Small sample size. Looking through the boxscores of your first 6 games, where Raines had 1 SBA, and McGee had 3 SBA, Raines isn't getting many opportunities to steal. In those first 6 games, despite a .419 OBP, he had only 2 opportunities to steal. He was caught in his only attempt, against an A+ catcher in one of them; In the other, he drew a leadoff walk against a C- catcher, so I'm not sure why he didn't go there. However, those were his only 2 times on base where there wasn't someone in front of him or he was already on 3B. I didn't look at the games after those, but it looks like he had a rough stretch over his next 25 games as his seasonal OBP dropped from .400 to .354 and he had just 5 attempts in that span. Then, in his last 19 games he had 12 attempts as he started getting on base more again and his OBP went back up to a .385 on the season.
Likewise, for McGee, he started hot, had a number of SBA in the first few games, then cooled down quickly and though he wasn't on base much. In the 6 games I checked for him, seemed to have an open base virtually every time he reached. Since you've posted this, Raines has more SBA than McGee (10-8).
170-175 PA isn't a big sample size... alot can change from that point on. Trends can be noticed like you did, and when you look to the underlying cause (open base in front) you can start to assess whether it's more valuable having Raines stealing or moving the 7-9 hitters around the bases (as it looked like he was doing early on), and then make any adjustments if you feel necessary from there.