As described in Wired's pop piece, the average body temperature for healthy adults isn't 98.6℉. Instead, data suggests that it is 97.7℉. Here is a link to the original study by  Hausmann, Berna, Ggujral, Ayubi, Howekins, Brownstein, & Dedeoglu . 1. This is an excellent theoretical example for explaining a situation where a one-sample t-test could answer your research question. 2. I created fake data that jive with the results, so you can conduct the test with your students. This data set mimicked the original findings for healthy adults (M = 97.7, SD = .72) and was generated with Andy Luttrell's Data Generator for Teaching Statistics . 97.39 97.45 97.96 97.35 96.74 99.66 98.21 99.02 96.78 97.70 96.90 97.29 97.99 97.73 98.18 97.78 97.17 97.34 97.56 98.13 97.77 97.07 97.13 9...

Explain validity to your students AND earn some "I'm still hip!" street cred using this The Tonight Show clip that features a Buzzfeed quiz AND exactly one Jonas brother. Nick Jonas took a "Which Jonas Brother are you?" Buzzfeed quiz. He scores as Joe Jonas. Ergo, the Buzzfeed assessment measure is not valid. It does not properly assess what it purports to assess. Watch the video for yourself. If you want to take this example a step further, you could have your students take the original quiz, discuss the questions and their ability to discern which Jonas Brother is which, you could describe Nick Jonas as a Nick Jonas Subject Matter Expert, maybe Social Desirability got in the way of Nick answering the questions honestly, etc. Another thing I've noticed as my blog and I have aged together: There are now generations of Buzzfeed quiz assessments that provide great examples for different age groups: Gen X: Shirley Manson did not score as Shirley ...

Emily Watson, writing for NPR, describes medical research by Ogilvie, vanNiekerk, & Krajden . This research provides a timely, topical example of false positives, false negatives, medical research, and gets your students thinking a bit more flexibly about measurement. This research provides valuable information about debate in medicine: What method of cervical cancer detection is most accurate: The traditional Pap smear, or an HPV screening? The Pap smear works by scraping cells off of a cervix and having a human view and detect abnormal cervical cancer cells. The HPV test, indeed, detects HPV. Since HPV causes 99% of cervical cancers, its presence signals a clinician to perform further screen, usually a colonoscopy. The findings: Women over 30 benefit more from the HPV test. How to use this example in class: - This is a great example of easy-to-follow  research methodology and efficacy testing in medicine. A question existed: Which is better, Pap or HPV test? The questi...

This is a remix of a bunch of posts. I brought them together because they fit a common theme: Examples that use actual data that researchers collected but still manage to lie or mislead with real data. So, lying with facts. These examples hit upon a number of themes in my stats classes: 1) Statistics in the wild 2) Teaching our students to sniff out bad statistics 3) Vivid examples are easier to remember than boring examples. Here we go: Making Graphs Fox News using accurate data and inaccurate charts to make unemployment look worse than it is. Misleading with Central Tendency The mean cost of a wedding in 2004 might have been $28K...if you assume that all couples used all possible services, and paid for all of the services. Also, maybe the median would have been the more appropriate measure to report. Don't like the MPG for the vehicles you are manufacturing? Try testing your cars under ideal, non-real world conditions to fix that. Then get fined by the EPA. Mis...

A tweet straight up solved a problem I encountered while teaching. The problem: How can I explain why the confidence interval area for a regression line is curved when the regression line is straight. This comes up when I use my favorite regression example.  It explains regression AND the power that government funding has over academic research . TL:DR- Relative to the number of Americans who die by gun violence, there is a disproportionately low amount of a) federal funding and b) research publications as to  better understand gun violence death when compared to funding and publishing about other common causes of death in America. Why? Dickey Amendment to a 1996 federal spending bill. See graph below: https://jamanetwork.com/journals/jama/article-abstract/2595514 The gray area here is the confidence interval region for the regression line. And I had a hard time explaining to my students why the regression line, which is straight, doesn't have a perfectly rectangula...

So, Fortnite is a super popular, first-person-shooter, massive multi-player online game. I only know this because my kid LOVES Fortnite. With the free version, called Battle Royale, a player parachutes onto an island, scour for supplies, and try to kill the other players. Like, there is way more to it than that, but this is my limited, 39-year-old mother of two explanation. And, admittedly, I don't game, so please don't rake me over the coals if I'm not using the proper Fortnite terminology to describe things! Anyway, my brain thinks in statistics examples. So I noticed that for each Battle Royale match starts with 100 players. See the screen shot: This player is parachuting on to the island at the beginning of the skirmish, and there are still 100 players left since the game is just starting and no one has been eliminated. Well, when we introduce our students to the normal curve and percentiles and z-scores and such, we tell them that the normal curve represen...

Patti Neighmond, reporting for NPR, wrote a piece on how the medical establishment's method for assessing patient pain is evolving . This is a good example of why it can be so tricky to operationalize the abstract. Here, the abstract notion in pain. And the story discusses shortcomings of the traditional numeric, Wong-Baker pain scale, as well as alternatives or complements to the pain scale. No one is vilifying the scale, but recent research suggests that what a patient reports and how a medical professional interprets that report are not necessarily the same thing. From Dr. John Markman's unpublished research: I think this could also be a good example of testing for construct validity. The researcher asked if the pain was tolerable and found out that their numerical scale was NOT detecting intolerable. This is a psychometric issue. One of the recommendations for better operationalization: Asking a patient how pain effects their ability to perform every day tas...

I'm such a sucker for beer-related statistics examples ( 1 , 2 , 3 ). Here is example 4. Now, I don't know about the rest of you psychologists who teach statistics, but I ALWAYS show the ol' Yerkes-Dodson's graph when explaining that correlation ONLY detects linear relationships but not curvilinear relationships. You know...moderate arousal leads to peak performance. See below: http://wikiofscience.wikidot.com/quasiscience:yerkes-dodson-law BUT NOW: I will be sharing research that finds claims that dementia is associated with NO drinking...and with TOO MUCH drinking...but NOT moderate drinking. So, a parabola that Pearson's correlation would not detect.  https://twitter.com/CNN/status/1024990722028650497

This is a fun question perfect for that first or second chapter of every intro stats text. The part with data distributions. And it works for either the 1) beginning of the Fall semester and, therefore, football season or 2) the beginning of the Spring semester and, therefore, the lead-up to the Superbowl. Anyway,  Ben Jones   tweeted a few bar chart distributions that illustrate different descriptive statistics for NFL players. https://twitter.com/DataRemixed/status/1022553248375304193  He, kindly, provided the answers to his quiz. How to use it in class: 1) Bar graphs! 2) Data distributions and asking your students to logic their way through the correct answers...it makes sense that the data is skewed young. Also, it might surprise students that very high earners in the NFL are outliers among their peers. 3) Distribution shapes: Bimodal because of linebackers. Skewed because NFL players run young and have short careers. Normal data for height because even...

If you are unaware, Nextdoor.com is a social network that brings together total strangers because they live in the same neighborhood. And it validates your identity and your address, so even though you don't really know these people, you know where they live, what their name is, and maybe even what they look like as you have the option to upload a photo. Needless to say, it is a train wreck. Sure, people do give away free stuff, seek out recommendations for home improvements, etc. But it is mostly complaining or non-computer-savvy people using computers. One of the things you can do is create a poll. Or, more often, totally screw up a poll. Here are some of my favorites. In the captions, I have given some ideas of how they could be used as examples in RM or psychomtrics. This is actually a pretty good scale. A lesson in human factors/ease of user interface use? Response options are lacking and open to interpretation. Sometimes, you don't need a poll at a...

As described here, legislatures in Puerto Rico attempted to take independent authority away from the Puetero Rican Institute of Statistics (PRIS), a governmental watch dog in charge of double checking statistics and research methods used by the government . This decision was made in order to streamline government, which is understandable. But it was also problematic because watchdogs need independence in order to have the power and safety to say unpopular things. Anyway, the legislatures ended up NOT streamlining PRIS's and taking away its authority, in part due to an outcry from other scientific agencies. How to use in class: -Statistics in real life, informing decisions, informing funding, being controversial. -Why do organizations like American Statistical Association and American Association for the Advancement of Science exist? Well, for a lot of reasons, one of which is t o publicly protests moves like the one PR tried to execute. -Statisticians and scientists aren...

Jodi Cohen's story about research ethics violations, and the subsequent pulling of $3.1 million in grant funding , is a terrific case study that shows your students what can happen when research ethics are violated. It is also an excellent example of good, thorough science writing and investigative reporting. Short version of the story: UIC psychiatrist Mani Pavuluri was studying lithium in children. She was doing this on NIHM's dime. And she violated research protocols. The bullet points, copy and pasted out of Cohen's article, are a summary of the biggest ethical shortcomings of the study: So NIHM asked for their money back ($3.1 million) and the university and research are now being investigated by the government. This example also highlights that IRBs are NOT just some rubber stamp for researchers. They are in charge of enforcing federal rules for research. Another interesting fact: UIC tried to block ProPublica from publishing the story. This w...

The Ingrham, writing for The Washington Post, used data to investigate the claim that undocumented  immigrants are a large source of crime.  You may hit a paywall when you try to access this piece, FYI. Ingraham provides two pieces of evidence that suggest that undocumented immigrants are NOT a large source of crime. He draws on a  policy brief from the Cato Institute and a research study by Light and Miller  for his arguments. The Cato Institute policy brief   about illegal immigration and crime is actually part of a much larger study . It provides a nice conceptual example of a 3 (citizenship status: Native born, Undocumented Immigrant, Legal Immigrant) x 3 (Crime Type: All crimes, homicide, larceny) ANOVA. I also like that this data shows criminal conviction rates per 100K people, thus eliminating any base rate issues when comparing groups. From: https://www.washingtonpost.com/amphtml/news/wonk/wp/2018/06/19/two-charts-demolish-the-notion-that-i...