Are you a health nut (aside from the cookies)?
No, no really. For one thing my exercising tends to come and go in phases that last a few months at a time. I guess when I am in a phase where I am running a lot I am pretty healthy overall, but as it stands now I haven't gone for a run in about 3 months. As for my diet, I have a few habits that are generally considered to be healthy -- e.g., I don't eat meat, try to limit my consumption of animal products, and I gave up drinking alcohol a little while ago -- but the relevant health issues were not a major factor in most of these decisions. Overall I would say that I am somewhat health-conscious, but quite far from being nutty about it.
Favorite equation and why?
Well this is more of a "conceptual" equation than a proper equation, but:
\(DATA = MODEL + ERROR\)
This equation summarily represents the enterprise of data analysis.
\(DATA\): A set of observations representing the thing we are trying to predict
\(MODEL\): A set of rules or formulas that make a specific prediction for each observation in DATA
\(ERROR\): The amount by which MODEL mispredicts DATA
Top 5 songs and top5 books of all time?
This is obviously a tough one but I'll try to adhere to the top5 format.
SONGS
1. Beethoven - Piano Sonata No. 14 (Moonlight Sonata)
2. The Beatles - Strawberry Fields Forever
3. Dave Brubeck/Paul Desmond - Take Five
4. Paul McCartney - Band On The Run
5. The Flaming Lips - Do You Realize??
BOOKS
1. Richard Dawkins - The Selfish Gene
2. Malcolm Gladwell - Blink
3. Jack London - The Call of the Wild
4. JRR Tolkien - The Hobbit
5. Judd, McClelland & Ryan - Data Analysis: A Model Comparison Approach (my introduction to \(DATA = MODEL + ERROR\))
wondering if there was any person (prof usually but could be, i dunno... TA, colleague, etc.) that had a particular influence on you which made you start focusing on stats, methods and data analysis more? i dunno, someone who once told you that you're talented at it or a particular course that blew your mind and you just had to say "woha! this is *so* me!"
Chick Judd, who, among other things (see book #5 above!), first planted in my head the idea of writing a stats/methods paper, ultimately expanding my view of what kinds of research I can do and leading me to finally see myself as, at least in part, a "methods person."
What statistical methods would you like to know more?
1. Time series analysis
2. Bayesian statistics
3. Not sure if this counts as a statistical method or methods per se, but I would like to delve deeper into the areas of measurement theory. I have some basic facility with item response theory but would one day like to be an expert in this area. I also am fascinated by what little I know (and it is just a little) about conjoint measurement theory.
What is your favorite probability distribution?
Uniform.
Just kidding... how boring would that be?
Probably the binomial distribution, simply because it is one of the only distributions that I actually feel like I really understand (e.g., how it is derived).
Are there any cookies you don't like? If so explain why they are horrible (and go into as much detail as possible)
Snickerdoodle... I won't go into the detail that you requested as I am afraid it will upset me too much.
I am on board with Bayes, although I have primarily only a high-level, conceptual understanding of Bayesian statistics. There are some who hold up Bayesian methods as the only sensible way of analyzing data, the solution to all of science's problems, etc., and while I am rather skeptical of the ideal portrait that I think some would like to paint of Bayes, I do want to learn more about the Bayesian statistical approach and am open to the prospect that this might really be a more sensible way to do things in a lot of cases.
With that said, what area of statistics would you say you're most weak at? What would you like to learn more about (if not merely methodological, so as not to repeat Dason's Q here)?
I like to think that I have at least a basic, passing familiarity with most of the major statistical methods that someone in my position is likely to come across, but there are a few areas that I am frighteningly ignorant about. For example, I have no idea what a "random forest" is, nor a "graphical model." However, I think the biggest and most conspicuous gap in my statistical knowledge is in the area of multivariate linear models, that is, linear models with multiple response variables (e.g., MANOVA, multivariate multiple regression). I simply have never spent any appreciable time learning about these methods and I don't have any good excuses for why not.
