Mankiw's post gives the best example (emphasis mine):
I am pretty sure Paul would not find this line of argument persuasive. As far as I can tell from reading his commentary over the years, he does not believe that the distortionary effects of taxes are particularly large and so they do not figure much into his policy analysis. But many other economists (and I suspect many stimulus-skeptics like the tea-partiers) believe that taxes have significant incentive effects and can prevent the economy from reaching its full potential. Their argument seems logically coherent, even if it relies on a different set of parameter values for the relevant elasticities than Paul believes to be true.
Beliefs, beliefs, beliefs. Damn the evidence - or even better, argue that the evidence is wrong, mismeasured, or inaccurate.