If an idea lines up with the ultimate conclusion a man wants to reach, he will believe it even if there is little real evidence that it is right.
The other day, a friend and I were going mountain biking at a trail that we knew probably had rattlesnakes. The best time for us to fit it into our schedules was at about 4pm. We reasoned that the rattlesnakes would probably go back into their holes around this time, so it would be a safe time for us to go. Did we have any real evidence to lead us to believe that rattlesnakes really do go back into their holes at 4? No. Did we want this to be the case? yes.
One issue where I think that this principle can be seen over and over again is with regard to global warming. I don't know for sure if global warming really is a problem (I think that it probably is), but on both sides people grab onto any data that seems to support their conclusions, even if it is barely valid and not conclusive. The argument has devolved from considerations of facts leading to conclusions to a battle between people with entrenched conclusions selecting data to try to support themselves.