Orlando's moves this offseason have been both praised, and highly criticized.  That's what happens when you're active in the offseason after making the NBA finals.  Some will feel your team that made the finals should be brought back the same, and some will feel that you should upgrade, given the fact that you didn't win.

The biggest move of the offseason has so far been trading for Vince Carter.  It's been shown here through numerous posts that Vince is a clear upgrade on offense.  He bests the player he's essentially replacing (Turkoglu) in almost every statistical category.  However, it has also been mentioned that he is not a very good defensive player.  First, let's compare Vince and Hedo in terms of some composite defensive statistics for the 2008-09 campaign:  (All data taken from basketball-reference.com)

As you can see, Hedo clearly played better on defense last year than Vince, as measured by defensive rating, defensive win shares, net defensive plus minus, and opponent PER.  However, Hedo played on an elite defensive team, while Vince played on a young team with plenty of defensive deficiencies.  Many people have made the argument that given the chance to play on an elite defensive team with the reigning DPOY he will improve significantly.

I thought it might be enlightening to look at other players who have been traded in similar circumstances, and how moving from a weak defensive team to an elite defensive team has impacted their own defense.  Note, this is by no means a comprehensive analysis, I am only going to look at how this has impacted a few players to get an idea of what we might expect from Vince next year.

The players I'm going to look at are Brent Barry, Paul Pierce, Mario Elie, Tracy McGrady, and even Hedo Turkoglu.  Here are their compiled statistics:

As should be apparent, each of these players went from mediocre or even bad defensive teams to very good defensive teams.  In the case of Paul Pierce, it wasn't so much that he was traded as the rest of the team was, and the end result was the same.  Some of these players were not known as good defensive players, yet managed to perform well when placed on an elite defensive squad.  Even Hedo Turkoglu, the player we are comparing Vince Carter to, performed somewhat better when going from the Kings (who were actually fairly good defensively) to the Spurs (who were excellent).  His career in Orlando illustrates this better; as the Magic improved on defensive, so too did Turk.

Given the above data, I think it is by no means unreasonable to expect Vince to improve significantly defensively, and replicate the level of defense from the starters he is replacing.  He has one of the best coaches in the league, an elite shot blocking big man behind him to protect the basket, and the motivation to buy into committing to being a good defensive player.  While I don't think that these composite defensive statistics tell the whole story, I think they give a good understanding of the subjective notion that players can improve when moving from weak defensive teams to elite defensive teams. 

If someone wants to do a more thorough analysis, tracking player moves and changes in defensive production, they would be welcome.  It is certainly possible that my results are influenced by selection bias, and I'm sure not every player has improved when moving to a better defensive team so there must be some counter-examples as well.

Update:  I updated the post to reflect additional statistical measures that were requested, namely net defensive plus minus, and opponent PER.  Regarding the validity of these statistical measures, all of them have some shortcomings.  As I mention in the comments, net defensive plus minus (and plus minus stats in general) exhibit a high degree of variance, which makes it difficult to rely on.  On the other hand defensive rating is so tied to team performance that it also exhibits a degree of bias.  Tied into all of this is the notion that defense is generally a team effort, so any aggregate measure will have some error from individual performance.  Personally, I prefer a consistently biased measure to a measure with a high degree of variance because it contains an error that can be understood if not always accounted for.  However, it always helps to have all the data available for a more complete picture.  For more discussion on the bias-variance problem in estimation theory, see: http://www.aiaccess.net/English/Glossaries/GlosMod/e_gm_bias_variance.htm

A note of caution regarding opponent PER for Hedo Turkoglu's 2002-03 season: The data was not directly available to I calculated it myself manually, but I was in a bit of a hurry so I can't guarantee that the result was correct.  If someone want's to double-check me that would be great.