David Morganstein

Most of us became statisticians because we knew statistics is an exciting and stimulating profession—one in which curiosity and initiative intersect and we can make a difference in almost any industry that sparks our passion. But how many times have we been at a social event where someone asked us what we do for a living? When we replied “statistician,” we almost certainly got “The Look.” You have seen this look many times! It’s the look that says, “There is no way I could ever do that for a living.” And it’s the look that is almost always followed by stories of an incredibly boring or outrageously difficult high-school or college statistics class.

To address this all-too-often negative perception of our profession and help ensure that many students’ sole encounter with statistics is a more pleasant experience, Dick De Veaux accepted a homework assignment: He and his committee are creating “Stats 101.” It’s dynamic, relevant, real-life case studies for use as teaching examples, a kind of life raft for instructors that the ASA will offer to schools and universities.

Too often, the person asked to teach that introductory course in a nonstatistics program such as psychology, medicine, engineering, or business is someone relatively early in their career. They may have little or no hands-on experience actually conducting a data analysis. They may rely primarily on a textbook and present statistics as a set of formulas and procedures.

Dick De Veaux

Dick and his team are gathering exciting examples and describing the process of looking at and learning from the data. They’ll capture the students’ interest with the kind of detective work many of us have learned to do and that makes our jobs so relevant and interesting. Read on to learn more about what Dick has in mind.

What do you hope to accomplish with STATS 101? Who do you think will be most interested in it?

As we know, a huge number of introductory statistics classes, particularly in nonstatistics departments, are being taught not only by people who don’t have PhDs in statistics, but often by people who have neither taken a course nor worked in statistics. These instructors have had to learn statistics “from the book”—sometimes just before their students—and have never had the experience of carrying out a statistical analysis on a real problem. So, we’re developing a series of about 10 case studies that will show statistics in “action,” solving real problems and how all those techniques taught in intro stats work together. We’re hoping these will give the instructors not only a sense of how statistics is practiced in the real world, but also some extra examples to use in class. Our target audience is anyone teaching an introductory statistics course who has no training or experience in our field.

Tell us about the content. What principles will you include, and will you present them in a unique way?

There’s really nothing new here. The Data and Story Library (DASL) did something similar nearly 20 years ago. What we’re hoping to do is provide fewer, but deeper, stories and show how all the techniques one learns in an intro statistics course are actually used. Some of the studies will go further than the typical intro stats techniques—using multiple regression or logistic regression, for example—which will give the instructors a sense of where the intro course leads, as well.

Why did you take up this challenge? Who are your partners, and what interests do they bring?

A wonderful independent schoolteacher, Joe Cleary of the Loomis-Chafee school, lives part of the year in Williamstown, where I teach. I didn’t know him, but he came up to me one day and said he was using my AP book and confessed he really didn’t know much about statistics other than what he’d learned from the textbook. He asked if he could “shadow” me for a day or two and see what I really do. I thought it was a really great idea. After spending the day watching me do some data analysis he said, “You have to show this stuff to other high-school teachers. Now I see why you care about things that I didn’t think were important and don’t really care about some things that I did.” So, we developed a workshop for high-school teachers to give them some experience with complex data sets using both traditional and data mining techniques so they could really see what statisticians do today.

My colleagues for these case studies are Dave Bock, retired from Ithaca High School; Nick Horton of Amherst College; Danny Kaplan of Macalester College; Julie Legler of St. Olaf College; and Deb Nolan of Berkeley. They are all wonderful teachers with many years of experience who know the value of in-depth applications for motivating students in both introductory and AP Statistics.

Can you give us a preview of some of the case study topics?

We tried for a mix of applications. We have examples that range from a look back at the famous butterfly ballot of 2000 to examining patterns of bike use in the Washington, DC, bike-share system to modeling how likely your next flight is to be delayed. In each case, the choice of example has been based on availability of data, pedagogical impact, and general interest.

How is STATS 101 different from the “usual” introductory course?

Let me be very clear: This is not a “course.” There are many interesting introductory courses out there. We’re just trying to provide some examples for instructors who haven’t had the benefit of doing lots of statistical analysis to both give them some motivation and context and possibly provide some real-world examples for their students. As one of my colleagues in operations research who had to teach intro stats once said to me, “You make the course really interesting, but it’s unfair because you have all these great examples.” We’re trying to level the playing field just a bit by putting some on the web.

Will there be any video included?

We’re hoping to eventually have some videos going through the examples, as well, but that’s getting ahead of ourselves a bit. We first need to finish the case studies.

What do you consider as modern methods for teaching statistics, improvements over the traditional?

As we all know, statistics in practice is done using computers with some sort of statistical package, whether that be R, SAS, Minitab, JMP, or something else. I’ve never seen anyone in industry using statistics in their job computing summary statistics on a calculator or drawing a histogram by hand. Yet, we spend time in the introductory course teaching some of these skills. With students going through the Common Core, most of the beginning part of the traditional intro stats course will soon be redundant. We need to get on to the exciting parts of modern statistics—building complex models; simulation and resampling for inference; and discussing causation, reproducibility, and data collection in the Internet age.

Unfortunately, many people who had a single course in statistics came away with the impression that it is primarily about formulas and theory. We believe statistics is a vibrant profession because it is primarily about practice and application. We’re hoping to provide instructors with exciting, interesting examples that will make the subject more real to them and help make their teaching come alive and be more attractive and stimulating to their students.

Will you touch on Big Data in any way?

In a way, yes. We hope these examples provide a precursor to statistical thinking and get instructors and their students thinking about statistics as “problem-solving.” We want people to see statistics as part of the solution when confronted with big unstructured problems. All the data sets are “real,” and most are “modern” in the sense that they’re the kind of data now routinely collected.

The bike-use data are a great example. As people return bikes, several things are sent to the database—the station the bike was taken from, the destination station, demographic information on the user, and the length of time the bike was out. If you’ve ever used a system like this, you know one of the big problems is that too many stations are either empty—so no bikes are available—or full—so you can’t leave the bike. I’ve been late to a meeting because I couldn’t find a place to put the bike. How can we help the city do a better job at placing bikes and figuring out how large the stations need to be?

Are you hoping students who see these examples will “jump ship” and take up statistics?

I’m always hoping students will jump ship and study more statistics, and from what I see across the country, it’s happening even before we wind up posting these examples. I’m just hoping these provide some help for the instructor who feels a bit overwhelmed trying to understand a subject with which he or she has no real world experience. If we can help them understand what it is statisticians really do and how we think, I’ll be thrilled.

Dick and his committee are storing their materials on the ASA community site. They welcome your review and comment. If you have any reactions, suggestions, or just feel supportive, contact Dick at deveaux@williams.edu