Twitter is my muse. This blog post was inspired by this Tweet:    In a study comparing blood concentrations of caffeine after coffee or energy drink consumption, blood caffeine levels peaked at about 60 minutes in all conditions. Plan accordingly. https://t.co/cWWakGGtHe pic.twitter.com/c5Nn3x3w1f — Kevin Bass (@kevinnbass) November 30, 2021 This study is straightforward to follow. I, personally, think it is psych-friendly because it is about how a drug affects the body. However, it doesn't require much psych theory knowledge to follow this example. Sometimes I'm worried that when we try too many theory-heavy examples in stats class, we're muddying the waters by expecting too much from baby statisticians who are also baby psychologists. Anyway. Here are some things you can draw out of this example: 1. Factors and levels in ANOVA The factor and levels are easy to identify for students. They can also relate to these examples. I wonder if they used Bang energy drinks? They a...

Flowingdata is great. They create lots of exciting data visualizations and share other people's visualizations.  This visualization from Flowingdata is especially significant.   I think it illustrates base rate fallacies beautifully. Moreover, it is applied to a very crucial issue: Immunizations. The base rate fallacy has been used repeatedly to attack the efficacy of vaccines . In particular, instances when vaccinated people catch diseases for which they have been vaccinated. Frequently, such arguments fail to consider base rate data regarding how many more people are vaccinated.  This illustration from Marc Rummy is elegant and straightforward and explains a mathy/sampling/statsy concept without any actual math. I love it.  Also, this illustration has been updated recently with a bit more text to explain everything: Apparently this picture I made that was part of a post 4 months ago recently went viral. Here's a new & improved version that includes the explana...

 OK. I'm proud of this one. I think these are good chi-square examples. And they are about Americans' beliefs about various supernatural beings. Presented with commentary because I couldn't help myself. 1. Supernatural beliefs by age Already. I fully intended to tease my traditional UGs about their beliefs about vampires. Because I do believe that would probably be a significant chi-square...but... ...IS GEN X OK? This survey asks about ghosts, demons, psychics, vampires, and werewolves. What are the "OTHER" this survey is talking about? Aliens? Dolly Parton? I'm intrigued. 2. Werewolves: The Unity Horse. The Unity Wolf? Both Trump and Biden supporters are united in their belief that werewolves do NOT exist.  Anyway. This survey is intriguing. There is a lot of material to work with.  Go read it here .  PS: Hey! If you like this idea and would love a whole stats textbook from the brain of the person who came up with this idea, sign up for more information abou...

There are plenty of stats calculators all over the interwebz. Power calculators, calculators for every test statistic under the sun, etc. BUT: Omni Calculators   I would try to list all of the different calculators, but I just don't have that sort of time on my hands. Descriptives? Sure. Inferential? Why not. Risk? Certainly. For more on the company that has created this calculator, read up on Omni Calculator here . ALSO: There are plenty of non-stats calculators available on the website, with a total of 2,122. 

This blog post was inspired by the brilliant Leslie Berntsen. Her ACT 2021 presentation,  " I'm not a therapist: Mental health education and advocacy for non-clinicians," was about teaching and sharing sound practices in mental health care when you, yourself, are not a mental-health-care-type-psychologist.  Anyway, she had this really great idea about sharing user/trends/etc. data from the Crisis Text Line  (CTL) with students. CTL is a text-based mental health crisis hotline. It is staffed 24/7. I include it on my syllabus as a resource for my struggling students because, unlike my uni's very hard-working counseling center, CTL is always available.  https://www.crisistextline.org/ Why share data from CTL in your stats class? Because you can use it to a) teach a bit of stats, b) introduce students who need this resource to the resource, c) introduce potential volunteers to the service. Here are my ideas for how you could use this data in class. 1. Data visuali...

As I write this blog post, Francis Haugen testifies on Capitol Hill and sheds light on some of Facebook's shady practices. TL;DR- Facebook realizes that its practices are support terrorism.  This led to a public relations blitz from Facebook, including Monika Bickert, who appeared on CNN . Of particular relevance is repeated reference made to an Instagram survey of 40 teens ( here is the documentation I was able to find ). I saw this tweet from Asha Rangappa Reaction about one of those interviews: https://twitter.com/AshaRangappa_/status/1445487820580081674 LOL I had to listen to this twice to make sure I didn’t mishear: This @Facebook exec repeatedly refers to a “survey” of FORTY teen Insta users— as in 4-0 — to support her assertion that the “majority” of teens have a great experience on the platform. For real. Listen to it https://t.co/Ye7ocWcnzG — Asha Rangappa (@AshaRangappa_) October 5, 2021 This example packs a lot of punch. It is a good one for the youths because it is ...

Maybe this is the sort of thing you will like. I guarantee you, it is spook-tacular.   https://docs.google.com/presentation/d/1Gu4VYsYTNJ28RtBf6cNp1xszjo-bfItSHAp-C0PiEbY/edit?usp=sharing

A man after my own heart. This is of minimal educational value but maximal stats humor. David purchased a Pure Drink water bowl for his cat. He wanted to know if it actually resulted in his cat drinking more water.  This wee (hahahaha) little study could be used on the first day of class to demonstrate: 1) A hypothesis 2) Operationalized variables 3) Within-subject research design  4) p (HAHAHAHHA)-values 5) What a god damn stats nerd their instructor is 6) The power of data visualization

Okay. I know there are so many COVID examples out there, but this one is maybe a tiny bit amusing (it involves Berlin dance clubs). It also demonstrates a within-subject research design and ecological validity. It is also a very tiny example that is easy to understand and doesn't require students to understand any psychological theories. Yes, many of you are psychologists teaching statistics, but I think it is vital that we use various examples to ensure that at least one of them will stick for every student. Emma Hurt/NPR Anyway. Berlin has a famous dance club culture , which has been under tremendous financial strain due to COVID-19. Since winter is coming and outdoor options will no longer be possible, the government has sponsored a pilot project to study whether or not clubs can be opened safely if everyone at the club has tested negative for COVID-19. NPR reported on this applied, within-subject design study  (a three-minute-long news story you could use in class): In addition...

Nicole Gilbert Cote, Jared Schwartzer & Natasha Matos created a great website, The Accessible Toolbox,  filled with ideas for creating an accessible statistics class. In addition to advice, they offer A FREE (thanks, APS!) 3-D Tool Kit for teaching stats , which includes :

HEY. Here is a whole bunch of data, searchable via Google.                       https://datasetsearch.research.google.com/ h/t: Samy ! 

First, a pun: American athlete Athing Mu broke the American record for the 800m. I guess you could say...that Mu is anything but average!! HAHAAAHAHAHHA. https://twitter.com/Notawful/status/1409456926497423363 Anyway. It is late June 2021, and my Twitter feed is filled with amazing athletes qualifying for the Olympics. Athletes like Sydney McLaughlin. That picture was taken after McLaughlin a) qualified for the 2021 Olympics AND b) broke the 400m hurdle world record. Which is amazing.  Now, here is where I think we could explain effect size interpretation. How big was McLaughlin's lead over the previous record? From SpectrumNews1 McLaughlin broke the world record by less than a second. But she broke the world record so less than a second is a huge deal. Similarly, we may have Cohen's small-medium-large recommendations when interpreting effect sizes, but we always need to interpret an effect size within context. Does a small effect size finding explain more variance than any pre...

The Pudding  took a data-driven approach to test a popular hypothesis: Women's pockets are smaller than men's pockets.  Authors Diehem and Thomas sent research assistants to measure the pockets on men's and women's jeans. They even shared supplemental materials, like the exact form the RAs completed. https://pudding.cool/2018/08/pockets/assets/images/MeasurementGuide.pdf And they used fancy coding to figure out the exact dimensions of the jeans. Indeed, even when women are allowed pockets (I'm looking at you, dressmakers!), the pockets are still smaller than they are in men's jeans. They came to the following conclusion: Amen. Anyway, there are a few ways you can use this in the classroom: 1) Look at how they had a hypothesis, and they tested that hypothesis. Reasonably, they used multiple versions of the same kind of pants. If you check out their data, you can see all of the data points they collected about each type of jeans. They even provide supplemental mat...

One thing I like to emphasize to my students is that just because a scientist is using math and science and statistics, it doesn't mean they are unbiased. I usually describe how Sir RA Fisher love statistics, smoking, and white folks and, shock of shocks, produced data that supported both the safety of smoking and the soundness of eugenics.  For more on that: How Eugenics Shaped Statistics, by Clayton for Nautilus Magazine . And now I have another plug-and-play, easy-to-implement example of checking your bias in your research. http://journals.sagepub.com/doi/abs/10.1177/0098628320979879 The article makes a sound argument: While social justice/bias issues may be present in other psychology courses, they need to be addressed in our stats classses as well.  This journal article from Teaching of Psychology suggests that a lecture that highlights Samuel George Morton's "research" that investigated skull size and intelligence, as well as more modern examples of bias, leads ...

 Hi! If you want to hear what my actual voice sounds like as I talk about the teaching of statistics, check out this interview I did with Eric Landrum for the Psych Sessions podcast.

 https://today.yougov.com/topics/lifestyle/articles-reports/2021/05/13/lions-and-tigers-and-bears-what-animal-would-win-f Sorry that I haven't been posting as often lately. You would think that with the summer, I would have more flexibility, but I am working hard on some writing deadlines (for a stats textbook!), and my kids' activities have picked up considerably with soccer season starting. This example illustrates fun data visualizations as well as a t-test. YouGov is a polling company, sort of like Gallup. They collect many Very Serious polls and silly polls like  this one, where they asked participants to state whether or not they could beat 34 different animals (from rats to grizzly bears) in an unarmed fight. Their graphic designer deserves a raise for this bar graph, including several tragic humans vs. animal memes/movie clips. Here are a few lessons you can draw out of this funny data. Paired t-test example: They took the participants identified as men and women and...

It occurred to me that I haven't shared my absolute most precious professional development and stats teaching resource in the blog.  That resource is the Society for the Teaching of Psychology. Non-psychologists reading this post, don't stop now. Keep going.   1. There are multiple free e-books available from STP. Some are about teaching, broadly. Some of them have a chapter or two devoted to the teaching of statistics. But there is at least one exclusively devoted to the teaching of statistics,  For the Love of Teaching Undergraduate Statistics . I wrote a chapter in the book, so I'm partial.  2. The STP journal, Teaching of Statistics , includes pedagogy research about the teaching of statistics . Full disclosure: I'm a consulting editor at this journal. 3. If you are a member of STP and come up with an excellent teaching idea or study idea related to teaching statistics to psychology majors, STP has got some money for you . They have several grants, reviewed...

Hi! Here is all the material I shared at my keynote address at Teaching of Statistics in Psychology. I included three exercises, with slides, you could use to teach confidence intervals, chi-square, and t-tests. Here is the talk:

Dr. Morton Anne Gernsbacher has freely shared her WHOLE Psychological Statistics class with the world. No paywalls, no log-ins.  Divided up sensibly, including sections on effect sizes and Bayesian. With some assistance from Chelsea Andrews, she created this phenomenal resource for all of us to use. While I'm focusing on her stats class, be sure to check out her amazing Research Methods and Psychological Effects of the Internet courses. DO NOTE: I'm not primarily recommending this directly as a resource as a class for students. I'm recommending it to you, instructors, as a source of great, free statistical readings and teaching ideas that you could incorporate into your own classes. This is especially helpful if you are trying to teach without a textbook or if you are just looking for additional ways of explaining tough statistical concepts to your students. Here are a few things she shares that you could use: 1. Plenty of guidance for teaching via Excel/Google Sheets/App...

I've blogged about Our World in Data before. There is a lot to appreciate at this webiste, but I would like to draw your attention to the wealth of interactive COVID visualizations you can create. Many of these visualizations include a toggle button that changes the graph from a logarithmic graph to a linear graph. Which really, really helps illustrate log data transformations to our novice statisticians. There have been occasional dust-ups over the last year with people not understanding the difference  or  being unfamiliar with log transformations , or graphs not being appropriately labeled.  I also like this example because most of my examples skew towards American content, but this data visualization tool lets you select from many countries .  ANDPLUSALSO: There is data to be had at the website. Data for days!

 I found this Tweet, which leads to a brief report on BBC. A recent report from the World Obesity Federation shows COVID death rates are higher in countries where more than half the population is overweight. Cause and effect, or bad statistics? @TimHarford and @d_spiegel explore - with some maths from me. You can listen on @BBCSounds https://t.co/hevepmz8RC — stuart mcdonald (@ActuaryByDay) March 14, 2021 The BBC has a show called "More or Less," and they explained a recent research finding connecting obesity to COVID 19 deaths.  Here is the original research study . Here is a pop treatment of the original study . For more stats news, you can follow  "More or Less" on Twitter . And they cram, like, a half dozen lessons in this story. It is amazing. I've tried to highlight some of the topics touched upon in this story. How can you use it in class? I think it would be a good final exam question. You could have your students listen to the story, and highlight ...

Go follow Statsystem ( Facebook , Instagram ). They are a smarty pants who dreams and thinks in stats memes. I know these are silly and hilarious and fun. I think these are good for a nerd chuckle for ourselves. And who doesn't need more nerd chuckles? I also think these are light, funny, accessible ways to back up a more complex stats lesson with a succinct meme that conveys the guts of a lesson.  See: Power Regression: Central Limit Theorem: T-tests: And don't forget the big picture:

This teaching example: 1. Is psychology research. 2. Features the actual data from the generous and helpful Dr. Bray . 3. Features GIFs. EVERYTHING is better with GIFs. 4. Includes puppies. 5. Includes a good ol' Psych Statistics standard: The one-sample t-test. Okay, get ready. I first learned about Dr. Emily Bray's dog cognition research via Twitter . Never let it be said that good things don't happen on Twitter. Occasionally.  1 Dogs are known for their ability to cooperate with humans and read our social cues. But are these skills biologically prepared? To find out, we tested 375 puppies at 8.5 weeks on 4 social cognition tasks (task descriptions: https://t.co/aETequNBce ) #AnimBehav2021 #Cognition pic.twitter.com/7vN2lp82Dp — Emily Bray (@DrEmilyBray) January 27, 2021 This is such a helpful way to share your research. This example works for your Cognitive or RM classes as well as your stats class, since this thread illustrates not just her findings but her methods. T...

When I teach inferential statistics, I think it is helpful in providing several conceptual (no by hand calculations, no data analyzed via computer) examples of experiments that could be analyzed using each inferential test. I also think it is essential to use non-psychology examples and psychology examples because students need to see how stats apply outside of psychology. At times, I believe that students are convinced that a class called Psychological Statistics doesn't apply outside of psychology.  So I like this quick, easy-to-follow example from medicine. Thomas, Patel, and Bittel (2021) studied how different vitamin supplements affected outcomes for people with COVID-19. The factor (COVID intervention) has four levels (usual care/control, ascorbic acid, zinc gluconate, and ascorbic acid/zinc gluconate). And the four groups acted pretty much the same. Bonus stats content: Error bars, super-cool Visual Summary of a research study that really highlights the most essential parts...

Stuart Ritchie wrote an excellent book about the problems (and solutions to those problems) in science called Science Fictions . Illustrator and author Zach Weinersmith summarized and illustrated those problems in science in the form of a short webcomic . Both the book and the comic are great and have a home somewhere in your psychology curriculum. The comic is a quick, digestible primer on the problems with science. Meanwhile, the book goes into great depth, including many of the problems related directly to p-hacking, fishing expeditions, etc.  For more, read Stuart's book, Science Fictions. Aside: I am American, and he is Scottish and reads the audiobook version to me while I walk my dog, and it is soothing. I'm creating a one-credit course for our Honors program based on this book.  Also, follow Stuart and Zach on The Bird App.

FlowingData shared some interesting info on how much cars cost version their environmental footprint .  TL;DR: Low emission cars tend to be cheaper in the long run. Hooray for the free market! The data is also available via the New York Times , along with a much more in-depth conversation about the actual cost of high/low emission cars, but it is behind a paywall. The original data, presented in a fun (nerd fun) interactive website , is available here.  How to use it in class: 1) It's a correlation! Each car model is a dot with two related data points: Average cost per month and average carbon dioxide emissions per mile.  2) It's Simpson's Paradox! Note how electric cars (yellow cloud all have similar emissions, but the average cost per month varies. Same for Diesel cars. Overall, you still see the positive correlation in the data, but if you break it down by class of car, the correlation isn't present for every level. 

I gave one of the keynote addresses at the STP Teaching Pre-conference this year. I'm so sad that we weren't all able to come together face to face, but I am so glad that I was invited to give this talk at a pre-conference I've attended off and on for years. Here is the PowerPoint. There are a bunch of links. TL;DR: You should do discussions in your stats classes. No, don't talk about degrees of freedom or bar graphs for an hour. Instead, help your novice statisticians see data (and your class!) IRL. A few details on my discussion days, and how it goes: I spend a full 55-minute period on the discussions. My students must submit a brief reflection piece about the readings. Here is how I describe the reflection piece and Discussion Day in my syllabus: Here is how I present the Discussion Day materials to my students:

If you are trying to explain the Chi-Square Test of Independence to your students, here are some timely examples that are political and not polarizing. Well, I don't think it is polarizing. I'm sure there are people out there that disagree. Maybe some of the questions are polarizing? Regardless, it is nice to have an example that uses a current event with easy to understand data.  The example comes from  CNN. The network conducted exit polling during the 2020 presidential election . I'm sure they didn't intend to provide us with a bunch of chi-square examples, but here we are. Essentially, CNN divided Biden and Trump voters into many categories with not a parameter to be had. I have included a few of the tables here, but there are many others on the website .  They illustrate different designs (2x2, 2x3, 2x4, etc.) and different magnitudes of difference between expected and observed values.