When it comes to science, perhaps the most dangerous form of bias is simple wishful thinking. Oftentimes the data are extremely hard to come by in the first place, requiring huge investments of time, money, and hardware. After all that effort you probably really, really hope that your hypothesis works out. If it does, you get to publish, speak at conferences, and have a shot of getting your grant renewed. If it doesn't, well, maybe you can come up with a new idea?
So I'm sympathetic to scientists who unconsciously try to make their results look as good as possible. They're working in a broken system, with the number of their papers and the size of their grants determining their fate. But that subtle incentive introduces an ugly bias into the body of scientific research: apparent results that simply aren't there.
It's the job of the rest of the scientific community to a) be aware of that bias, and b) fight against it. You may have heard that some fields of science are facing a "replication crisis", where the majority of results are not holding up to further scrutiny. This is good! It's a sign that the field is self-policing and actively working against the implicit biases that lead to junk research. Imagine a world where we didn't even know these fields were facing a crisis, and the results were simply taken at face value. Is that world better or worse than our current situation?
For example, the CalTech astronomer Mike Brown and his colleagues argue that there is evidence for an additional massive planet in the outermost solar system, but many other astronomers argue against that hypothesis and present their own lines of evidence to counter. And he doesn't blame them - he even admitted during a recent talk that had he seen a competitor publish those same results, he would've scoffed and worked to disprove them.
Fighting bias in science isn't some lofty high-minded virtue. It's a result of natural human grubbiness and our desire to prove our opponents wrong and to demonstrate our own correctness. Progress in science relies on a community of healthy skepticism and open debate - those are the core principles that drive advances in understanding.