Today's post is one, small, to-the-point example of sampling bias. On May 27, 2020, my family and I were awaiting lift-off for the (subsequently grounded) NASA/SpaceX launch. To no one's surprise, I was following NASA on Twitter during the hoopla, and I noticed this Tweet: https://twitter.com/NASA/status/1265724481009594369 And I couldn't help but think: That is some sampling bias. Admittedly, their sample size is very impressive, with over 54K votes. But this poll went out to a bunch of people who love NASA so much that they follow it on Twitter.  What is a less biased response to this question? As always, Pew Research Center had my back. 58% of Americans responded that they definitely/probably weren't interested in traveling into space: https://www.pewresearch.org/fact-tank/2018/06/07/space-tourism-majority-of-americans-say-they-wouldnt-be-interested/ If you want to expand upon this example in class, you could ask your students to Google around for information on the ...

The Washington Post , and numerous other media outlets, recent wrote about an unintended consequence of COVID-19 and the sudden drop off in commercial flights: Fewer data points for weather forecasts ( PDF ). Due to the coronavirus, commercial flights are down: How does this affect weather forecasts? Data is constantly being collected from commercial flights, and that data is used to predict future weather: Ways to use in class: A conceptual example of multivariate modeling : Windspeed...temperature...humidity...lots of different data points, from lots of different elevations, come into play when making our best guess at the weather. This is a non-math, abstract way to discuss such multivariate models. A conceptual example of effect sizes/real-world effects: In the article, they clearly spell out the magnitude of the data loss. That is pretty easy to track since we can count the number of flights that have been canceled. More complex is determining the effect size of this data loss....

Hi! This post is a change of pace. Instead of providing an example to use in stats class, I'm going to share how I incorporated text-message based class discussion in online courses with the GroupMe App. Doing so was a big win for me during a hard semester, I hope it is a big win for anyone who happens to read this post and use GroupMe in the future. My experience using GroupMe App to facilitate class discussion during The Rona My goals for OL SP20: I wanted my students to learn. I wanted to preserve the best parts of my classes. I didn't want my classes to be another burden in a stressed out world. During March 2020, I wast teaching Introduction to I/O Psychology. It was a class of 20. My students were mostly Juniors and Seniors, who were either psychology majors or minors. On our last day of f2f class, when we knew that we were going to transition to OL, I asked my students to reserve our normal 12:20-1:25 MWF meeting time for the class. I wanted my students to continue to h...

Katherine Baicker ,  Oeindrila Dube ,  Sendhil Mullainathan ,  Devin Pope,  and  Gus Wezerek created an interactive, data-driven piece for NYT . It provides a new perspective on how we should proceed with re-opening businesses during the COVID-19 pandemic. They argue that we must consider 1) how long people linger in different types of stores, 2) how often they visit these stores, 3) the square footage of the stores, and 4) the amount of human interaction/surface contact associated with how we shop at different stores.  How to use this in class:    1) Show your students how data can inform real-life problems. Or crises, like how to safely re-open stores during COVID-19. 2) Show your students how data can be used in creative ways to solve problems. The present argument uses cellphone location data. 3) Show your students data viz in real life: Here, scatterplots that really improve the #scicomm potential of this piece. 4) Show your students the rese...

This brief video from NPR (they make videos, what?) describes how the CDC will be randomly sampling Atlanta residents to test for COVID-19 antibodies. The efforts hope to provide a better estimate of the spread of the disease. H/t to Sy Islam for sharing this with me. I think you could use this in class as a super-fast example of how we use samples to generalize about larger populations. The CDC is sending out employees to conduct antibody tests on a random sample of Atlanta residents. The tests are meant to show how many people have been infected with the coronavirus. pic.twitter.com/mXqznHUJmV — NPR (@NPR) April 29, 2020

Hi! Here is an easy-to-use review of regression, with a regression example. I posted about using this catty tweet to teach regression previously: Right? Do you see the regression? Every wife has two data points: Year she gets married, and the year she was born...and from that, you can perform a regression to predict when Quaid would actually marry someone who has not been born yet (2052). Well, I decided to make it into a whole Google Slides presentation for this example, with links to data, to use as a regression review during the quarantine. Admittedly, the example is ridiculous, and the sample is far too small to run a reliable regression. That being said, I think the example is vivid and sticks. I also think it does an excellent job of illustrating how the equation can be used to make predictions. Additionally, I genuinely find meaning in helping out my fellow statistics instructors in good times, and doubly so during this challenging semester. Feel free to view my pres...

Here are two good review examples for chi-square, one for goodness-of-fit, and one for the test of independence. Here is my Google Slide presentation, which includes links to data sets for the examples. One features Taco Bell . The other features actual Developmental Psychology research, as featured on NPR . When I use these in class, the students have already been introduced to chi-square, have been walked through examples of both chi-squares, and then they analyze the data on their own using JASP. Y

I hope everyone is hanging in there. Here is a pretty straight forward one-way ANOVA example that is interactive, based on for-real personality psychology research, and interesting. I blogged about this previously but whipped up a Google Slideshow you can download and edit to suit your own teaching. Also, I uploaded data that you can use with your students.  TL:DR- A bunch of researchers gave the NEO to 1.5 million Americans to determine if different regions of the US have different personality trends (see research here ).  Original Study They do. Then Time magazine reported on the study . And the scicomm was beautiful. They accurately described the research AND created a fun interactive portion in which you students can take the NEO-Short Form and be matched with the state that best matches their personality (Hi, I am West Virginia because I'm high in neuroticism and low in openness to new experiences, which are great qualities to have during a pandem...

I apologize in advance if you are pandemiced out. It is just that my brain won't stop seeing stats examples in information related to the COVID-19 pandemic. For instance, researchers are looking back at the 1918 Flu Pandemic in order to forecast how social distancing (or lack thereof) will affect mortality rates now. And these patterns, as illustrated by National Geographic, demonstrate different data distribution shapes . The data comes from a reputable source, is scaled to deaths per 100,000 as to allow for comparison, and the distributions are related to very important data. Other lessons your students can learn from this data: This is what good scicomm looks like. Also, sometimes a good data visualization is better than an accurate-yet-filled-with-jargon version of the same information. For instance, much has been shared about NYC vs St. Louis in terms of timing of quarantine. Here is the comparison yet again, but in an easier-to-follow description: There is a ton of...

TL:DR: Imma send my students to YouTube to watch three MythBuster clips that approximately illustrate t-test research designs. Then, they will identify the t-test research design that is illustrated by each of these clips.  More detail. MythBusters is a show that gleefully creates research to test urban legends and random questions that arise in day to day life. The questions that my clips tackle are: a) how badly do people drive when distracted by hands-free cell phones, b) could Indiana Jones have really made it through the chamber at the beginning of Temple of Doom and c) what is faster: Weaving in and out of lanes or staying in the right-hand lane when driving? So, they will watch the clips, and I will ask them questions (they will submit their answers via Google Forms Quiz) to make sure they can tell which sort of t-tests you would use to analyze the data, given research design. Here is the PPT I will use. I've never used this exact clip in class before. I di...

Hello, my friends. We live in strange times. I will try to share everything so that this transition to online education is more manageable. My university is pausing teaching for this week (3/16-3/20), so we can prepare. I've prepared my first lesson. I've used this lesson before. I like this lesson because you can use it with ANY of the inferential statistics taught in Intro Stats. Here is my PPT for the projec t. I used it for an Independent t-test, but you can use it for any test. What you need: A FB account, a way to share a lecture with your students, a way for your students to share ideas about what sorts of variables you can glean from FB. What your students need: A FB account (I suggest you create an extra data set to share with students who don't have one), a way to conduct statistical tests (by hand, calculator, JASP, Excel, R).  Tips:  1. I don't think this is an excellent way to introduce a statistical test, but I think it is a usef...

Psi Chi, the psychology honor society, is sponsoring a fun, free, low-commitment way to help your students (and maybe you?) learn R. I talked with Jordan Wagge, one psychologists spearheading the project (along with John Eldund and April Staples), and my understanding is that there will be multiple cycles of this class, using different data, research questions, and inferential statistics (so, if your students can't do this right now, a new cycle will start Late Spring). Each cycle will run the course of three months. There will be an assignment due in the middle of each month. This class would be great for any graduate school-bound undergraduate. Here is the formal intro from Psi Chi . A good place to get started is this Google Doc that outlines the whole contest and the process you/your students will go through, step by step. Here are all of the materials , hosted on OSF. Also, if you successfully complete it YOU GET A STICKER. And I find that the UGs love stickers. N...

Hi, friends. There is a free, new edited volume that contains 18 terrific chapters of ideas and hints and guidance for teaching undergraduate statistics. Drs. Beyer & Peters but so much work into gathering ideas from great teachers and now we can all enjoy it.  I wrote Chapter 1. In it, I make an argument that we need to teach statistics to our psychology UGs in a way that better matches how they will use the statistics. TL:DR: 4% of psychology UGs get PhDs in psychology. So maybe we should reconsider how and who we are teaching. Download it here:  https://teachpsych.org/resources/Documents/ebooks/love-teaching-undergrad-stats.pdf