9. When data is too limited, no conclusion can be drawn.

The companion to "I can be wrong" is "I don't know." Speak confidently when you have a strong argument backed by verified data. Otherwisesay "I don't know."

What qualifies as too limited? The answer is a resounding "it depends." The importance of the conclusion, the chances of acquiring other data prior to any time a decision has to be made, and even the presence of others more qualified can affect when you decide to withhold judgment.

For example, let's return to the medium, who after explaining their ability to see ghosts, asks you if you believe that ghosts exist. Let's assume you knew very little about ghosts other than what you've seen in the movies. So the data you have to work with is: a few works of fiction, plus someone's unverifiable statement that they can see ghosts. Neither is really data, so your data amounts to exactly nothing. This certainly qualifies as too limited.

You might assume that this means that your answer should be, "I have no data to prove they exist, so no, obviously they don't exist." But if you made this argument, you would be falling into a common reasoning mistake, typically identified by absence of evidence is not evidence of absence. This catchy little phrase tells us that without data, we cannot make any judgment at all. Just because we cannot prove something is there does not mean that it is not.

So, as a critical thinker, your answer will be something along the lines of "I really don't know enough about it." They may attempt to provide other data, such as tales of their past encounters with spirits (unverifiable, just like the spider story), or may even take you to a house they consider to be haunted so that you can observe some of the strange goings-on directly. Perhaps strange noises, unidentifiable lights, or even objects moving of their own accord.

Is this enough data? Not really. The existence of ghosts is a pretty huge claim. Another catchy phase to remember: extraordinary claims require extraordinary evidence. This means that strange noises, although notable, are not a particularly overwhelming piece of evidence, in the face of such an enormous assertion as the existence of ghosts. No matter what your beliefs, it's likely that a major change in your worldview will be necessary to accommodate such a conclusion. If you believe in an afterlife, then are ghosts spirits that didn't go to that afterlife? If so that raises the difficult question of why they didn't qualify, and how you can shape your life so as to avoid being trapped on the earth as a ghost, stuck making noises and nudging objects around pointlessly for all eternity. If you don't believe in an afterlife, than that calls that whole belief into question. If you think ghosts are not spirits of the deceased, then there's a whole other giant bag of questions to be asked.

The point is, some overpowering evidence would need to be presented for the medium to make you think seriously about the existence of ghosts. Evidence of some activity not created by a human or a machine, or credible scientists who create devices capable of measuring (and eventually, explaining) the energy or matter which composes the spirits are two possibilities. Supporting a claim as ambitious as the existence of ghosts would require a great number of very solid pieces of evidence.

We should also clarify that the terms "limited" and "enough" are not satisfied by sheer quantity. Ten mediums telling you they can see ghosts is really no better than one. Neither is a hundred or a thousand. Each of these individual pieces of information is unverifiable, and so all must be rejected. To be sure, a thousand people who claim they can see ghosts seems like something more worthy of investigation than just one. But just like a hunch, this only makes it more worthy of investigation. A large quantity of subjective data is still of no direct use for creating an objective conclusion. Having "enough data" means enough real data: data that has passed our careful scrutiny outlined earlier. When taking the census at your space colony, you only count people that live there, not how many people knocked on the door but were not admitted last month.

One important qualifier to this tenet is that, in the real world, it's often necessary to make a decision without enough data. Unfortunate though it is, not acting is often worse than making a bad decision. For example, when in a strange city and faced with a number of restaurants, you may have no good method at all for choosing one. The smells are all strange to you, the menus are in a language you can't understand, and you know nothing of the local culture. It's fair to say that you don't have the necessary data to make a good conclusion; but suspending judgment means you'll go hungry. Since the consequences of a bad choice are unlikely to be dire (probably just the wasted time and cost of the meal) you can simply go with the one that has the nicest looking exterior.

In order to justify making a conclusion or decision without sufficient data, you must determine that the effect of not choosing will be worse than the aggregate of the available options. Otherwise, say "I don't know."