. One difference that sometimes occurs, however, is that if you are the one writing the quiz, you may occasionally find that your question is inappropriate. Even if you are not writing the quiz, but simply (remember, nothing is ever simple) choosing a problem from the text, you may choose one that requires a piece of information that you actually didn't lecture on. In that case, common sense should take precedence over pure grading issues. Did you ask a question students couldn't answer with current methodology, say? Then maybe you should give everyone full credit for their valiant efforts, with extra credit for the one or two who may have actually known how to solve the problem.

Exam grading is also in many ways like homework, although in this case careful preparation before grading can save much time. There seem to be two models of mathematics exams: Those that are given to classes of up to thirty students, and those for five thousand (Well, maybe five hundred). In the first case, you end up grading all problems on all the student papers. In the second, you tend to grade only one problem but you must grade until you drop, and then get up and grade some more.

Uniformity with fairness and speed are keys to grading exams. Nothing is more disconcerting than finding at 3 a.m. that you have graded 347 papers, an undetermined number of which were done incorrectly. If you are grading 500 papers, carefully doing the problem yourself before grading any papers is central to uniformity. When you have a complete solution, make up a credit scheme before you grade any papers. (A sample problem solution with grading scheme is shown at the end of this section.) While you are proctoring the exam, you can show your answer and grading scheme to other TAs for comparison although it doesn't hurt to remember that this will be your problem, so the final approach is your call, as well as being your responsibility. My point here is that, within reason, you need to find a grading scheme you are comfortable with, one you can defend.

As you grade the first few papers, occasionally review your scheme to see if it still seems to fit what the students actually knew and did. This review will also help avoid grade inflation or deflation that seems so inevitable over ten hours of work ("This is the same mistake that I've seen a hundred times now well, this time you get a zero!")

Uniformity and fairness are related to one another. You may be a harder grader than your officemate, but if you can defend your methodology to other TAs and students, they will "generally" accept it. (Note that last generally. Some may not; see the section on being a good colleague.)

Most TAs see the "speed" part of "grading with speed" as only being of benefit to themselves--"I want to get this pile of papers done and out of here!" But speed with accuracy also benefits students, because they get to have their problems back while they still remember what the questions were.

To aid in speediness, try some of the following:

If you are grading a full class of exams--thirty students, the entire set of exam questions, say--grade problem one for each student, then go on to problem two, etc. In this way, you will ensure more uniformity. Also, try to grade each individual problem in one sitting; take a break only after you have seen all the unique, exotic methodologies the students can come up with. (By the way, I find that I can sometimes bribe myself into grading by promising that I'll take a break as soon as I finish these last eight copies of problem 3. Not being very bright, I'm usually able to use that argument to convince myself to work ten more minutes.)

Uniformity has other benefits. Among them: It leads to fewer re-grades, which take a lot of time. It also makes for more defensible scores, so that students consider the grading (and the grader) fairer.

After you have graded as many homework, quizzes and exams as you can stand, you will have to assign final grades. Each department of each university and college seems to have allowed its own system to evolve and each of these systems is like each other, but not quite. For the bare-bones description of one such system, check out the grading section of What Should be on a Syllabus. Note, however, that this section is not completely forthcoming as to how allocation of final grades is done in an standard class. Well, let us lift the veil.

I am occasionally in charge of a large number of calculus sections, for instance 26 sections averaging 20 students each. Thus, by the end of the semester, we calculus instructors have in the range of 500 grades to assign. Assume that we give three exams during the term (these are called prelims where I come from), each worth 100 points. We also administer a 150 point final exam (yes, it is called a final); and additional materials, such as homework and quizzes, add up to 50 more points. Thus students can earn a total of 500 points. If the exam is scheduled for a Tuesday morning, we will spend that afternoon, Wednesday and Thursday grading. By late Thursday afternoon or Friday morning at the latest, each instructor will have collected finals, recorded grades, and totaled raw scores. (Of course, there are always one or two instructors who have failed to do the above; they should read the section Get Along with Colleagues.) We then have a meeting at which instructors put up the raw scores of their students. This we do in ten point intervals, from 500-491 to 210-201. Numbers below that fit into the 200-0 category.

We find the median grade (not the mean), and assign to its ten-point interval the set of last B- grades. Working up and down the intervals, we then assign an A range, a C level, and a D range.

Interval Total Sum
491-500 3 3
481-490 8 11
471-480 14 25
461-470 20 45
451-460 16 61
441-450 20 81
431-440 16 97
421-430 22 119
411-420 28 147
401-410 etc.

Having put together a curve based on class scores, we now assign letter grades to each student in each section. We are not done, however. In each section there are grades that are anomalous. Some students have one grade that is much lower than the rest, say. Others have a rising set of scores, e.g., 49, 62, 87, and 130/150 on the final, showing that they maybe have caught on later than others. Occasionally, students will have personal problem. I discuss each such anomaly about fifteen- percent of the total -- with the individual instructor, and we come to some sort of consensus. We seem to end up raising about half the grades, but no single grade ever goes up more than one level, e.g., from C+ to B-.

One of my general feelings about grading is that students always do less than or equal to their best on individual exams; but still, there must be two or three exams where they perform to expectations theirs or mine. Further, good homework and classroom questions may show interest, but they are a precursor of good exam performance, not a substitute for that performance. And finally, I have a thing about the grade of A+; I will never raise a total below 490 points out of 500 to an A+. To my mind, an A-natural is a perfectly wonderful grade, and I won't apologize for giving it.

Return to Top

Cooperative Learning

There are various ways to approach the methodology called cooperative learning. You can suggest that students do their homework together. You can offer them weekly study sessions where they can sit together and work problems while you circulate through the room offering hints and suggestions as to how to solve problems. (For more details of how such a program worked for Myrtle Lewin and me a few years ago, see [17].)

After a short introduction of a particular topic, you can stop lecturing so as to let students try two or three of the exercises together, after which you can have them present their solutions at the board. You can make up worksheets for students to use to discover mathematical concepts for themselves some sample worksheets are provided at the end of this section.

You can teach a project-oriented type of class, and then make up some really hard worksheets that the students will need a few days and some help from you to construct solutions for. Then the students can work in small groups writing up their solutions. This method of cooperative learning is very labor-intensive for all concerned, and is not one you should simply blunder into. There are books available, however, to help you with the details see, for instance, [8] and [15]. At the end of this section, I have offered some of Matt Horak's calculus projects as samples.

You can assign major projects (or final projects, or individual projects) in place of some exam or final. Students can then report to each other on what they have learned and they can evaluate each other's projects.

I have no doubt that you can think of two or three other ways to encourage cooperative learning in your classroom.

In addition to the question of how cooperative sessions can be done, there is the more interesting question of why you might want to do them. Faculty often express a desire to have students actively engaged in the learning process; what better way to do this than to get them to work the problems and construct the examples?

An objection that is sometimes raised to this last is, "But when six people work together, I can't usually tell which of them are really working, and which are copying." OK. "And, can you tell when they hand in their homework whether they did it themselves or had someone else do it for them?" When you get them working in class, you can walk around observing the dynamics of groups; you can ask questions like "Where are you stuck?" and make such comments as when your group has a solution, I'll ask one of you [not saying which one] to present it on the board.

Return to Top

Technology

Technology is being used more often in the mathematics classroom, from low-tech aids like overhead projectors and microphones through mid-tech calculators to high-tech computers. The more time you spend in teaching, the more you will be called on to use some of these materials.

Perhaps it is just my own bias, but I find it particularly annoying when an instructor comes to class unprepared to use the necessary equipment. Ho ho, well, I brought these transparencies, but I see they don't fit the projector. [You didn't bother to check beforehand?] "And, I can't figure how to turn it on... can anyone help me? Oh, and I see that my data disk isn't compatible with the software and no, I didn't prepare any backups, and..."--well, you were at that talk too, so we both remember it well, right?

Rule One: If you are using technology in your class, test it beforehand. Have a backup in case the worst happens.

Practice with the equipment before your talk or class; turn it on, check the displays for visibility, set audio equipment for sound clarity without feedback, make sure computer and calculator displays are visible from the back of the room, check to see that you have the proper cables and plugs for your laptop.

Rule Two: Make sure your examples justify the technology you are using.

Why do you need a calculator with an LCD display projector just to draw the graph of a parabola? Do you have to load up Minitab or Datadesk to compute means or deviations for a sample of size six? In other words, choose your examples to fit the equipment. Let the calculator graph the function y = sin(x)/(1 cos(x)) so that you can find all max and min for that function. Have the calculator graph y = ax2 + 1 for various choices of a; then ask students to describe how changing a affects the graph. Do the same for y = x2 + k, for various k.

Use Minitab to find means and deviations of census data, so as to eventually construct and test hypotheses for yourself and the students to defend or disprove.

Rule Three: Make sure that your overheads and displays fit.

A transparency with writing too small to read may not (quite) be completely useless, but it certainly is frustrating to the audience. It is not always possible to know how a particular room or hall is going to be configured for a talk or class, but there are many good reasons for putting less information on each individual transparency and making the size of the type, font or print larger than you think you will need. The next rule will discuss more details of speaking from overheads; for now, let me continue with another nuts-n-bolts comment.

Another frustration for the audience that is simply solved is that of the shrinking violet who refuses to pick up the microphone. He thinks his mumbling is sufficient, while the audience knows that the solution is available right at the lectern, if only he would use it.

Rule Four: Realize that teaching with technology is not the same as lecturing.

The audience does not have time to take notes or absorb the ideas being shown on transparencies. They are barely able to listen to your presentation. Such problems especially occur in good talks; it can be very frustrating for a listener who thinks that he or she has just seen a remarkably good lecture but can't really reproduce any but the most minimal parts of it. Such an audience has been more entertained than taught. To alleviate this problem, you could consider bringing individual copies of transparencies and displays for handout.

Speak slowly and allow ample time for questions while overheads are still on display; recall that all but the experts in the room need time to absorb what you are telling them. Also, do not play "peek-a-boo" by covering over parts of transparencies. If you don't want the audience to get too far ahead of you, put less material on an individual transparency and write larger. Another method of keeping interest is to offer people a related exercise at the start of your presentation that (you can claim) that they will be able to solve by the end of the talk.

Rule Five: Be prepared for total system meltdown.

I once gave a lecture in Japan on the day a typhoon hit. Five minutes after I began, all the electrical systems failed. When I asked what to do, the audience said, "Just go ahead with your talk." So I used chalk, wrote large on the board and spoke loudly. Some of what I said must have gotten through, because afterward a listener came up to me and very courteously pointed out an error I had made.

Return to Top

Writing Assignments

A variety of small writing assignments are usable in the mathematics classroom.

Have a supply of 3x5 index cards in the back of the lecture hall for students to use to write questions about the lecture. Answer the best or most frequently asked questions at the start of the next class.

Ask an occasional quiz question in class: "What's so fundamental about the fundamental theorem of calculus?" "Describe one application of today's topic."

At the end of a solution of a mathematical exercise, ask the students to describe the real-world implications of the answer they just got.

Have the students write out a description of the topics covered since the last exam, as well as why those topics might be important or useful.

None of the above assignments takes a long time to construct, nor is it difficult to grade. Yet each enhances the students awareness of the usability of the classroom material. Further, each asks the students to think a bit more holistically and carefully about the somewhat deeper meanings of the materials they are studying.

Of course, the above are only a small sampling of the possibilities of writing assignments in mathematics. For lots more, check Countryman [9] or Meier-Rishel [19].

Return to Top

Making Up Exams and Quizzes

Making up exams is both an art and a science. If you do it properly, you get an honest appraisal of your students understanding of the course and the material and approach you have taken. At the same time, by constructing good exams, you can avoid the pitfalls that make examinations time consuming to grade and difficult because of post-exam complaints. If you don't do it well, you can probably guess what I'm about to say.

So what are some of the steps you can use to avoid pitfalls?

First, make a list, for yourself and for the students, of the topics you have covered since the last exam. If some of these topics are too time-consuming or not interesting enough to test, say so. If you feel you need to test one of the time-consuming topics, e.g., Newton's method, or Riemann integrals evaluated by summing and using induction, you can consider testing them by assigning a special overnight take home project.

Now that you and the students know which general topics are to be tested, it is time for you to decide, without the students help of course, how many and what kinds of problems to assign in an average examination time period.

Let's say you have ninety minutes. I sometimes tell students that I try to design a one-hour exam and then give them ninety minutes to do it. This rather meaningless bit of information seems to relax them. Generally speaking, stress reduction before the exam is not a bad idea; there is a difference between making an honest and fair, yet difficult, exam, and making one which is simply filled with tension.

Suppose you have decided on a five question exam, based on the fact that there were seven major topics since the last test, and one of those is easy enough to skip, while a second can be embedded in a later, more important topic. Make about forty percent do-able by anyone who stayed awake long enough to watch you show some examples on the board. (This is one of the reasons I don't want to pass students who cannot get a 40% average on my exams. See the section on Grading for details.) The forty percent do not have to be just like trivial homework, by the way; you might split some of your five problems into easy, moderate and difficult sections, thereby spreading the easy stuff around the exam.

Now you have sixty percent of the test left for more challenging material. Half or a bit more of that can be similar to some of the more interesting examples and homework problems the types of problems that make students think, but this group had a chance to do that thinking last week while they were doing their homework exercises. I basically never give the students assigned homework problems on the exams, by the way, although I do know some people who do. I just feel that using old exam questions as homework problems often makes the students feel that the instructor didn't really put an effort into the preparation.

Assuming that about twenty percent of the exam is still to be constructed, it's now time for you to think of a more challenging question -- or parts of questions. Now is the time to think, "What is the essence of the material I have been teaching for the last four weeks, and how can I ask the students to show whether they have absorbed that essence?" This does not necessarily mean asking them to formulate a proof; rather it should indicate that you could quiz them about some fundamental points that you have been making repeatedly during your excellent lectures. One effect of putting such questions on the exam is to increase attendance at the rest of your excellent lectures--"Wow! If I go to class, it might help my grade on the next prelim!"

For many more details on how to make examination questions that hit the mark, try the section Using Cognitive Levels to Make Appropriate Problems.

Let me say a bit more about finding challenging problems. Early in my career, I used to expend real energy trying to fashion a problem that would force students to use current knowledge to discover something new. For instance, I might be inclined to write, "You've seen exponential growth. Well then, now I'll ask you to find out about logistic growth all on your own." These well-meaning attempts almost always turned out very badly. The "numbers" would turn out to be too messy, and the concepts were too far from the students current awareness. Further, thirty minutes or so was simply too little time for serious thought. So, eventually, I came to the realization that at best I could formulate a couple of problems that proceed from easy to difficult, with the difficult part counting maybe only five points. ("You couldn't get that part? Well, good thing it was only worth five points. OK, let me show you how...")

How do I do this? Glad you asked. Let's go back to the exponential growth into logistic growth problem. We split it into four parts, each worth five points:

In problem four, you found the rate of growth of a strain of bacteria. Now let's suppose that the bacteria are growing in a lab on a circular Petri dish whose area is 5 cm2. Thus it is fair to assume that the area, A(t), covered by bacteria in the tube at any time t is governed by the equation
a) dA/dt = k(5 A(t)). If you know that A(0) = 1 and dA/dt = 0.2 at
t = 0, what value do you get for k? Is this k value positive or negative, and what does it tell you about dA/dt?
Now writing b) dA/dt = k dt
5 A(t)
for your value of k,
solve this equation for A(t). Your solution in part b) will have an arbitrary constant in it. Calling that constant D, find its exact value. Using your final solution to part c), make a reasonable argument that A(t) is never larger than 5. What is your reasoning for this?

Notice that the above is still not an easy problem it wasn't supposed to be. However, the first part should be manageable for any student who understands what you have taught about exponential growth and decay. Part b) is harder, of course, except that you have already separated the variables in the equation the students have to solve. Even if students found a wrong answer to b), you can still grade part c) as if b) was correct. Thus they can still receive credit for part c) without getting very many points for b) at all although they do have to get some kind of reasonable answer for b). That leaves part d). It's not easy, "but hey, at least it's only a five-pointer, right?" If you now design one more problem with a hard five-pointer as part d) and you are done. On this exam, it'll be easy to get at least 40%, the average should be around 70 or 75%, and more than 90% shows that the students have worked.

Before moving on, let me make a comment about quizzes: I tend to make them relatively easy. For instance, if I lecture on the chain rule on Friday, and maybe I've shown the students how to find the derivative of sin2(3x), I might then ask them to use the last five minutes of class to find the derivative of cos3(2x). Once the students see that the quizzes are reasonably easy, they have incentive to come to class and listen carefully to what I am teaching. Further, the quiz is then easy for me; I can sometimes finish grading in the fifteen minutes between classes, if no one stops me to ask questions.

Return to Top

Using Cognitive Models to Make Appropriate Problems

What makes examination questions easy or difficult? To some extent, it is the students level of preparation and their attitude toward taking tests. These are internally applied forces coming from the test takers themselves. But, there are also external stimuli at work here, such as the difficulty of the examinations we instructors construct.

Some years ago, a variety of individuals began to study what are now called cognitive patterns--the ways by which individuals learn information. By now, a number of books and articles have been written offering models of what are termed levels of cognition; i.e., the levels of difficulty of concepts, ideas, questions. Most of these texts are more applicable to the arts than to the sciences or mathematics (c.f. [1], [2], [3]), but one which has been successfully used in mathematics is due to Benjamin Bloom [3].

The Bloom model splits cognition into six levels, from lowest to highest. These levels are: Knowledge, Comprehension, Application, Analysis, Synthesis and Evaluation.

Knowledge, Bloom's lowest category, pertains to whether the student has absorbed and can successfully reiterate the concept being taught. For example, in a first-semester calculus class we instructors might apply this principle to teaching the definition of the word limit. To see whether the students have absorbed that definition, we can ask the logical but not particularly inspired question, "Can you define the word 'limit'?" This is an acceptable question; for the students to master the concept, they must be able to articulate it.

Let us now aim for higher levels of questions by going to the next level in Bloom's model.

A note here before we go on: Bloom's use of the word "knowledge" in reference to the lowest level of cognition has sometimes been called into question. Many people have suggested that they believe the word knowledge indicates more cognitive awareness than Bloom seems to ascribe to it. Alternatives such as memorization and recitation have been suggested. We believe that the word recall would adequately express the concept that the student is trying to reproduce what the faculty member has taught.

A second level of cognition described by Bloom is comprehension. Do the students have some understanding higher than mere memorization of the concept under consideration? Here we mathematics instructors may ask, "When you think of the word limit, what do you see and how does this relate to the standard definition?"

Once the students have shown ability at the level of comprehension, they can be asked an application question: "If you were trying to extend the definition of limit to three dimensions, how might you try?" With an assignment of this sort, we are attempting to get the first-semester calculus student to extend their intuition to third-semester concepts. Notice, of course, that since the topic of functions of two or more variables has not generally been introduced in the first-semester course, such a question is probably not a good one for an examination; it is, however, a reasonable discussion topic for a classroom situation.

An even higher level of cognitive thought that we might ask of a student is that of analysis. Here the student must take apart the concept in question and put it back into context while considering the implications. For example:

1. Use words and pictures to investigate what happens to the function
z = f(x,y)=xy/(x 2 + y 2) as x and y head toward zero along the two distinct paths x = y and x = 0.

This will not be an easy question for first-semester students to answer; they will need help in finding the way because the thinking requires an inductive leap. At this level of complexity, we might ask the student to write out the thinking in natural language as a way of articulating the mathematical concepts. This could be done at the board, as homework, as a class project or in an individual paper, depending on time constraints. Note, however, that for a third semester or higher calculus student, the above question might simply be one at Bloom's comprehension level.

Synthesis refers to the coalescing of analysis into an argumentative claim, a difficult and complex cognitive process for most students, but one that can guide them to a more sophisticated level of mathematics. Topics of this type can include hypothetical definitions, various historical treatments, or differing interpretative views. A typical synthesis topic: Discuss historical factors that led from CauchyÕs definition of limit to Weierstrass's definition. Such topics best lend themselves to writing assignments.

Of course, a synthesis question can also be quite sophisticated, and we would probably not want to suggest it to students in a first-semester calculus course. However, we might be interested in spending a few minutes discussing such a topic in class so as to give the students an understanding of how mathematics is not written in stone how it changes over time, how intuition solidifies into definition and theorem, and related ideas.

In a synthesis assignment, we can modify the expectations we hold for students. Claims can be low-level or high, depending on whether the thesis and subsequent proof reach back to the earlier levels of critical thought those of application, comprehension and knowledge or up to a high level like evaluation and argument. Generally, a synthesis paper includes most, if not all, cognitive categories; since the response will be complex, the student will address at least one high-level idea.

The highest level of comprehension in Bloom's model is called evaluation, where the student shows that she can embody the concept itself as well as all the cognitive levels described previously. At this cognitive level, we might ask for a major paper, since obviously these questions require a substantial effort from the student. For instance, we might ask: "Why is it so hard for people to absorb the epsilon-delta definition of limit?" An evaluation discussion is sometimes called an argument. Argument papers make for quality major projects, as in "Discuss how Newton's and Leibnitz views of calculus differed, and the implications of each person's approach for the history of thought and for the teaching of mathematics."

2. An Example from Infinite Series.

To find out if students have been listening in lecture, I sometimes ask a quick quiz question at the recall level; for instance:

Find the limit of the series: 3/2+3/8+3/32+3/32+

This simply asks the students to substitute the appropriate numbers into the formula for convergence of geometric series.

To gauge the students at the level of comprehension, I propose that they:

Find lim[1 - 4 - 6 + Σ(3/2)(1/4)n]

Here I am testing whether students know the above formula. At the same time, I can find out whether they have absorbed the fact that, as I have said a number of times in class, adding or subtracting a finite number of terms to a series does not affect its convergence or divergence.

The following qualifies as an analysis question in a freshman class:

Find all a, r such that Σarn converges.

This last is a possibility for either a longer quiz question or an examination problem.

The next question, a "synthesis" problem, was used as a major part of a final exam:

Discuss the infinite series you have seen in this course. Include convergence tests, and provide examples of series for which each test does and does not work.

Although students did not exactly appreciate the above question, I found out a great deal about what they had learned about infinite series.

An "evaluation" project, much too involved to be considered a mere examination question, is one given in a mathematical exposition course: Trace the historical development of the concept of convergence. Include some analysis of Euler's work on defining the exponential, along with a discussion of the Riemann integral, Cauchy's tests versus Cauchy's "proofs," Weierstrass' work, and G.H. Hardy's contributions to series tests.

3. Levels of Cognition and Tutorials.

Once we understand levels of cognition, we see that we have been using them in everyday situations:

A student walks into our office and says, "I can't factor 10x2 x 3." In Bloom's model, this is a knowledge question. We show her how to factor the polynomial, then suggest, "Now you try the next one, 15x2 x 1." We are testing comprehension.

Next, we wonder if she can find a way to factor a cubic. Here, we are asking for application, and also an analysis. The student now wants to know, "Is there a formula for finding the roots of all possible polynomials?" She's having us analyze and synthesize our knowledge of algebra.

Finally, when she blurts out, "Why do we need to know this? Can you give me a real world application? What's algebra good for, anyway?" she's asking for evaluation; she wants a reasonable argument as to why she should learn the subject.

Of course, we have been dealing with such students all our academic lives. The point here is that when we understand the cognitive levels of the questions and answers we are dealing with, we will better understand the cognitive levels of the students. Then we can design our curriculum, questions and exams to fit those levels.

4.Course Design and Cognition

Although the content and methodologies of calculus and algebra courses are usually (thought not always) quite well defined, the same is not necessarily true in other areas of mathematics, such as history of mathematics or geometry. Here, too, knowledge of cognitive levels can aid in course construction and application.

Tom has previously, but less completely, described a geometry course for students who "know no geometry" [23], which he has constructed around the Bloom model:

Start by asking the students what geometric words they know or remember. At this point, they are working on the level of knowledge, offering words like point, line, plane; names like Euclid and Pythagoras; such concepts as theorem or axiom; vague terms like shape and solid. Spend some time asking the students to catalogue their randomly chosen terms into categories; what makes words like point or Pythagoras different from surface or theorem? This allows students to show how much they comprehend the terms of geometry, and to analyze differences and similarities in the concepts.

Now ask: "What is geometry?" This is a question that requires a good deal of synthesis and evaluation. Allow students to find a definition that appeals to them and suggest that as a "working definition" of the concept. Then give them an assignment that challenges their definition.

For instance, students almost always describe geometry as earth measurement. A good assignment which confronts this definition is to have them measure the height of a large building--an application of their knowledge of geometry--and then find out by subsequent discussion (analysis of their methodologies) that the technique they have used does not work on the surface of the earth. Then go on to the move sophisticated problems that occur because of the counterexamples that can be constructed, questions like: "What do triangles look like on the surface of the earth?" Such questions are very analytic and give the students a feel for how the rules of mathematics are constructed.

The course is now opened out upon a plethora of analytic questions that lend themselves to writing and discussion:

"What is an analogue to a straight line on the surface of the earth?
"How would you measure angles on the earth?"
"What is the difference between the surface of the earth and the earth itself?"

Further, you can address synthetic and evaluative questions:

"How many surfaces are there? And what does the word surface mean?" "If Euclid knew the earth was round, why did he say he was studying geometry? What was his definition of geometry, and how might it differ from ours?" "What are 'straight lines' in space? Can we even talk about them when there is no 'grid' to compare them with? And if we cannot talk about straightness in space, how do we know how to get to the moon?"

None of these questions is trivial, yet all come out of only a few hours of simple discussion and seemingly trivial writing assignments given to moderately skeptical students who are supposedly not at all mathematically inclined.

Return to Top 5. Some Final Considerations

Although we have chosen Bloom's vertical model of cognition, there are others that can be useful in the processes described in this paper. For instance, see Chaffee [7], who offers a horizontal model based more on student writing strategies. Another model, more linguistic in format, is offered by Vygotsky [26]. See also Piaget [21] for a discussion of cognitive strategies in young learners.

Another simple cognitive model from composition courses, somewhat overlapping with Bloom's, consists of just three classifications of writing: personal, informational and argumentative. One personal paper might be "I chose to do mathematics because I found it as creative as art"; another is "My mind is more inclined to algebra than geometry." Most mathematical papers fall into the informational category: the "large-scale geometry of the universe can be partially explained by curvature in two and three dimensions"; or "Messages can be sent with (almost) complete security." Thesis statements formed around business applications of linear programming and fractals and fractals dimensions can also fall into this category. Argument papers, as much rhetorical as mathematical, have either soft or hard theses. A soft thesis requires only information to prove its claim through a lower level of cognition; a hard thesis, unless it employs sophisticated statistics, must be supported by more analysis and interpretation. Examples of soft theses might be the following: "Calculus students are ill-prepared by their high school experience" or "Writing assignments provide an effective form of evaluation in the mathematics classroom." Some examples of hard theses, depending on how they are approached, are: "Statistical analysis shows that cancer rates highly correlate with cigarette smoking" and "Better socialization in middle school leads to higher retention rates of women in mathematics and science programs."

By using cognitive models as guides in our mathematical teaching in ways that our writing colleagues have long done in composition, we can move students to deeper levels of understanding of mathematics. In the classroom, where students often ask, "What's it good for," the use of cognitive techniques can help them, and us, find answers to this kind of question.

References.

Bloom, Benjamin, ed. Taxonomy of Educational Objectives. New York: Longmans, Green, 1952.
Chaffee, John. Thinking Critically. Boston: Houghton Mifflin, 1997.
Piaget, J. The Child's Conception of Number. London: Routledge and Paul, 1952.
Rishel, T. Writing in the Math Classroom, Math in the Writing Class; or How I Spent My Summer Vacation Using Writing to Teach Mathematics, ed. A. Sterrett. MAA Math Notes 16. Washington: The Mathematical Association of America, 1992.
Vygotsky, L. Thought and Language. Cambridge: MIT Press, 1986.

Return to Top

The Active Classroom

Show up early, maybe by five minutes. Say hello. Cheerfully. Start handing out homework or new handouts. Ask a general question, like "How's it going?" Or, "Was the homework too hard this time?" If the response is "I couldn't do number seventeen," say, "OK, I'll do that one on the board when class starts." If it's "Yes, the homework was too long," then tell them, "OK, I'll do some at the beginning." If you're doing a recitation and not lecturing, you can ask if the lecturer is up to date on the syllabus--you may well know the answer to this question, but at least you will elicit a response. This is preferable to telling the students where they're supposed to be in the text and which homework you'll be discussing today. By this time, more students are filtering into the classroom, and you can bring them into the same preclass conversation.

If the students want to talk about their flu, the last class, or last night's basketball game, that's fine--until the prescribed time for your class to start. Then say, "Well, it's about time to start. Did anyone have a homework question, or something from the last lecture that you didn't understand?" Or tell them, "Jenny said she couldn't do number seventeen, so I promised I'd do that one. Did anyone have a question before that?" Suppose no one does. Ask Jenny to remind everyone what question seventeen says--that way she can talk--then ask all the students present if "anyone has an idea how to start the problem." If no one says anything, don't just start solving it, offer a hint, like: "...this is a section on parametrically defined functions; what might that have to do with the problem?" At that point, someone usually suggests, "Well, I started with the formula for 'parametric derivatives'." Respond, "Right. That's the one that goes dy/dx = (dy/dt)(dt/dx), just like the chain rule, right? So how does that fit the problem? " In this way, you are getting the students to tell you what they know, not just about this problem, but also about their comprehension of the recent material from the course.

There are a number of general ways to keep the classroom active: Ask leading questions--ones whose answers are not simply yes, no, or "square root of two." When a topic depends on some earlier concept, ask the students to provide the earlier information and formulas so as to show you what they remember. Once you show how to do one problem, choose a similar one and ask students to work on it communally. Then, when they have all had a chance to begin solving it, go to the board and write down what they tell you is the method of solution.

You can also tell the students a bit about the history of the topic or one application of it, and then ask them to bring in more of the history or another application for the next class. Then you can start off the next time with what you have found out. Don't rush through your own answers as if time were the enemy. Give everyone time to think of whether they understand your calculations and whether they need to ask about the seemingly trivial steps you thought too easy to write down there you wrote 2+1, then over there you wrote 3, why?

Every topic was new to each of us at some point; we had to think hard about what made it work. Then when we found out, we began to think it was trivial. But it isn't--not for the new student, not for the person who didnÕt see it well the first time out--so we should give everyone a chance to ask all the questions the smart ones, the wrongheaded ones, the ill-thought out ones, the ones we should have asked the first time we saw the material. If we can set up a classroom atmosphere where our students can ask all these questions, then we will be a long way toward being a "good teacher" whatever that concept may ultimately mean.

This has only been a minimal presentation a first--case-scenario--of methods for making a classroom more responsive. More interesting and advanced techniques include the use of worksheets and cooperative strategies like having students work together. Students can also make presentations, both small, like individual problems, and large, like final projects, of work they have done.

As you get further into this topic, you will want to consult such references as Bonwell [4], Cohen [8] and McKeachie [18] for more advanced suggestions.

Return to Top

"What Was That Question Again?"

When you lead recitations, you will find that you open yourself to all sorts of questions. This situation is one of the most anxiety-producing ones in teaching; "I have absolutely no idea what they will ask. How can I handle that?" Let's examine some of the types of questions you will be called upon to answer:

The standard question.

An example: "Can you do number twelve?" If you prepared to before going to class, you can even answer, "Sure." Just make sure that you then do it.

The question that makes no sense.

Everyone asks such questions on occasion; resist the impulse to put the questioner down. Instead, think about how to turn the question into a good one, maybe by responding with, "Maybe what you are asking is...?"

Of course, it is also possible that the listener asked a meaningless question because what you thought was a perfectly clear explanation was opaque to him or her. Or else he or she was daydreaming though part of your previous discussion. In any case, you now have been given an opportunity to reinforce points you (thought) you made earlier.

The silly question.

Don't make a big deal of it. Act as if it's an honest question. Answer it quickly, then move on.

You will occasionally have a student who seems to specialize in asking silly questions. Other students will roll their eyes as soon as they see his or her hand go up; resist the impulse to "side with" the others by smiling, joking or answering with a smirk on your face. Such behavior on your part is simply unprofessional, even if you know that some of the students are going to downgrade you on evaluations for your "allowing too many stupid questions."

The unintelligible question.

You might simply say, "I don't quite understand. Could you rephrase that?" Or, "...Are you asking about...?" Then try to rephrase the question into something sensible.

Alternatively, you can ask someone else to try to rephrase the question.

The "challenge to your authority" question.

I often get these in first semester calculus. "Let's see if he knows what they taught me in my last week of intensive calculus at my high school." Or, "Let me ask him how to do the hardest problem in this section, even though it wasn't assigned for homework."

I answer these questions slowly and carefully, if I can. If I don't remember the answer, I will respond, "That wasn't part of the assignment, but I'll be glad to show you in the next class." Then I make sure to do so. I resist the impulse to turn the question back on the student by asking, "Did you try it? Then what's the answer?" This last might set up an adversarial situation, one where you are either perceived as knowing the solution but unwilling to show it to the students, or as being someone who isn't really open to answering student questions.

The "good question."

Hooray! A good question. Say, softly, honestly, "That's a good question." Then answer it. By publicly recognizing good questions, you encourage more of them.

The question you don't have any answer for. This is everyone's nightmare, and this nightmare will sometimes come true. So what? Just respond, "I don't know." Then ask the audience what they know about the topic. You might just learn something new. Remember, anyone can ask anyone else a question they can't answer.

On the topic of questions and answers, I am reminded that for some years I have had a large cartoon poster on the back of my office door. A large beaked avian in a dress--"Ms. Bird," perhaps--is standing amid a circle of cute, fuzzy, small animals, who are looking wide-eyed at Ms. Bird as she intones, "There is no wrong answer, Malcolm, but if there were, that would be it."

Questions and answers are an integral part of learning. Our method of handling them is important to our effectiveness in our teaching, and ultimately in our careers. It behooves us to get used to them, to think about them, to encourage them, and to enjoy what they can teach us about ourselves. Sometimes we will even be surprised at how much we actually know!
Return to Top

Motivating Students

Too many faculty interpret the word "motivating" as "pandering"; dressing up as Isaac Newton, say, telling silly jokes that are out of character, or giving out A's as if they were jellybeans. Now, far be it from me to claim that I've never told silly jokes, and I don't give out A's like jellybeans, but I have given out Pringles potato chips to illustrate hyperbolic manifolds, and...but I digress.

To me, motivating means addressing the history, culture, and usefulness of mathematics.

You don't have to get a second bachelor's degree in history to insert a bit of information into your calculus class about Newton and Leibnitz, or about Bishop Berkeley and his feud with Newton over infinitesimals. You can also spend a few minutes reading up about Euler's treatment of exponentials, Cauchy and Weierstrass on limits, and Bolzano about continuity. There are a number of references you can use for this material, for instance, Boyer [5] and Klein [16]. For a more advanced approach, try Edwards [13].

Cultural aspects of mathematics are also related to the history. Students enjoy hearing about the ancient Greek approach to infinity, and how it would have affected their willingness to accept the eighteenth and nineteenth century approaches to calculus, infinity and the infinitesimal. Further, the fact that such ideas continue to be questioned and refined into the twentieth and twenty-first centuries makes the students feel that their own skepticism about these concepts is relevant and valuable.

Students are also highly interested in how mathematics applies to their own fields of interest. Here, the principal of "Show, Don't Tell" takes over; start a section on second order linear ordinary differential equations with a model of a shock absorber from a car. Discuss the principle of damped oscillation. Then derive the differential equation for the model, discussing possible benefits and shortcomings of the assumed linearity of the system of equations. Once you have solved the system, don't stop there discuss the meaning of the various constants as it applies to the comfort level of the ride of individual automobiles.

The same principle applies to using Fourier series in solving the heat and wave equations, to using linear algebra in describing inventory control, to exponential functions in drug prescription, And but you get the idea. If you can't think of examples, just look at some of the harder problems in your current textbook; chances are that there are some good applications there. In closing, recall that students are always asking for motivation; you are likely doing the same with respect to your first-year analysis course. If you are skeptical of what I just said, simply take note of the number of times you hear -- or, in your analysis class, you think -- that famous question, "What's this good for?"

Return to Top

How to Solve It

One of the ways in which you can be helpful to your students is to offer general suggestions as to how to solve the types of mathematical problems they will be seeing throughout the semester.

Many years ago, George Pólya wrote a book called How to Solve It [22], in which he addressed the same question. Many people have used Pólya's model since then (only a few with attribution). Pólya suggests the following. To try to solve a mathematical problem:

  1. Read the problem.
  2. Read the problem again.
  3. Draw a picture or diagram.
  4. Find and label the unknowns what are you looking for.
  5. Find and label the known quantities.
  6. Write down all the formulas and relations between the known and unknown.
  7. Solve the problem.
  8. Check the answer.

And here I might add a suggestion:

Think about how you might generalize the problem.

In his book, Pólya offers a number of individual problems -- geometry problems, word problems and related rates problems from calculus, and others that he solves by using his methodology. Many calculus texts, for instance, Stewart [24] and Thomas-Finney [25], do the same in their discussion of word problems. Instead of my taking up space showing how they do this, I will just recommend that you take a look at these books for details.

Let me make a few suggestions about how you might approach the teaching of such difficult topics as word problems and related rates problems for the students:

Exercise:

Choose two problems from the textbook you are using next semester. Solve these problems slowly and deliberately using Pólya's method--no shortcuts allowed! Does his method fit these problems? Is the fit perfect, just adequate, or not at all? (For instance, people often complain that "a picture just isn't needed for this problem.") How would you modify Pólya's approach to fit your problems?

Return to Top

Course Evaluations

No doubt you have been on the student side--the "giving side"--of course evaluations. Now you will see the "teacher side."

Course evaluations can be extremely useful in telling you how your individual class of students has perceived your teaching during a specific semester. Did the students see you as "organized"? Even if neither they nor anyone else can tell you exactly what "organized" means, the students have an opportunity to offer an opinion.

Were you "helpful"? Does that mean, "did you answer questions without insulting the students' intelligence?" Or that you had lots of office hours, even though most students never came? Or that, heaven forbid, one afternoon you showed them a "preview copy" of the exact same exam as the one exam administered that evening? Now that last would be "helpful"--but not in the sense that a good faculty member would like to see.

Were you "knowledgeable" about the material? Of course, you could be successfully completing a course in complex analysis with a grade of "A+", and still have the students in your calculus section saying that you weren't.

In view of all the above skepticism, what is the function of the student evaluation process?

Well, first of all, the process does have uses other than the ones just discussed. Take, for instance, the question of "knowledge." If you know lots about functions of several complex variables yet can't give an intuitive response to the question, "Why does the ratio test work?" then your calculus students are of course going to see you as someone who doesn't know much about calculus. ("He's really a nice guy, but...") Alternatively, if your response to the question about the ratio test is to give a rigorous proof of the test, then the students are likely to give you high marks for knowledgeability--and low ones for helpfulness. ("She knows a lot of math, but she can't bring it down to our level.")

As to the question of "organization," this one is tricky. I have personally thought a lot about what it means, because I am consistently rated lower in this category than any other in my over thirty years of college teaching. I believe that when students talk about organization they seem to mean, "He has a plan in his head as to where we'll be at the end of each class, he tells us what that plan is, and he gets there almost every time." I must admit that, if this last is the students' "definition," then I don't conform to their ideal. Instead, I am willing to take questions at (most) every opportunity; I am happy to revisit earlier concepts if students show that they don't know them; I sometimes make up more than one lesson plan in my more "nonstandard" courses, and then let the students questions and interest dictate which one I use on a particular day. In return for this last, I occasionally begin class by outlining where we have been, and I do periodic reviews to show the students where we have come to.

I also take some comfort from the fact that one of our college's previous teaching award winners once told me, having seen my evaluations, "Don't worry, Tom, I always get 'low marks' for knowledge, too."

Student course evaluations are thus useful, without being a complete determiner of teaching ability. They tell us what the students expect of us. They teach us something about the expectations of our audience. Maybe most importantly, they allow us to find out our own classroom goals, and how these goals conform to or conflict with the ones enumerated on our evaluation forms.

In short, we need to know how we are being evaluated, where those evaluations go, and what each of the qualities listed on the evaluation form has to do with our approach to teaching.

At Cornell, we give a nuts and bolts evaluation form to all TAs and faculty. They are asked to hand out and collect this form very early in the term, usually around week three, but they are specifically encouraged not to turn it in to any official or semiofficial entity in the department. The function of this evaluation is to let the instructors find out how I'm doing in various categories, from "Do I speak loudly enough?" through "Do I give enough time for questions?" to "Are my answers intelligible?" The full form is given at the end of this section, along with a few other typical evaluations.

Return to Top

Get Along with Colleagues

If there is only one piece of advice you should take to heart from this entire volume, it is that embodied in the title of this section: Get along with your colleagues.

When you are told to come to the "calculus one" meeting, do so. Grade in a timely fashion, and do it in such a way that others will not have to field two-hundred student complaints.

Do your share of the busywork involved in administering the course; offer to give occasional makeups; to run off five hundred copies of exams; to teach once or twice for a sick colleague. Show up for office hours; don't leave it to the other TAs to offer excuses for you and then have to help your students in your place.

Be nice to people, even if they aren't nice to you who knows who they are, or what kinds of problems they may be having at the time. And, if they really aren't nice, then at least be formal and professional. You have every right to choose your friends, but colleagues are more like family you have little or no choice in the matter. If you really can't stand someone, ask yourself why. If you think you have a good reason, fine; chalk it up to experience, and move on to get the job done as quickly as possible so as to be out of the way.

Don't speak ill of fellow TAs, about faculty or about administrators to your students. As a friend of mine said the other day, "Oh my goodness. I didn't realize that that person I was complaining to about the dean was his spouse!" If you have a professional problem with a colleague or co-worker, take the problem to a supervisor if that person is trustworthy. If not, try your graduate student director or the chair. Say something like, "I don't want to cause a problem or get anyone into the middle of an argument with another TA, but something has come up about (say) an ignoring of some possible cheating [say] on the last exam, and I need to talk it out."

By the way: Yes, I have seen a (very) few TAs who were successful in the classroom, but who were so difficult to deal with as colleagues--always arriving late, leaving all their TA duties to others, skipping classes, etc.--who were told to leave. So "brilliance" is no guarantee of support for a teaching assistant. And, believe me, it usually gets even more difficult for an uncooperative faculty member. So learn to be cooperative now, or expect to pay a lot more later.

I have also seen a number of TAs and faculty who are "selectively nice." If a tenured faculty member is asking them for a favor, that's fine, but not a secretary. (You have seen this too, where a faculty member treats you badly because you're "just a TA.") Please don't be like this. In many ways, the staff runs the university. They were at their desks before the chair became chair, and they will be there when the chair has gone back to being a lowly tenured professor. They know how to cut corners, and they can be helpful to people who are courteous--and coldly uncooperative to those who are not. More to the point, they too just want to be treated like human beings, just as TAs do. So treat them that way.

'Nuff said.

Return to Top

What is a Professional?

The first time you step in front of a class, you cross an invisible line. You don't see it, but the students do; you are no longer one of them. That's why they look at you quizzically when you ask, What should we do today? They also don't appreciate your little jokes about exam grades. And, when you write a cute comment on their homework about how this work is more like high school stuff they see the comment as acerbic, and they let you know.

A professional is one who speaks for and has responsibilities to the discipline he or she is teaching and to the other practitioners of that discipline. Some of these responsibilities have been described in the section called Get Along with Colleagues, but not all. There is more to being a professional than speaking courteously to an officemate, as important as that is.

You have responsibilities to students:

You will also have responsibilities to the faculty and other TAs:

Most importantly, you have responsibilities to mathematics itself:

Return to Top

Teaching Methodologies for Various Types of Classrooms

Without trying to be exhaustive in my list, let me say that there is a variety of teaching methods, among them:

Each of the above has positive and negative aspects. For instance, the lecture method imparts lots of information in a short amount of time; in the proper hands it is usually organized well; and it makes use of the expertise of the lecturer. On the other hand, students can easily "nod off" in a lecture, and information imparted is not necessarily information received.

Socratic dialogue is often touted as an active, open form of learning that gets students involved in the educational process. On the other hand, such discussion can lead nowhere. It is also often "falsely democratic" in that a few speakers can dominate discussion, either crowding out other viewpoints or allowing for participation by only a few.

I claim that a guided discussion is one that has more structure than the Socratic model, perhaps by using information sheets or reading lists. Handled properly, such guided discussion can overcome many of the objections to Socratic dialogue listed above, and can achieve the objective of finding out what the student knows. Yet, students and faculty sometimes complain that such a technique isn't fast enough in imparting knowledge. Better prepared students, especially, often object to having to wait for students who "just don't get it."

Then there is "recitation," or student-guided learning. This methodology makes maximum usage of the student and of the text, putting the "burden of proof" (not to mention "burden of effort") on the students. This method doesn't necessarily make effective use of the expert, and if guided discussion operates slowly, this method can really operate slowly. On the other hand, say those who advocate this technique, once students do "get it," there is no doubt they have it--whatever "it" is. A final argument usually offered against student-guided learning is that it can miss the most salient items involved in a course.

Thus, what we need to ask ourselves is: What is the goal we want the students to achieve, and what is the best methodology for achieving that goal?

Return to Top

Lesson Planning

As a TA,

While doing homework before class, think about:

Also as a TA, you may also be called on to discuss:

Some pitfalls TA's can experience:

Generally speaking, students just want to get through the course with a good grade and go on to what's important -- to them. They will accept, for the most part, work that is relevant to their goals, and their questions will reflect those goals: "Why is this important?" "Will this be on the exam?" You, for your part, should see both these questions as fair at any time in the course, and should be prepared, within reason, to answer them. Possible answers to the first question: "It's used later, in..." "There are engineering (business, economics, architectural) applications, and if you learn about this technique, it will make you better in that field." "The author needs it later in the chapter (or in the text), when we solve...." And the last "fair" answer: "I'm really not sure. Let's see if we can find out why the author and instructor want us to do this."

If you think about irrelevance while you are preparing your recitation at home, you will be ready to answer most of the questions that come up in class. Plus, you will be active in your preparation, which will make you more interesting to the students, and your material more interesting to all of you. You'll eventually become the kind of instructor the students "inadvertently praise": "I didn't think I'd like the material or the course, but he/she showed me why it was useful. I still don't like it that much, but now I see why it's important."

Return to Top

Problems of, and with Students

Chances are, if you are a sensitive, caring instructor, students will begin to see you as a "lifestyle advisor." A typical first reaction is to feel honored--but then an almost immediate response sets in; "How can I give advice to this person--or anyone, for that matter?"

Good. You should feel skeptical. Of course, you are an intelligent person and a trained mathematician. And that's the point; you are trained as an intelligent mathematician, not as a counselor or therapist. When someone comes to you to say, "I'm afraid I'm suicidal," how should you respond? How about as the intelligent person you are: "Thank you for coming to talk to me about it. It's pretty obvious that this causes you some pain, and I'm glad youÕre willing to speak to someone about it. I'd like to help you--would you mind if I called the university counseling service for you? I think they can provide some real support."

The vast majority of student