Skip to main content

Posts

Mark Rober's 14 minute long primer on machine learning

I'm a fan of former NASA engineer and current YouTuber/science comm pro  Mark Rober . He meets the sweet spot of containing YouTube content that is safe for kids but also engaging for adults. You may know him for creating obstacle courses for squirrels in his backyard and holding the world record for the tallest elephant toothpaste explosion .  Recently, I discovered that he made a stats-adjacent video  explaining machine learning by studying baseball signals and creating a way to de-code baseball signals . Anyway, if you touch on your topics in your classes, this is a great, quick explainer. It is well-edited, well-produced, and has captioning. You don't need to be a baseball fan to follow this example. 

University of Pittsburgh's National Sports Brain Bank

 I have written about the NFL's response to concussion data as a case study of how to obfuscate data. This has been covered in many places, including in The Atlantic and on PBS . In my experience, concussions are a prime source of conversation for traditionally college-aged students. Many of them were high school athletes. Fewer are college athletes. Most college students have personally experienced a concussion or loves someone who has. Now, the University of Pittsburgh is opening the National Sports Brain Bank . This is for athletes, not just football players. Two former Steelers have promised their brains, as have two scientists who played contact sports.  Here is a press release from the University of Pittsburgh . Here is a news report  featuring the two Steelers who have promised to donate their brains. However, as described by Aschwander, we still don't know how many football players have CTE (please read this piece, it is such good stats literacy from Aschwander...

"Why randomized controlled trials matter and the procedures that strengthen them" from Our World in Data

Looking to freshen up your readings for Research Methods? Or for a good, brief RM primer for a stats or psych class? Check out Our World in Data's "Why randomized control trials matter and the procedures that strengthen them" . Added bonus: Our World in Data dived into their data archives to illustrate each piece with their own research. I don't know about you, but my brain far prefers abstract concepts paired with concrete examples.  Some of the classic include: -Why we need RCT. https://ourworldindata.org/randomized-controlled-trials#what-are-randomized-controlled-trials -Why causal inference is hard. https://ourworldindata.org/randomized-controlled-trials#the-fundamental-problem-of-causal-inference -Why we need control groups. https://ourworldindata.org/randomized-controlled-trials#the-control-group-gives-us-a-comparison-to-see-what-would-have-happened-otherwise

A simple tool operationalizes post-childbirth hemorrhaging and saves lives.

 https://www.npr.org/sections/goatsandsoda/2023/05/10/1175303067/a-plastic-sheet-with-a-pouch-could-be-a-game-changer-for-maternal-mortality https://www.bmj.com/content/381/bmj.p1055 I love this study, in and of itself, because it is based on research that will save women's lives without spending a lot of money. I love it.  Here is a link to the original study . I learned about it from an NPR story about the research by Rhitu Chaterjee . I also love it because it is an accessible example of a bunch of statistics things: Dependent variables...operationalizing variables...why cross-cultural research and solutions aren't just lip service to diversity...how control groups in medical research are very different than control groups in psychology research...absolute vs. relative risk. -Dependent variables/operationalized variables: This study clearly illustrates the power of measurement and operationalization. The researchers wanted to create a way to better assess post-childbirth h...

CDC Mental Health Data

It shouldn't come as a shock that the CDC shares data on rates of public health issues in the US.  However, you may be unaware of the available data and interactive visualizations provided by the CDC and the different ways you can use them in class . 1. Teach your students a lesson about good sources for mental health data. 2. Show your students how data visualizations can help present and simplify complex data. https://www.cdc.gov/nchs/covid19/pulse/mental-health.htm 3. Get into the research methods. Everyone has heard of the census, but fewer have heard of the Household Pulse Survey (https://www.census.gov/data/experimental-data-products/household-pulse-survey.html). The US Census collects much information between the 10-year census, including mental health data. https://www.census.gov/data/experimental-data-products/household-pulse-survey.html 4. Talk about how the government assesses depression and anxiety. For example, you can show how the basic methodology uses a valid, relia...

MCU regression, revisited

I think it is important to emphasize how regression can be used to make future predictions using trends in existing data. Most psychology books use psychology examples to illustrate this, which makes sense. Still, I think explaining how regression is widely used in business to make financial decisions, and predictions is important. But that can be boring. But I found one example that uses the Marvel Comic Universe to do this. I already blogged about this , but I'm sharing exactly how I used this in class presently. ASIDE: This data is being regularly updated! Here is a Google Drive folder with 1) my version of the data (CSV and I turned all the percentages to decimal points for JASP) and 2) my PPT . Which includes photos of the scientists of the MCU. ALSO: While your students are doing their exercise, totes play the soundtrack from Guardians of the Galaxy. Do it. 

A rank ordering of the Taylor Swift songbook.

File under: End of the semester stress blogging about a person who brings me joy. Taylor Swift (see: sampling error with Taylor ). Here is a new, VERY accessible example of ordinal data . Rob Sheffield, writing for Rolling Stone, rank-ordered ALL of Dr. Swift's songs.  https://www.rollingstone.com/music/music-lists/taylor-swift-songs-ranked-rob-sheffield-201800/bad-blood-2014-196114/ Also, introduce your students to Methods Section 😁. This rank order is based on the variable "Taylor genius". You could even use this as an example of anti-interrater reliability. This ranking comes from exactly one person. AND YOU'RE ON YOUR OWN KID DESERVED BETTER. Each ranking includes the best lyric from the song as well as a brief description of the Taylor Genius on display. Is this also an example of qualitative data? https://www.rollingstone.com/music/music-lists/taylor-swift-songs-ranked-rob-sheffield-201800/the-great-war-2022-1234617639/