Having just sat through a Sunday school lesson where science was the bad guy because scientist "know everything" and "change their minds about things all the time," this particular discussion comes at a perfect time to annoy me. I'm going to frame my discussion in terms of models, where models are the mental images we have of the world around us. In science it is understood, first of all, that ultimately our explanations are just models, and second of all, that our models are in many ways incomplete and wrong. We expand and correct our models using hypotheses, which are tested by experiment, observation, and measurement. It is also understood that there exist multiple working hypotheses--multiple, equally probable explanations that we should take into account and, in a well-conceived experiment, distinguish between. Scientific models, because of their connection to experiments and observations, must be directly tied to the physical world around us (except for physicists--they're special. J/K, Clark). Does that connection make us conceited about our knowledge? Maybe. I'm sure it looks that way from the outside. We really do have a lot of confidence in a lot of our models, including evolution, and that probably comes off as conceit. Evolution has withstood an awful lot of testing, so we're really rather confident it's correct. As confident as we are in many models, every scientist acknowledges that there are likely to be refinements, or even occasionally replacements of widely believed models when a new theory comes along that has better explanatory power. I'd say evolution is unlikely to be overturned completely by a new theory simply because it does do such a good job explaining what we see in the natural world. That said, if sufficiently convincing evidence were to come forward that something else were behind the diversity of life we observe, scientists would eventually accept the new model.
Religion is also a model, but unlike a scientific model, it need not be verifiable by independent experiments. In fact, the nature of faith precludes the possibility of independent verification of most beliefs. Whatever evidence underlies your faith (and physical evidence is certainly a valid part of faith for many!) your model is not one that can be verified or invalidated by someone else. Just isn't possible. Religious belief is still a model, though, and there are certainly aspects of your model, my model, everyone's models, that are incomplete or incorrect. Our culture seeps into our religious models, staining those models with baggage that has no eternal significance, and yet seems terribly important to us standing at this time and in this place. Because we are not privy to the thoughts of Deity, we have no concrete, independently verifiable way of removing those unnecessary bits for ourselves and for everyone else. It's also more or less impossible to get everyone to agree on a religious model--even within a religion! If you doubt that, just try getting a bunch of people to agree on what the true nature of God is.
Unfortunately, a lot of non-scientists seem to think it's appropriate to apply a religious model to a scientific model. It's not. In fact, it's incredibly conceited to think that it's appropriate. Because scientists are aware that our models are incomplete and incorrect, we are willing to modify, and even throw out models when they fail to accurately reflect reality as measured by experiment and observation. I find such mental flexibility much more often lacking in those of a more religious persuasion, particularly when it comes to a religious model.