Good science

Five diagnostic features of good science are: repeatability; economy (simple and aesthetically pleasing); mensuration (proper measurement); heuristics (stimulates further discovery); and consilience (connected and proved consistent with other good science). The work of real science is hard and often for long intervals frustrating. If you choose an academic career you will need 40 hours a week to perform teaching and administrative duties, another 20 hours to conduct respectable research and still another 20 hours to accomplish really important research. Advice to the novice scientist: There is no fixed way to make and establish a scientific discovery. Throw everything you can at the subject, so long as the procedures can be duplicated by others. Consider repeated observations of a physical event under varying circumstances, experiments in different modes and styles, correlation of supposed causes and effects, statistical analyses to reject null hypotheses (those deliberately raised to threaten the conclusion), logical arguments, attention to detail and consistency with results published by others.

Edward O. Wilson
American Scientist 86, Jan./Feb. 1998