We want to support the argument that the unique USMC Recruit Training--what is special about this one specific program--creates this special bond among Marines.

One way to say it is that (B) cites something specifically true about the USMC Recruit Training itself, while (D) cites something that seems more generally true. We absolutely know there are lots of conditions in the real world in which a specific group of people might be relatively isolated from others: students at a relatively isolated rural college, mountaineers going up one of the Himalayas, sailors on a sea-faring vessel, etc. If that's a general cause of bonding psychologically, then it would provide an alternative explanation and be a weakener. We don't know what it would mean in general for a group if it experienced relative isolation from others, so we don't know whether (D) would be a valid weakener, but being a possible weakener certainly doesn't help its candidacy as a strengthener!

Also, think about it. If the recruits are relatively isolated from outsiders, do they bond right away with each other, or do they remain lonely and awkward for a long time and start to bond after the training is over? We don't know. We don't know whether the content of the training program itself causes the bonding, or whether simply isolating them would work regardless of the content. (D) allows for a lot of wiggle room. Maybe the content of the training program itself causes the bond, maybe---in other words, maybe (D) is a strengthener. Maybe.

By contrast, (B) is tight. Recruits are more independent at the beginning of the training; later on, they are not acting separately, independently, so they are acting in connection with others. The decrease in independent behavior is an increase in cohesive behavior. If this is a consistent change during the training program, then it strongly suggests that the program systematically creates cohesive behavior. That's a clear strengthener.

"maybe a strengthener" is not a strengthener. A strengthener must be crystal clear and unambiguous. Similarly if the question were asking for a weakener or inference or any other kind of CR question. Maybe doesn't cut it. Maybe is a lily-livered failure of a choice.

Also, there's the tricky issue of outside knowledge. It's absolutely true that if, say, you had read an article or book about the USMC and had specialized knowledge about them from that reading, or if you knew personally someone who was in the USMC, then that would give you outside knowledge that shouldn't be used in answering the CR question. That kind of specialized outside knowledge should not be used and is not relevant. But general common sense outside knowledge (e.g. "are groups of people ever in relative isolation?") can and should be considered.

Answer: B