Timeline for Transforming variables for multiple regression in R
Current License: CC BY-SA 3.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 9, 2013 at 22:45 | vote | accept | zgall1 | ||
| Jun 9, 2013 at 22:45 | |||||
| Jun 8, 2013 at 19:42 | comment | added | zgall1 | Career WAR is the sum of offensive and defensive contribution for every play the player is involved in over the course of his career. The details can be found here - fangraphs.com/library/misc/war | |
| Jun 8, 2013 at 19:13 | comment | added | Affine | Might be helpful to know how career WAR is calculated then (aka understand the data generating process). | |
| Jun 8, 2013 at 16:22 | history | edited | Nick Cox | CC BY-SA 3.0 |
deleted 249 characters in body
|
| Jun 8, 2013 at 15:44 | comment | added | zgall1 | This is career WAR. It is Alex Rodriguez. | |
| Jun 8, 2013 at 15:35 | comment | added | Eric Peterson | how did someone get a WAR of near 100? To the best of my knowledge, there has never been a WAR (for an entire season) even approaching 20. Mike Trout was around 10 last year, I believe, and that was the highest in the modern era, with the exception maybe of 1 or 2 of Bonds' best seasons. | |
| Jun 8, 2013 at 15:28 | history | edited | Nick Cox | CC BY-SA 3.0 |
added 738 characters in body
|
| Jun 8, 2013 at 14:41 | comment | added | zgall1 | You are correct that WAR is my dependent variable. I will look into a GLM with log link. Thanks for the advice. | |
| Jun 8, 2013 at 14:39 | comment | added | Nick Cox | It seems that WAR is your dependent variable. You provide evidence for my assertion, disputed elsewhere on this site, that the two terms are often confused. My advice is not to recode negative values to zeros (maltreats the data) but to use a GLM with log link. Please assume zero interest in, or knowledge of, baseball minutiae on my side. | |
| Jun 8, 2013 at 14:09 | comment | added | zgall1 | With regards to the negative independent variable (WAR) issue, for reasons that are a bit complex, it is reasonable to re-code those as zeros if that makes the transformation process easier. Within the framework of this dataset, this is a justifiable procedure. If you would like me to go into more detail (warning - baseball jargon required), I am happy to do so. | |
| Jun 8, 2013 at 14:08 | comment | added | zgall1 | As you might be able to tell from the scatterplot posted above, I am using a baseball statistics data set. The independent variable, WAR, is essentially a cumulative measure of the value contributed by a player over their career at the major league level. The independent variables, AdjSLG, SOPct and BBPct are minor league statistics that are commonly thought to predict success at the major league level. The Age variable is the age at which the player produced the minor league statistics. The dummy variables are used to indicate the minor league level at which the statistics were produced. | |
| Jun 8, 2013 at 13:52 | history | answered | Nick Cox | CC BY-SA 3.0 |