Falsification and Fabrication

Falsification of data and fabrication of data are both forbidden in science, but until today I didn't know that there was a difference between the two. Falsification, apparently, is tweaking results you have to make them resemble the results you want to have. This could be removing outliers or changing standard error values or something like that. Fabrication, on the other hand, is just making stuff up, producing data for experiments or studies that never happened.