Most of this post (except my contributions in italics, is all from a single answer in a Quora thread which resembles the kind of answer I often write at Politics.Stackexchange, but it is so good, and most of it is straight borrowed from other sources like Pew, so I don't feel too bad about stealing; I did upvote the author of the answer and in defense of fair use am providing some significant academic and political commentary of my own). The question, which is also riffing off a Pew article, is:
In US why do only 6% of scientists identify as conservative?
For what it is worth, I know and correspond with several professional scientists who are conservative (Razib Khan, Lubos Motl, and Greg Cochran come immediately to mind, and have represented a conservative gay academic librarian whose work is intimately related to the real world scientific process, and I have some fairly conservative (for Oberlin College) friends who were STEM students at the same time that I was a math major who took lots of physics classes, although not quiet enough for a double major or a fourth minor).
In academic campus life, scientists tend be less political and less liberal than their peers in the humanities and social scientists, but more liberal than their peers in the business school and the engineering school, with lawyers typically having a bimodal distribution born out by campaign finance data. Most scientific disciplines are viewed as being intrinsically apolitical, but apolitical and culture free aren't the same thing, as a recent post at 4 Gravitons (by a scientist who is mostly quite politically moderate but probably leans slightly left).
The most recent data I have on this precise question of the ideological identification of scientists is from 2009, unfortunately.
For me, the starting point is this:
Scientists are much less likely to believe in God. All else being equal, this will make them more likely to be Democrats.
Party affiliation and partisanship tend to be mutually reinforcing in an age of ideological polarization. If you find a Republican who is for gun control, chances are that they will start moving closer and closer to their party’s ideological consensus over time. Scientists, for instance, overwhelmingly accept evolution and global warming, which are still a matter of debate for the American public.
I particular like this development of a causation hypothesis, indeed a quite plausible one, which most people who study political identity aren't willing to even attempt. And, from personal experience I am inclined to say that causation runs mostly from a scientific worldview to a less religious or at least less fundamentalist religious identity, rather than the other way around. This direction of causation is one reason that the long term trend is towards more secular societies in the developed world, although it is more complex than that.
It is a more nuanced hypothesis than the more common but crude view that political views favored by some high IQ occupation should be considered more valid because high IQ people are more likely to have correct understandings of things, which is true to a point, but underestimates the influences of socio-economic conditions, race, and personal self-interest, in my humble opinion (something that I addressed a bit in a Politics.SE post about the downsides of meritocracy is a system of social and political organization.)
Since skepticism of these concepts is more likely to be a feature of conservatism, people like evangelical protestants who tend to reject them are a going to be viewed as an out-group by most scientists.
Even if you disagree that this process occurs, people who are scientists in the US tend to have left-of-center positions on all manner of political questions:
Scientists are less likely to be skeptical of government, more likely to be skeptical of business, more likely to support the welfare state, more likely to support the adoption of further measures for social justice, and less likely to be pro-military.
I would note that essentially all scientists have college educations and by definition they are professionals, mostly working on salaries, and that most scientists who identify as such work for governments or nonprofits on a salaried basis, even though some do work in the for profit sector where professionals with a similar skill set who identify primarily as inventors or engineers, even if they end up doing some science, are more common. This occupational setting could influence their views as well.
Now, there are some instances where scientists are more likely to be in agreement with conservatives, but not enough of those to alter the overall liberal balance:
I don’t have any data on opposition to animal research, but I would guess that to be a liberal position. The left also tends to be more opposed to nuclear power, which scientists overwhelmingly support.
As for vaccines, opposition to vaccine mandates tends to be higher in conservatives, in spite of of the popular perception to the contrary. I surmise that this is more due to libertarian ideology than to a rejection of the science.
Interesting nuances, I wonder how much the form of the question influences the partisan breakdown, however.
Still, overall, in the American political context, scientists espouse liberal positions. And this is reflected in how scientists self-identify:
Notably, the premise of the question (only 6% of scientists identify as conservatives) differs from the 9% figure in the data in this 2009 study, although given the likely margin of error in the survey data and the shift of attitudes over the last 11 years, which college educated professionals tending to the left, the 6% figure may be accurate now.