This list is abstracted from "Why People Believe Weird Things", written. by Michael Shermer. Reading this book is highly recommended. Particularly, chapter 3 of Shermer is recommended here. It provides an excellent outline of fallacies in thinking.
1. Anecdotal evidence is not useful.
All an anecdote tells you is what happened in one case. It tells you nothing about the general population and you cannot draw any general conclusions from it. You must have well-designed and controlled experiments to get enough data to reach real conclusions.
2. You need more than scientific language.
Words and phrases must have precise operational definitions. All hypotheses must be testable.
3. Bold claims need evidence.
Extravagant claims require a lot of evidence. The boldness of a claim does not make it true. A far-out claim will not be accepted until it has been successfully tested many times. The bulk of evidence must support it.
4. Radical (heretical) claims can be wrong.
Surely the Wright brothers got laughs concerning their attempt to fly. Alfred Wegener was scorned when he proposed that Earth's continents actually move around. These ideas survive because they stood the test. The Wright brothers' airplane actually flew, and a mass of evidence has shown that Wegener was right. But - there is a large number of other radical claims that did not withstand the tests and have been forgotten.
5. Where is the burden of proof?
Who must prove what? The person making an extravagant claim must prove, via experiments and evidence, that the new claim is actually more valid than current ideas. The new hypotheses must make better predictions and successfully explain more phenomena better than current theory. The current experts are not obligated to prove that their idea is better.
6. Rumors are not necessarily real.
You have almost certainly heard some wonderful story and later wondered if could really be true. Large numbers of such stories fall into the category of "urban legends," meaning that they never really happened. It is wise to take these stories as amusing fiction until you can find some confirmation of them.
7. Unexplained does NOT mean not explainable.
The fact that you have never seen or cannot explain some phenomenon does NOT> mean that it must be some unexplained supernatural thing. It would be quite arrogant to assume that you know everything.
8. Watch for rationalization of failures.
Pseudoscience cannot tolerate failures; they will be rationalized or explained away in some manner. True science must accept negative results as part of the search for the truth.
9. Look out for "post hoc, ergo propter hoc" reasoning.
If event B follows event A, that does NOT prove that A caused B. Event B could follow A purely by chance. You must have well-designed and controlled experiments to show that B always follows A. A single occurrence is not sufficient.
10. Beware of coincidence.
Truly random behavior can produce some interesting coincidences. Causal relationships do not always exist. Some interesting combination of events may be nothing more than chance. The fact that you have never heard of it before may mean simply that the probability of it is very low and you don't expect to see it often.
11. Check the misses as well as the hits.
Is the thing you are looking at really representative of its population? If one prediction of a "psychic" appears to be correct, how many others were not correct? We tend to remember the hits and forget the misses.
None of this is intended to say that there are no problems in real scientific thinking.
1. Theory influences observations.
What you see is often influenced by what you expect to see. Observations will be interpreted according to current knowledge, which can obscure important implications of the observations.
2. Observations change the observed quantity.
The classic example of this is found in the measurement of the motion or position of a sub-atomic particle. The process of measuring perturbs the particle.
3. Instrumentation influences results.
The basic idea here is this: that which your instruments cannot detect does not exist. Spectacular advances in knowledge often occur when detection capabilities improve so that previously unseen things or phenomena can be seen.