Thinking about it, in real life having different people from different backgrounds on a project has the potential to increase, instead, the research speed. With different backgrounds, they can look at a problem from many different vantage points and catch things the others wouldn't have thought of.
For example: the current description of plasma rifles. Sure, it seems like its all high-energy physics and magnetic fields and thermodynamic plasma fields and stuff. However, it'd likely be someone like a biotechnologist who discovers one of the key parts of the rifle: the spinners that make the plastic shells.
That's an argument against separate skills, but I do like the idea of separate scientist skills. I get the impression that the range would only be, say, 80% to 120% of average - so if you don't pay attention you won't notice a difference, but if you manage things right you'd get small bonuses to research production. In fact, if you only research one item at a time with all your scientists, there'd be essentially no difference at all. The only way to get this bonus would be to split projects up into teams.
Example:
Suppose there are two projects that need 800 research points to complete, but one is a biology project, the other is physics. You have four scientists: Two with 60bio/40phys, and two with 40bio/60phys.
In a four-man team, the scientists have a total of 200bio and 200physics, so researching each project would take 4 days. Researching both projects would take 8 days. This is the exact same situtation that we have now (each scientist gives an equal value).
In a two-man team the bioscientists would have a total of 120 bio points, and research the bio project in 6.7 days.
The same thing happens to your physicists: They can research 120 points together, and finish the physics project in 6.7 days.
In this case, the two teams will finish at the same time.
If the physics project was a vital armour that you need to get on your troops ASAP, you'd stick all four labcoats on it, so you can finish it in 4 days. However, the next project done will take another 4 days to complete.
If your goal is to research the entire tech tree as quickly as possible, then it's best to split your projects into smaller teams (even if you only have 4 guys available). You'd have to wait an extra couple days for your first tech to come out, but the second tech will be done a day ahead of of if you just tasked all your guys on them both.
This way, players are rewarded for thinking ahead and micromanaging, without forcing them to or penalizing them if they don't. In fact, it's a trade-off between getting a project done now, or getting more projects done sooner later on.