This issue came up a bit at Portrait of the scientist as a young woman a few weeks ago. I am not really one to socialize with my group and never have been. Research is hard enough without personal issues getting in the way. I am not really all that social to begin with, and I like to keep the personal separate from the professional. I think it can be hard to maintain an appropriate degree of professionalism when personal relationships get involved (this is human nature). It gets worse when this leads to "Golden Child Syndrome", where the more social/better connected group members get more resources or professional opportunities than everyone else.
Although I never socialized much with my own research group, I can see the temptation for a new Assistant Professor. This has played out in our department a few times. A new professor is probably in a new city where they may not know anyone. They spend a lot of time with their research group, and may be fairly close in age if they got a TT position after a two year postdoc (2-3 year postdocs are the norm in my field, but I was older after my time at National Lab). As students and postdocs, our life experience is that we find friends in the groups we spend a lot of time with. It is a new experience to be suddenly in a position where making friends this way also involves a power differential. In my experience, this seems to be much more of an issue with new Assistant Professors, and
seems to fade as the age gap between professor and student increases
and the new professor makes local friends outside their research group. I've also seen it lead to powerful resentments between the first cohort of students, who helped set up the lab and were friends with their advisor, and the next cohort, who came into a working lab and a situation where the PI was not actively searching for friends. Even so, some of my colleagues remain pretty social with their research groups.
My own students don't seem to socialize with each other as a group (I can't be certain, because I don't discuss personal lives with my students unless they bring it up). I wonder if my anti-social nature has caused this, but I don't have a problem with it. Some students seem to prefer a more social group (I say "seem", because I haven't discussed this with any students, and the power differential makes it unlikely to ever have such a conversation), but others don't, so recruitment-wise, I think it all balances out in the end.
One of the reasons I am kind of happy to not have a very social group is that I am not sure that the decision to attend group social events is ever really truly voluntary. In particular, someone from a different culture may not experience this as a choice at all. I think it is particularly problematic when the invitations come from/are issued on behalf of the PI, which can make the social event feel like a group obligation. Even absent the PI's direct involvement, if someone regularly chooses not to attend group social events, it may have a negative impact on their working relationships with group members, since people naturally gravitate towards helping people they are close to at the expense of people they know less well.
Worse when group social events perpetuate inequalities or send the message that only certain types of people are welcome in the group. At conferences, I've been in a group of attendees who decided to continue discussions at Hooters (yes, really). I've seen people at meetings organizing mixed professional/social outings to strip clubs and other non-inclusive venues. Such outings would really be problematic in the context of research group social events. Even things like research group contests can be exclusionary. It is one thing to set up an NCAA basketball bracket pool, where anyone can fill out a bracket just for fun even if they don't follow college basketball (or want to bet for money). It is another to organize a research group fantasy sports league which requires a large time commitment to following a specific sport for a long time.
I am not sure students consider the social atmosphere of the group when selecting a group to join. I wasn't very social as a student, so a group that met up every weekend would not have been a good fit for me. As a student, I spent a lot of time in the lab, and enjoyed time away from my group, though we all had pretty good working relationships. I had some friends in my group, but we did not socialize as a research group much outside of work. As a PI, I am probably too much in the non-social direction. We don't do much more than an annual group lunch, but so far it has worked for me.