I am getting a little weary of hearing people talk of science as a religion - it is not a religion. Richard Dawkins addresses this notion as well as various religious thinkers (see this article as an example). Essentially, religious people see tenants of science as inexplicable and therefore akin to faith while scientists are consumed with evidence and therefore faith is irrelevant. In fact, Dawkins views faith as vice - in either science or religion.
For me, this question is simple because of perspective. Religious practitioners generally begin with the conclusion and then argue why this conclusion is true while scientists begin with a model (akin to the popular notion of theory) and then test it, which can lead to either validation or falsification. As evidence mounts, this model may gain more credibility and become a theory or even a law. This distinction (#1) alone convinces me.
Still, both do offer evidence to support their positions; this is to be expected and may confuse the bystander. These rules of evidence are both different and often incompatible (although I believe that any evidence that science amasses may be appropriated by a religious argument). Generally, because religion is a personal phenomenon, personal anecdote may be offered as proof for a claim while science demands evidence that is empirically verifiable and repeatable by a test or tests. Proof for a religious argument may be identified after the fact while evidence in a scientific context must be predicted before hand and demonstrated. This distinction (#2) alone convinces me.
Also, religion often uses a deductive logical structure which can be understood as a series of "if/then" statements. For instance, one may deductively reason that if God created a world and man is part of this world then God created man. This deductive reasoning was valuable to the origin of science (ie Aristotelian to Cartesian schools) but Bacon and others rejected this method in the 17th century for evidence-based inquiry. Inductive reasoning looks for evidence and, like a detective, the researcher follows the evidence to where it leads. Ironically, fictional detectives like Sherlock Holmes practice a kind of inductive reasoning although these observations are often referred to as "deductions." Essentially, religion tends to be deductive while science tends to be inductive. This distinction (#3) alone convinces me.
While these three distinctions are enough for me to decide that science is not a religion, I understand that they may not convince someone that has already decided the contrary. Feel free to leave me comments on this matter.