I have trouble getting the idea of a "preliminary experiment" across to my students. They want their first data to be perfect and want to know what to expect before they even take that first measurement. There's also the issue of not knowing what variations to expect after you've only taken 1 (or a few) measurements. I wish all students had taken statistics before they enter a physics lab.
They also seem to expect more of physics than of other sciences. I recall in biology we had trouble even finding some given organ in a frog, fish, etc. And in chemistry we just wanted to see a reaction occur or a color change, etc. But in physics lab they don't just want to see two objects fall side by side they want g=9.80 not 10.
There is also the thought that if you just had better equipment your results would look better. And they would, but when you have better equipment you are also looking at things that are harder to measure. It's a sort of arms race, as your research gets more advanced your tools get better but what you are trying to do gets harder. As a result you mostly "run in place" on a "tredmill."