Nobody is claiming that scientists are infallible, and yes, of course they can be individually biased, and thus their conclusions can be wrong. But the thing to remember is that Science is not an ideology, nor a collection of facts. Science is a method of inquiry.
And as such, the Scientific Method is the most effective and powerful concept we humans have come up with so far. I'm not sure if everyone reading this tread are fully aware of exactly how the Scientific Method works, so here is a short rundown, just so we can all agree on what we are talking about:
1. Ask a question: That is to say, find a natural phenomenon and ask yourself, why does this happen and how.
2. Do some research: This includes studying the avaliable data already collected, as well as the collection of new data. Data comes in many forms, depending on the subject being studied and the discipline of science. For instance, the Theory of Evolution is supported by evidence from genetics, anatomy, ecology, animal behavior, paleontology, and others.
3. Construct a hypothesis: Here it might be useful to define exactly what a hypothesis is. Put plainly a hypothesis is a proposed explanation for an observable phenomenon. A simple hypothesis could state something like: If I do A, then B should happen. (This is a gross simplification thought. Proper scientific hypothesis often include vast amouts of conditions and potential results, among other things).
It is important to note that the hypothesis must be both falsifiable (that is, it should make predictions that, if shown to be wrong, will invalidate the hypothesis) as well as testable (that is, one should through experimentation or observation be able to determine if the predictions made by the hypothesis are correct or not). A hypothesis is still a far cry away from what we consider a Scientific Theory.
4. Test the hypothesis: Again, how a hypothesis is tested can vary a lot depending on the phenomenon being studied and the discipline of science, but it usually boils down to either performing experiments and observing the results, or by making independent observations of the phenomenon itself.
5. Analyze the results and draw a conclusion: If the hypothesis was constructed properly one should now, through experiment and observation be able to determine whether the hypothesis was right or wrong. If the hypothesis was right, that's great, but if the hypothesis was wrong, then it's back to the drawing board. And even if the conclusion was that the hypothesis was correct the scientists might want to go back to test it again in a different way. More information is always better than less.
The conclusion is valuable in either case. Edison supposedly said that he had not so much invented the lightbulb, as he had invented 10.000 ways of -not- making a lightbulb. This gives some idea of how scientists work; through careful and tedious testing and re-testing until they are sure their results are correct. The individual scientists could still be wrong though, which leads us to the next step...
6. Report the results: There are Scientific Journals covering every discipline of science in which scientists continually publish reports of their findings. These journals are usually not intended for the layman (although plenty of laymen read them too) but rather so that other scientists working within the same discipline can go over the data and look for faults and mistakes, as well as recreating the experiments to see if they get the same results. This is commonly called "Peer Review". Repeatable experiments and observations is essential to the scientific process, because as has been mentioned, individual witness reports and personal annecdotes do -not- count as evidence.
This last step is put in place to avoid biased results and to ensure that the hypotheses that eventually become Theories are supported by an overwhelming majority of the scientific community. This will to a large degree eliminate the problem of "humans making the conclusions" and replacing it with "lots of experts in that particular field of science agreeing on a conclusion".
Science is not dogmatic and scientists don't present their results as absolute everlasting truths. They are well aware that future technology and developements in science can, in time, produce evidence that show their results to be inaccurate, or even wrong. But the point is that the Scientific Method produces practical results, results that we see around us every day, and it is, thus far, the best and most efficient way we have of investigating nature and the world around us.