Kelly McConville is a survey statistician who develops estimation techniques that combine complex survey data with big data sources. Her work is used to estimate official statistics, related to canopy cover or occupational statistics, or to assess the impact of voter ID laws. She enjoys teaching her students how to learn from data and introducing them to R. She also involves her students in her work and co-chairs two national programs: the Undergraduate Statistics Project Competition and the Electronic Undergraduate Statistics Research Conference
Allan Rossman has worked in the Statistics Department at Cal Poly – San Luis Obispo since 2001. His primary classes include introductory statistics to students from throughout the university, and he has also taught courses in probability, simulation, and mathematical statistics.
Larry Lesser is a Houston-raised El Paso-based, award-winning educator, researcher, author, and speaker who integrates diverse backgrounds and multiple intersecting identities. For example, his passion for combining music and STEM made its way into his research, grantwriting, teaching, outreach, and service, and about 170 of his published poems and songs are STEM-related.
Episode Description
The Consortium for the Advancement of Undergraduate Statistics Education, aka CAUSE has held the United States Conference on teaching statistics, also known as USCOTS every other year since 2005. This conference enables teachers of statistics to exchange ideas and discover how to improve their teaching. The theme of this year's conference was communicating with and about data, a topic near and dear to us on the Stats and Stories podcast. Two sub-themes are explored as part of this conference, helping students to communicate the process and results of their statistical analysis, and helping teachers to communicate with students in order to develop their understanding of statistical concepts and their ability to implement statistical methods for conversations with leaders and speakers at the United States Conference on teaching statistics were recorded on site. And we are happy to feature these in a collection of episodes of Stats and Stories.
+Full Transcript
John Bailer
The Consortium for the Advancement of Undergraduate Statistics Education, aka “CAUSE” has held the United States Conference on teaching statistics, also known as USCOTs or USCOTS, every other year since 2005. This conference enables teachers of statistics to exchange ideas and discover how to improve their teaching. The theme of this year's conference was communicating with and about data, a topic near and dear to us on the stats and stories podcast. Two sub themes are explored as part of this conference, helping students to communicate the process and results of their statistical analysis, and helping teachers to communicate with students in order to develop their understanding of statistical concepts and their ability to implement statistical methods for conversations with leaders and speakers at the United States Conference on teaching statistics were recorded on site. And we are happy to feature these in a collection of episodes of stats and stories, where we explore the statistics behind the stories and the stories behind the statistics. I'm John Bailer stats and stories is a production of Miami University's Department of Statistics and media, journalism and film as well as the American Statistical Association. Guess where I am the US Conference on teaching statistics. Some say USCOTS. Some say USCOTS. Alan Rossman will certainly tell me why I did it wrong. I'm joined here by Kelly McConville of Harvard University and Alan Rossman of Cal Poly San Luis Obispo. They are the program chairs for the US Conference on teaching statistics 2023 Or as the primitives like me, say us cars. So, thanks. Thanks for taking the time out of your busy schedules to talk about this wonderful conference.
Alan Rossman
Thanks very much, John. I do prefer to call it uscots. But lots of people call it USCOTS. The people who call it USCOTS. Like the fact that US implies everyone in it together. And that's certainly one of the one of the attitudes and themes and poses we try to strike with us cuts.
John Bailer
So yeah, well, I wonder since I, this is my first time I hate to say this. So should I have called it you cots prior to this as a consequence?
Alan Rossman
No, no. Call it whatever you want. We're just glad you're here.
John Bailer
Absolutely. Well, I'm curious. You two are now the
Alan Rossman program shares, right?
John Bailer
This is this is what what a great thing. And thank you for the effort that you put into it. And for all of us that are able to join you were excited to see the program you put together. You had a theme that you pick this year.
Kelly McConville
Yeah. So our focus this year is on communication. So it feels kind of fun to be on the podcast given the theme. But specifically, it's about communicating with and about data. So thinking about communication, both, you know, extracting knowledge from data and then communicating that in an effective way to stakeholders or peers. But then communicating about data. So understanding difficult concepts like confounding and variation and being able to talk about, about those those difficult ideas.
John Bailer
Yeah, it's it seems that there's a lot of job security and miscommunication just because of so much of quantitative information. It seems like it's misrepresented and misunderstood. It's a very important part of what you do and what you're trying to encourage among the community. So Can Can you talk a little bit about, you know, as you think about this theme, why now, why did that come up is important now for this year in this program?
Alan Rossman
That's a good question. And I think communication is always important. And we haven't addressed that as a theme before. So partly, it was just the time to do it, because we haven't addressed it before. And it's critically important. Another aspect I think, is we try to appeal to a wide variety of types of teachers. We have teachers from two year colleges and four year colleges and research universities, some high school teachers, some graduate students who are thinking about pursuing careers in teaching. So we always try to choose a theme that's going to be relevant to statistics teachers, no matter what situation or environment they're in. And we think communication certainly fits that bill.
Kelly McConville
And maybe I'd add you know, as the data and the models get is messier and more complicated, maybe communication and being able to boil things down into simple terms becomes more difficult, but also just that much more important.
John Bailer
I'm curious as you reflect on your own work and on your on the practice of statistics, and in education and statistics over the last number of years, what have been some of the big changes that that you would note,
Alan Rossman
I've been at it for 35 years. So there have been lots of changes. One is certainly the power of computing and the accessibility of computing, that certainly changed how data is analyzed, that certainly changed how people teach you everything, and certainly how they teach statistics. I think in the last five to 10 years that the the advent of data science is starting to have a huge impact on statistics and on everything else. And I think that's going to be more and more impactful in the coming decade.
Kelly McConville
One thing I'd add to that is multivariate thinking. So I think a lot of introductory statistics courses used to focus on, you know, extracting knowledge from just a single variable or the relationship between two variables. But that's not the world we live in. Right. We live in a very multivariate world. So trying to help students grapple with finding relationships and in many variables and sort of sifting through the noise to find signals in those those multivariate situations.
John Bailer
Okay, so now I want you to put on your prediction hat, you know, so this is where you get to make some incredibly bold prediction of the future of the next big thing in teaching statistics and data science. What might that be?
Kelly McConville
Why that's tough, it is tough. What year was it? Alan, where the theme of us cards was the next big thing?
Alan Rossman
I thought it was 2011, maybe 2013. And that was the theme. And so all the speakers pretty much made their predictions, were they right? Well, the consensus then what consensus is a strong word, but lots of people predicted that simulation based in Prince would be the next big thing. And I think to some degree, they were right. I think simulation based in France has become much more widely used in the classroom in the last decade. And the idea has been around for decades and decades. But I think in the last decade, more and more teachers, even especially if introductory statistics, are using simulation to help students understand basic ideas of statistical inference. So I think to some extent, the folks who predicted that 1012 years ago were on the mark, I think it's harder to make predictions about the future than about the past. It's maybe I'll defer to Kelly for the
Kelly McConville
one actually looking towards the past, but a more recent past. So four years ago, so last time us cats was in person, the theme was around hypothesis testing and p values, right, this was around when the ACA had come out with their statement around key values. And I, I don't think we've radically changed how we teach hypothesis testing, even if that was the theme four years ago. And I wonder if that will, will come to pass in the future, right, more of an emphasis on talking about effect sizes or power, de emphasizing hypothesis tests and emphasizing more estimation and prediction. So maybe that is more of a hope of mine than necessarily what's going to happen. But I think another thing I would mention is just the the increases in modeling, right? Like the number of models out there, but also the widespread disciplines were using much more sophisticated models. And I mean, it's across the spectrum, right. So I think a very classical statistical models like a linear mixed model, I think are more commonplace all the way up to you know, those fancy predictive models like a neural network. And the more these become commonplace in other disciplines, the more I think they're going to need to bleed into our statistics classes earlier and earlier than they currently do so. So I would predict that to be another big thing is to think about how we can get more sophisticated modeling techniques into earlier earlier statistics classes.
John Bailer
Okay, I know I have one, one last sort of two parter question for you both. What is your favorite thing to teach? And what is your least favorite thing to teach? And no, no know, opting out and going back in time? You got it? You got a
Alan Rossman
good question. My favorite thing that I'm almost embarrassed to say this, maybe a statistician shouldn't say this, but my favorite thing to teach is probability. We have a course in Cal Poly that's called Introduction to probability and simulation. So we have students approaching probability from a simulation perspective and from a mathematical perspective, and I really enjoy helping students to look at Questions about randomness from both of those perspectives. They do a good bit of coding as well as doing a good bit of analytical traditional probability analysis. So as I say, I'm somewhat embarrassed to give that answer because probability is a mathematical concept. It's not so much a statistics concept. But that's, that's currently my favorite course to teach. Least favorite is harder, I'll ask tell it to go next.
Kelly McConville
All right, for favorite, I would definitely say it's data wrangling, which is a topic that's pretty new to my classes. But I think it's to me, I drank the Kool Aid. And I'm convinced that we should be teaching introductory students data wrangling because out in the real world, they're not going to see perfectly clean textbook datasets that don't have missing values and have all sensical observations. And so we need to teach them early on how to wrangle their data into a format that actually they can build models build visualizations, do summaries. And so I think it's a lot of fun. And I think it's fun, because there's like a certain subset of students, I think of these as the students who love to have a really tidy room who like dig into data, and get so much joy out of it. And so So yeah, I think for me, that's, that's probably my favorite. I would say, for my least favorite. So, you know, I talked about Kool Aid. One other drink of Kool Aid that I did take was related to the simulation based inference, Alan talked about earlier. So I teach simulation based inference in my intro classes. But I still feel like after that, I have to quickly teach my students about t tests and that sort of thing, because I know they'll be exposed to that and in other disciplines, but feels like this somewhat redundancy. That that I'm just not that excited about. I don't like half to talk about is n greater than 30. That's a silly metric to be checking and that sort of thing. So So yeah, that one gives me not much joy. All right. Now, I'm
John Bailer
hoping you forgot.
Alan Rossman
What I was thinking before you gave your answer about least favorite was pretty similar. Just getting into all the details of procedures. And, and frankly, I tried not to do that so much tried to emphasize the students the general principles and the basic reasoning approach. That's the same across multiple procedures. And like Kelly, I'd spend some time on the details, but I try to minimize time on the boring details.
John Bailer
Well, very good. I mean, I think I echo the same kind of responses. When I think about things that I really enjoy. I like probability to
Alan Rossman slightly less embarrassing
John Bailer
You and I can join together.
Alan Rossman data wrangling sounds more fun, it
John Bailer
was fun to Well, it's been a great pleasure to have Kelly McConville and Alan Rossman join us on stats and short stories, the very special edition where a roving reporter is wandering the halls of a conference center in State College, Pennsylvania to talk about this. Thanks again.