Wish there were some solid science on this related to MJ. All I can go by is my own experiences, with both crosses that I have control over, and commercial seeds I've grown.
I think there's likely more factors but the one's we typically talk about in growing MJ are genetics and environment. I have no doubt that environment can certainly play a part in producing herms, just like it can play a part in expression of the same strain/clone when grown under different conditions. This I've seen enough to believe it's a factor. The question is what type(s) of stress would tend to produce herms. I've stressed the hell out of plants, in some cases on purpose, sometimes from being lazy or not having enough time to take care of things properly. e.g. I leave any potential cross phenos, male or female, in small containers to the point of almost getting root bound to see how they behave. Males will tend to start flowering regardless of light schedule if left in too small a container, I've repeated this many times and I use it to cull males that do this (auto-flower).
At the same time, I've yet to see a herm in my crosses (myself), but the odd one has been reported. One recent one was reported by
@SSGrower on an F1 (BR x PL) first gen cross. I gave him a heads up that this was an F1 cross and I hadn't tested it yet. I try not to send out F1's unless I've run many with no issues, this is why I don't normally send them out. He had some other strains running at the same time and I believe he had issues with most if not all in that run. In that case I'd be looking for environment/process stressors, having one or two herm is one thing, having multiple strains I'd be leaning towards other factors than genetics.
On the genetics angle. The reason I got into my own crosses was that I kept seeing so much variation in commercial seeds. I figured if I'm going to pheno hunt through a pack of seeds and get multiple phenos, they may as well be my own and free(sh). Although I've yet to see a herm myself in commercial drops I've certainly seen many reported by others. When it's the odd one and the environmental factors are in question I can see some being caused by the environment. But it seems to me that too many breeders are putting out polyhybrid F1/F2/S1's which were never stabilized and bound to be more susceptible to herms. If I were to consider producing an S1 for instance, I'd wait until I'd crossed and backcrossed that line 3-4 times (e.g. minimum F2 back-crossed to an F1 male, with no herms showing). But every time I see a breeder producing too many new crosses/strains in the same year I cringed at the potential for unstable genetics. Some even advertise F1/F2's/S1's and charge full pop for what is essentially a chuck, worse still, people buy them(?). This is why I also think that genetics, particularly over the last few years with so many poly crosses, and little or no stabilization, also have a part in producing herms.