Oh, I'm aware of your expertise!--and I have a copy of Nonsense on Stilts.

However, even if a field as a whole can be conducted as a science in principle, the culture of a field can determine to what extent we actually hold practitioners to standards that make "science" a useful term to use to distinguish what they do from what people do all the time.

If a field is difficult, that of course slows progress. Systems neuroscience is about as difficult as it gets--ridiculously complex system, inadequate tools for the scope of the question, aspiration to bridge many levels of organization. But the culture is, overwhelmingly (save for some of the EEG and fMRI folks and the occasional computational-theory-of-everything crackpot), one of coming up with good evidential support in those few areas where the problem is surmountable and only holding fairly firm beliefs about those while maintaining an attitude of uncertainty about the areas where the difficulties are too great.

However, when the culture of a field faces difficulty by pretending it doesn't really matter, it changes the situation from slow progress to a stochastic walk through opinion-space. And I think in the case of psychology (though it's getting better) and sociology (not so sure), the biggest problem is that the culture admits far too much not-really-science so that the bulk of the effort goes into things that really don't much help our beliefs correspond to reality. It's not that the directions of belief-update are chosen completely without regard for evidence, most of the time, but that the quality control over direction is poor, and the scale of the update does not comport to the degree of certainty.