Michael J. Disney in American Scientist:
It is true that the modern study of cosmology has taken a turn for the better, if only because astronomers can now build relevant instruments rather than waiting for serendipitous evidence to turn up. On the other hand, to explain some surprising observations, theoreticians have had to create heroic and yet insubstantial notions such as “dark matter” and “dark energy,” which supposedly overwhelm, by a hundred to one, the stuff of the universe we can directly detect. Outsiders are bound to ask whether they should be more impressed by the new observations or more dismayed by the theoretical jinnis that have been conjured up to account for them.
My limited aim here is to discuss this dilemma by looking at the development of cosmology over the past century and to compare the growing number of independent relevant observations with the number of (also growing) separate hypotheses or “free parameters” that have had to be introduced to explain them. Without having to understand the complex astrophysics, one can still ask, at an epistemological level, whether the number of relevant independent measurements has overtaken and comfortably surpassed the number of free parameters needed to fit them—as one would expect of a maturing science. This approach should be appealing to nonspecialists, who otherwise would have little option but to believe experts who may be far too committed to supply objective advice. What one finds, in my view, is that modern cosmology has at best very flimsy observational support.
More here.