Let's talk about money, and how money relates to choice in research.
In order to execute a research project, you need to pay for:
- Your time
- Your staff's time
- Materials / subjects / equipment
If you are a professor at a US R1 university, you typically get a 9-month salary paid by the university due to your teaching. The summer salary should be bundled into your grants but isn't really strictly necessary: it's nice to have the money, though, and funders typically want you to pay yourself.
If an average professor makes about $100K, and you assume an overhead multiplier of 1.6, then that means you need between $0 and $40K per year to support yourself.
"Staff" typically means graduate student or postdoc, which, depending on your institution and circumstances will typically run you something between $30K and $100K for a graduate student or postdoc for a year. If you can use undergrads, then it's down at the lower end of that range or even lower. Of course, some graduate students can be supported by TAships, some undergrads will work for credits instead of money, and both grad students and postdocs sometimes come with their own fellowships, meaning that your staff may be free. So total of $0K to $100K per staff member.
In one of your comments, you state that a typically small study in your field will cost you $40-$80k to run: let's take the top of that range and consider $80K of materials / subjects / equipment cost for the year, and assume it requires one full-time person. Total project cost, then, is between $80K - $220K per year.
If you're doing the same project in industry, let's assume your salary will be doubled. But you no longer have the 75% teaching support and you don't have the option to not pay yourself. The overhead will typically be significantly higher as well: let's assume 2.5x, though the actual numbers in industry tend to vary quite a lot. Put those together and you need $200K * 2.5 = $500K/year just to support yourself, though we'll assume you need only half your time at $250K.
Likewise, your staff will be much more expensive too, since they're getting a similar salary to you. They'll also likely be much more efficient than a graduate student, though, so you can get the same efficacy with a smaller fraction of their time. Thus, if they're somewhat junior to you, the staff cost might only be about $200K/year. The materials / subjects / equipment cost stays the same, which leaves you with a project that costs $530K/year.
Look at those ratios: The same project costs 2.5x to 6.5x as much in industry! Government labs are similar, though they tend to pay their people a bit less and are correspondingly a bit cheaper.
These numbers are highly imperfect rough estimates, but they illustrate the major difference in cost scales between academia and industry. As a professor, every research action you take is heavily subsidized by the educational environment in which you are operating. This subsidy makes it far easier to explore novel high-risk ideas, just because it's way, way cheaper. On the flip side, you also pick up a lot of project risk because postdocs and graduate students are less well-known quantities than long-term research staff.
This tradeoff of risk vs. cost is a major component of the "research freedom" available at universities. Even if we hold all else equal, it's a lot easier to get hold of $100K for trying out an idea than it is to get hold of $500K.
