Skip to content

Quick advice for Sam and anyone else who’s teaching a new statistics class for the first time

1. Figure out your educational goals–what you want your students to be able to do once the semester is over. Write the final exam–something that you think the students should be able to do when the course is over, and that you think the students will be able to do.

2. Show the exam to some colleagues and former students to get a sense of whether it is too hard (or, less likely, too easy). Alter it accordingly.

3. Structure the course around the textbook. Even if it’s not such a great textbook, do not cover topics out of order, and avoid skipping chapters or adding material except where absolutely necessary.

3. Set up the homework assignments for each week. Make sure that these include lots of practice in the skills needed to pass the exam.

4. Figure out what the students will need to learn each week to do the homeworks. Don’t “cover” anything in class that you will not be having them do in hwks and exam. (You can do whatever you want in class, but if the students don’t practice it, you’re not really “covering” it.)

5. Write exams and homeworks so they are easy to grade. I also recommend having a short easy quiz every week in class. The students need lots of practice. This is at all levels, not just intro classes.

6. Every lecture, reinforce the message of what skills the students need to master. (To get an idea of how the students are thinking, picture yourself studying something you’re not very good at but need to learn (swimming? Spanish? saxophone?). There are some skills you want to learn, you want to practice, but it’s hard, and it helps if you are cajoled and otherwise motivated to continue.)

This is not the only way to go, of course; it’s just my suggestion for a relatively easy way to get started and not get overwhelmed. Especially remember point 1 about the goals.

Some helpful resources:

First Day to Final Grade, by Curzan and Damour. This is nominally a book for graduate students but actually is relevant for all levels of teaching. Especially since I think that “lectures” should be taught more like “sections” anyway.

Teaching Statistics: A Bag of Tricks, by Gelman and Nolan


  1. Bill Meisel says:

    I'm curious to know if you think teaching hypothesis tests the traditional way (i.e., wihtout discussing p-values) is sacrilege. I am teaching at a school where students do not have easy access to technology. I have been teaching the p-value approach for so many years that I am having doubts about dropping it, even though it will make things easier for my students (i.e., it's one less calculation for them to make. )

  2. Lisa says:

    There's a resource on the web that's useful for statistics teachers. It's It has lesson plans, applets, that sort of thing.

    As for the hypothesis testing question, I started out teaching only the traditional method. It works fine. But I've been moving more and more to p-values simply because that's what's reported more often in research literature. If you don't have access to technology, you can estimate a p-value (ie. between .01 and .05) with the old tables. But you might explore the use of TI-83s or web pages like StatCrunch that will do these calculations for you.

  3. anon prof @ tufts says:

    these are great tips, thanks a bunch. i have been teaching stats for three years and these encompass a lot of "unwritten" rules that are in my head.

    the "teaching statistics: a bag of tricks" has been very helpful as well.