Thank you Fencer for the reply! That's a lot clearer. However, I have to say that based on this calculation, for people who are interested in pursuing a >80% research career it seems like the salary does not differ as much as the pp salaries. For example, based on a 80% research calculation and assuming a 180k salary for pediatrics, that comes out to be 182k...not that much different from the 206k for a rad onc!
This is absolutely correct, and probably one of the reasons why there are still 80/20 careers in peds, and almost nobody in rad onc runs a 80/20 group. However, in individual case scenarios, your milage varies greatly depending on the specialty, institution, location. Some institutions are very "centralized", meaning that you have to work for them in both research and clinical roles--this protects you from having to find a job, but makes it easy for them to exploit you. Others are more "entrepreneurial", meaning they pay you a specific (low) amount from the grants and you can do whatever else you want to (or not!) make up the rest.
The other thing to realize is that this is all administrative. In reality, the 80% time research is not really 80%, and the 20% clinical may not be 20%. For example, you may work a lot of hours at night writing papers and grants, and work at a clinic or a faculty practice on weekends/coverage/on-call duties. The institution might assign you "20%" clinical duty but the "20%" might consist of nothing more than an hour every other week of seeing 1 patient (literally, because of no-shows). Or it might assign you a real 20%, like 2 months of attending coverage a year, PLUS weekend call coverage in an attending call pool. Or perhaps the 2 months inpatient attending coverage is half way covered by residents and all you have to do is to show up for 2 hours in the morning to round with them. Or perhaps the 20% is doing 5 expensive and fancy procedures per week. Or perhaps if you have the chops they might use your 20% in a more administrative/managerial role, which means perhaps not "clinical" in a strict sense, but say teaching, or quality improvement or practice management/supervision. As you can see, the 20% can generate a vastly different amount of revenue, even within the same specialty (i.e. procedures vs. clinic). Your role (and, correspondingly, your salary) depends on how the institution sees where your value is, and whether you have sufficient leverage, skill and relationship to advocate for yourself.
While being completely tied to the institution and getting a fixed salary used to be the rule for physician scientists, the "entrepreneurial"/no$-guaranteed system has become more common in the age of soft money/consolidation, etc. A not-uncommon scenario is that the clinical and research divisions have a Chinese wall, and the research division pays you a small salary strictly out of grants (i.e. 90k for a K-awardee), then you have a menu of choices (i.e. inpatient vs. outpatient procedure vs. non-procedure) to pick from (or apply to) in the clinical division, and make different salaries depending on what you decide (and what they offer). Whenever your grant runs out, you can work more within the clinical division to make up the salary differential (if that's what you want), or perhaps even work outside of the institution (for example, at an ancillary clinic, or even a different hospital, as long as there's no non-compete). As you can see, these two parts will pull you, because they have differing priorities--the clinical division will want to squeeze the most out of your supposed "20%" while the research will make sure you apply to more and more grants. The ability to manage and thrive in this tension is probably *the* most important skill for physician scientists.
This is all very complex and full of negotiations on an individual basis to be worked out when you are ready, and you should not base your career choice based on something as uncertain as your salary/work hours 10 years from now.