Pire ,

What I am trying to say is people (researchers ) may not be aware of the source of errors in their thinkings, or measurment. May be they are trying honestly. They get certain results based on their thinking  and their measurment style. But they may not be aware of the source of error, either in thinking process or in measurement process.

When they get couple of good  results , they send for publication. This is the procedure for publication. May be you get 2 good result in 10 experiments. You just hide the bad results and show the good results. This how the system works.