Firstly, if you're able to join a group openly, the problem of access can be fairly easily overcome.

Secondly, it also goes some way to resolving a major problem with this type of ethnographic research, namely that of how to
record data while observing people's behaviour. 

Finally, Overt Participant Observation makes it less likely (although it may never completely overcome the problem) the researcher will "go native" (that is, become so thoroughly-integrated into the group you're studying you cease to become an observer and simply become a participant...).