True evidence-based proofs, the way its done with science, requires that you test one variable. In the example above you'd have to have a parallel economy where all else was the same except the tax rate. Unfortunately this is impossible, but this does not mean we get to change what "proof" means. Economics would work much better if everyone came to terms with the fact that it is untestable and was honest with themselves that all opinions regarding it are just beliefs, regardless which side you fall on.
Also, one could argue that a property tax would be better than an income tax as a way to enforce maintenance of the commons and to provide infrastructure and an educated work-force. (I tend to favor use of taxes as "investment", rather than "money for nothing")
Assertion: "high" (greater than 50%?) marginal tax rates make economic growth impossible.
Evidence: Top tax rates between 1940 and 1980 were between 70 and 90%. Growth was higher than it was after 1980.
Conclusion: the assertion about the harmful effects of a high top/marginal tax rate is false. (I think this process is called "induction").
Note that this does not prove that high tax rates are beneficial, only that they are not harmful.
It's certainly possible that the general slowdown since 1980 is due to other factors, such as growing scarcity of cheap energy and other materials. But investigating alternatives will screw with the livelihood of somebody in the top 0.01% income earners, so one might argue it's harmful to your health to push too hard for such :-)