ORS, if we're faulting cases, let's look at yours:
1) How relevant are any of Lee's stats over three years past in evaluating his current value?
2) How relevant are home-road splits including those years from the distant past, especially when they're years in a very different ballpark?
3) But, since you bring up career trends, is 2005 really a representative year in Lee's career? BR credits Lee with 6.3 batting wins in 2005. His second-best year is 2007, the year you discount as an anomaly, where he had 2.7 batting wins. OK, his 2002 tied his 2007, and his 2003 was 2.6 batting wins, but in 1,413 games--nine 157-game seasons--he's accumulated only 18.3 batting wins. That's 2.0 batting wins per full season--and that's even lower than Lee's 2007, despite completely forgiving time lost due to injury and bench time as a young player. It looks to me as if 2007 is slightly generous, not misrepresentative in any other way.
4) And, before you look up Youk's batting wins, they're lower but they're not normed interleague...Youk's EqA normed for all time was .297 in 2007, while Lee's was .305, and that doesn't consider Youk's defense nor Lee's unusual home-road split (it does adjust for normal park factors, but Lee draws a higher-than-normal advantage from Wrigley Field). In 2006 Youk was better than Lee. In 2007 it was very close.
In 2008, the topic of discussion, I still would want Youk over Lee. BP calls it close, but they favor you, ORS, as does every other poster.
I'll wait for October to tally stats. I'm backing Kevin Youkilis, even if I stand alone.