An emphasis on personal experience rather than trust in others

Nullius in verba The 1660 motto of the Royal Society: “take nobody’s word for it”.

Something I saw with my own eyes, or an experiment I performed myself.

These are classic QS experiments. Do it yourself, with your own code and data, open-sourced as much as possible.

When writing about health, avoid subjects over which you have no direct experience.

I can write about my own weight, for example, or my own experience with exercise, but avoid speculating about diets or techniques that I haven’t tried myself.

It’s okay to discuss anecdotes from other people, of course, so long as you have direct experience

If you saw or talked with somebody, go ahead and pass along that information, but as much as possible try to add your own signal, some commentary or summary that shows you are doing more than repeating hearsay.

When using an external source, make it your own

Books and other publications can be an invaluable sources of knowledge and inspiration, but they can equally carry disinformation and falsehoods if used uncritically. Static words imperfectly capture reality, so don’t over-rely on them, especially if you don’t have direct access to the authors.

As much as possible, don’t simply quote another author. Summarize, expand, reform in your own words and add your own take in order to make it your own. Stand on the shoulders of giants; don’t simply look up to them.

If you read the results of an experiment in a peer-reviewed journal article, for example, assume that the data is incomplete, perhaps inconsistent, in the same way you know your own data and conclusions are tentative.

Avoid retweets from people you don’t know

Don’t ascribe undo authority to a newspaper or magazine, or even something in a peer-reviewed journal or book. You can summarize these sources, but avoid attributing more importance than they are worth.

It’s okay to pass along information, but be clear that you’re merely repeating something, and as much as possible try to add additional facts or commentary.

Here’s an example: A researcher whose work you have not personally examined says that mice treated with X turn out differently than mice treated with Y.

If you’re in a hurry to simply pass along the researcher’s claim, if you think the subject is noteworthy enough to deserve additional attention, then go ahead and retweet — so long as you make it clear that you don’t endorse the statement. In practice, this retweeting should be rare and mostly limited to situations that require immediate attention.

One rule of thumb: your posts should be timeless. Ask if you’re okay having this comment retweeted in ten years. Will future readers, with additional information, forgive any limitations in your analysis? Or will they dismiss your comment as uninformed?

If it’s my own data, based on my own personal experience, I can’t be wrong. For example, I can say that my weight at the beginning of a diet was X and at the end it was Y. That’s a fact that won’t change, even if later new evidence demonstrates that the diet has no effect. Even if I later discover that my methods were incorrect – say, my scale was off – the data point collected at the time still stands.

Humble Curiosity

The personal scientist never really knows for sure. This sense of humility may be unfamiliar to people in the modern era, but in fact much of earlier scientific progress depended on it.

Historian Warren Sussman refers to a shift from a Culture of Character to a Culture of Personality¹. As economic prosperity encouraged anonymous exchange, success came to be determined by attributes that could be immediately sized up. A first impression matters more if you are less likely to cross paths again. The whole point of, say, franchises is to project an image of familiarity in an unfamiliar location. McDonalds lives and dies, not by the personal relationship you have with a particular store manager, but by a well-articulated but anonymous “image” carefully crafted to make you trust their product. In a complicated world, we instinctively seek consistency. In a smaller world, where you know everyone, there is room for nuance because you have the luxury of getting to know each person, quirks and all. Important attributes like dependability, predictability, steadfastness can’t be measured in an anonymous, one-time-only interaction, so we seek proxies, the most obvious of which is self-confidence. If an anonymous peddler doesn’t himself appear confident, how can you be confident in his product, especially if you are forced to choose among many similar ones?

Our current culture puts a such a premium on the appearance of confidence that maybe we forgot how much of this is a show. How often do people really, truly know something to be true, and how much of it is bluster born of a cultural requirement to rise above the noise.

The people behind https://lesswrong.org perhaps know this lesson best. Instead of striving to be right, try to be less wrong.

There is some precedence in the general philosophy of science literature for this approach. As @nntaleb observes, Karl Popper doesn’t really “accept” any hypothesis, but rather merely fails to reject at some statistical significance.

But the flip side of – what shall we call the opposite of confidence? humility? – must not be a retreat into the familiar. A personal scientist is distinguished by a constant, intense fascination with the unknown, and especially with paradoxes that go against familiar wisdom.