The grading scheme was as follows:
Participation through iClicker 3% (drop 2 lowest)
Homework 12% (drop lowest)
Quizzes 15% (drop lowest)
Midterm 20%
Final 30%
Team Data Analysis Assignments (x2) 20%

I wish that I could leave separate reviews for Professor Garfinkel's lectures and the rest of the course (the homework and labs), because they were like night and day.

Professor Garfinkel is extremely intelligent, funny, and personable. He explained how modern statistics was made to justify racist, eugenicist views and how the downfall of good statistics led to the downfall of good science (meaning that almost all research findings are irreproducible). He was very engaging in lecture and helpful in office hours, and if the entire class was just sitting in lecture and listening to him talk, I would have gladly had 4-hour lectures.

What made the class so unpleasant was the coding. Professor McCully is the new coordinator for the LS30/LS40 classes, and I think it is largely her influence that contributed to the excessively long homeworks, labs, and team data analysis assignments. I spent at least four hours on the homework by myself and two hours in office hours seeking help on the homework every week. As another review noted, sometimes even the TAs weren't sure what was expected for the code. This culminated in a 118-point homework assignment that was designed to be extra long because "we hadn't had homework in two weeks" - because we were taking THIS CLASS'S midterm and midterm data analysis assignment.

The midterm and final were extremely conceptual (almost no actual math). This could be a good thing, except the midterm was graded extremely harshly. If you didn't have exactly what they wanted (which often included answering unasked questions), they'd take off points, which really started adding up. In office hours after the midterm, there were so many red marks on everyone's Gradescopes that it looked like Christmas.

Near the end of the quarter, we started getting two labs per lab section - that's right, labs A and B. We couldn't finish them during our discussion section, which then made completing the homework (and then the data analysis assignments) very difficult. We get it - it's difficult to check code on Campuswire, but many of McCully's responses were downright rude.

I know I shouldn't be complaining too much since I got an A+, but they didn't input grades until Spring Quarter had already started (and, I believe, ran out of time to grade the final data analysis assignment as harshly as they'd graded the midterm data analysis assignment).

Professor Garfinkel is great, but he is very little of the time you actually spend on this class. The homeworks and team data analyses were designed like LS40 expected itself to be the only class you're taking this quarter - I genuinely don't know if they thought we could code faster or if they just didn't care how long it took us.

Alan Garfinkel is one of the best professors I've had at UCLA. He's an extremely caring and empathetic professor who is genuinely passionate about the material that he teaches. This is the second class I've taken with him and he's been consistently good across the board.

Don't get me wrong. LS 40 isn't exactly an "easy" class. The content is a lot more conceptual and you'll need to focus on really learning and understanding what's going on in order to do well. You'll also need good Python skills that you've hopefully developed in LS 30A/B.

The exams are pretty fair imo, I Garfinkel doesn't do "trick" questions or ask anything that's outside of what you learn. If you follow the lectures you'll be fine.

Summary: This is a great class with a great prof. The skills you learn are really useful for getting research positions too (ie. p-val NHST, Confidence interval testing, Bayesian stats). Come and be inspired.