Restricting the effect to universities prevents me from drawing on a key factor I had in mind, external pressures that shape scientific research, since in popular opinion support for science competes with religion on average. More popular support for and deference to science could have significant positive effects (depending on your model).

Still, the Dark Side needs to be sustained, and when its emotional underpinnings are removed a superstructure can collapse suddenly. Moreover, the negative affect for their previous faith-based religions would tend to make ex-believers more susceptible to the idea of reliance on logic and evidence.