The problem is that 99% of your average person's interaction with "science" is by way of newspaper articles claiming "scientists say red meat is bad" and then three years later "scientists say red meat may be good". Another three years go by, and they see "scientists say red meat is bad for you after all". In reality, some scientists demonstrated an interesting correlation with a single marker (and often later studies are talking about correlations to something else entirely) that may or may not actually be causative, and either the media is spinning it into "a story" or some public health bureau is playing a numbers game that leads to the behavior "if it has any effect at all we should recommend it".
But, what the public sees is: "scientists say X" followed by "scientists say Y, previous scientists were somehow wrong", followed by "scientists claim that second scientists were wrong and they were right the whole time". We are actively training the public so that when we hear "scientist say X" the response is "engh: they're probably wrong". Hell, this article and my reading it is an interesting example: it's chock full of references to published journal papers and actual psychologists, but I honestly don't even care anymore as every time I've gone and pulled the research articles like this are based on I've come up with tenuous conclusions based on flawed experiments that are being trumpeted as the new truth.
Worse, much of the "science" being cited isn't really science at all: a cohort study, for example, has no controlled experiment and can't really be used to show anything definitive; it is, at best, a way to come up with interesting hypotheses on which to later run actual experiments. However, a good number of the "scientists say X" that people hear are from health research, where controlled experiments in humans are sufficiently rare and tied up in ethical conundrums that many of the statements being made (and almost all of the ones involving diets) come from new cohort studies or meta-analyses of previous cohort studies. It is no wonder, then, that every three years we hear something different. ;P
Now, with that in mind, imagine the average person hearing "scientists claim evolution is true". They often respond with things like "it's just a theory: one of many". I will contend that they do that because to them "scientists claim evolution is true" only has as much evidence behind it as "scientists claim red meat is bad for you" or "scientists now claim salt may not be bad for you after all" (http://news.ycombinator.com/item?id=4069613) and the result is now that they don't trust us: we can tell them anything we want, but they have no reason to believe us anymore, as we were continually wrong to them every time it mattered to them before.
I will then claim that if we want to be credible, we need to stop reporting every paper that comes out to the public as "scientists say X" and even further be careful about what we call "science" in the first place. It should be that when we claim that something is true, it is because we have sufficiently strong theoretical models and experimental evidence that can really claim "this is knowledge that may be imprecise, but is not inaccurate" in the way that newtonian mechanics is a little imprecise (it doesn't take into account relativistic or quantum effects), but isn't so wildly inaccurate that we should ever expect to see a newspaper headline "scientists report we were wrong all along: objects of greater mass actually repel each other".