The threat is not from strangers, but from fully vetted and trusted community members having some kind of psychological event that leads them to bad actions.
The various entities working at McMurdo do pretty thorough pre-deployment psych evals to try to catch the kinds of people who would break vital equipment (think furnaces, or water-treatment equipment or high-cost/one-of-a-kind research equipment) or equipment necessary to fix vital equipment. But there's really no guarantee until you're out there that any given person is actually trustworthy. To say nothing of emergent psychological events that would drive a person to do such a thing. And anyway, the kind of person who wants to go to Antarctica frequently doesn't conform to our expectations of what a well-adjusted personality looks like. When your threat model is "trustworthy community members who occasionally and unpredictably become threats", you lock doors.
I've been down with the USAP a couple times, most recently about a decade ago, and in that span the psych eval was required for winterover deployments but not summers. It was very questionable.
I like to think that a spoof "How to Pass the Psych Eval" video we made one winter might've somehow contributed to that particular system being torn down a few years later, though I don't have firsthand experience with the system that replaced it. My understanding is that the new one involves more "team building" type stuff before deployment, which sounds like a step in the right direction.
One issue is that by the time a person has gotten to that late stage in the hiring process, they are often deeply invested in deploying - having packed things in to storage, quit "real world" job, etc. and their nominal manager understands that. Another issue is that people behave differently in different environments, and Antarctica is quite unlike Denver.
My information is about 10 years old, so that makes sense. It also makes sense that such evals are of limited utility, as the people I know who worked down on the ice were ... difficult personalities to get along with.
The various entities working at McMurdo do pretty thorough pre-deployment psych evals to try to catch the kinds of people who would break vital equipment (think furnaces, or water-treatment equipment or high-cost/one-of-a-kind research equipment) or equipment necessary to fix vital equipment. But there's really no guarantee until you're out there that any given person is actually trustworthy. To say nothing of emergent psychological events that would drive a person to do such a thing. And anyway, the kind of person who wants to go to Antarctica frequently doesn't conform to our expectations of what a well-adjusted personality looks like. When your threat model is "trustworthy community members who occasionally and unpredictably become threats", you lock doors.