Greg and Patrick talk about median splits and other ways that continuous variables might be categorized to simplify analyses, and the often very high cost of doing so. Along the way they also mention Never Have I Ever, hand cuffs, gold hoop earrings, bungee jumping, Erik, Henry and Penny, having a mini, chug jugs, Hill […]
S3E11: The Cons and the Cons of Median Splits Read More »
Patrick and Greg answer a mailbag question about some foundational principles of regression and correlation, specifically the differences among correlation, standardized regression weights, semi-partial correlations, and partial correlations. They also check in with someone from Season 2: Amanda Montoya’s undergrad mentee, Kat. Along the way, they also mention: Ap Olo Gies, mocking by text, particle
S3E09: Semi-Partially Clarifying Measures of Association in Regression Read More »
What starts as a friendly check-in with our intern Ethan McCormick in the Netherlands turns into Ethan’s Pop Quiz Revenge around the themes of statistical paradoxes and distributions. Along the way, Greg and Patrick also discuss headlamps, airport carousels, prostrate vs. prostate, operatic ambulances, ungulates, how X-mas got its X, Cobra Kai, bromeo, metric time,
S3E05: Pop Quiz 3: Ethan’s Revenge Read More »
Patrick and Greg talk about the potential advantages and disadvantages of alternative approaches to analyzing two time-point data. They discuss traditional models for raw and residualized change scores, and describe how each of these remains baked into the soul of contemporary models for repeated measures data. Along the way they also mention: 24 inch pizza
S3E04: Two Time Point Data: What Is Your Quest? Read More »
In this episode, Patrick leads Greg and the rest of us through the incredibly cool and powerful topic of regions of significance, a way to help unpack under what moderating conditions treatments and variable relations are statistically significant. Along the way, they also mention: parallelism, Olivia Newton-John, going to the wrong church, laws in Vermont,
S3E02: Johnson-Neyman Regions of Significance Read More »
Greg and Patrick explore the many interesting issues related to centering predictor variables prior to analysis, both in the standard regression model and in the multilevel model. Along the way they also discuss honking Diet Coke through your nose, rashes, hostage negotiation, reading at seven months of age, dealer’s choice, Coors Light, Galapagos tortoises, friendly
S2E30: ‘Always Center Your Predictors!’ And Other Sh*t My Advisor Says Read More »
In this episode Greg and Patrick embark on a sometimes unhinged discussion of outlying and influential observations: where do they come from, how are they identified, and what do we do with them. Along the way they also discuss the outlier bell, stubbornness, pre-holiday grievances, Polly, Sesame Street, being more less likely, Greg’s freak brother,
S2E22: Outliers — And Then Things Got Weird Read More »
In this episode Patrick and Greg compare and contrast multiple regression and the structural equation model and argue that although regression has brought us far, there are numerous reasons to turn to the SEM to continue to drive us forward. Along the way they also address gad flies, the Louisiana Purchase, Kool-Aid, high school friends,
S2E18: Regression — Like That Old High School Friend You’ve Outgrown Read More »
Greg and Patrick ring in the new year with a rambling discussion of univariate and multivariate normality: what it is, when and why we assume it, and does it really matter. (Spoiler alert — it DOES). Along the way, they also discuss statistics history, supplement #2, Larry the Book Shop Guy, Isaac Newton’s BFF, platypuses,
S2E17: Embracing Your Non-Normality Read More »
Greg and Patrick discuss issues of controlling for extraneous variables, by randomization and by design, but especially the challenges and assumptions associated with statistical control. Along the way they also mention drag racing, Volkswagen Bugs, Seabiscuit, Harrison Bergeron, Diet Coke, Janet Jackson, John Denver, gastrointestinal distress, Sewall Wright, and babies and bathwater.
S2E14: Control (Variable) Issues Read More »
