Spelling Out Numbers vs. Writing Them Using Digits

Any thoughts on what the rule should be here? My sense is that numbers written using digits are much easier to quickly absorb, so I tend to write them that way whenever they refer to something that people might want to use in calculations or comparisons. I’d say, for instance, that “These books tend to sell for 20% below their list price of $8 to $10,” rather than “These books tend to sell for twenty percent below their list price of eight to ten dollars.” But when counting people or things in contexts where the count likely doesn’t need to be grasped as a number suitable for calculation or comparison, I spell out the number, for instance in “There are eight reasons why this law is a bad idea.” And I of course also spell out numbers when there’s no conventional numerical equivalent, for instance in phrases such as “millions of people.” The line is not always clear, and there are probably situations where I’d be on the fence about this; but that doesn’t trouble me much, just as it doesn’t trouble me that I’d sometimes use one synonym and others time another, based just on rules of thumb about what tends to sound better to me.

But I’ve found that many journal editors suggest that many of these numbers be spelled out, for instance in the “twenty percent below their list price to eight to ten dollars” context. My sense is that some common rules are to spell out numbers that are under 100, or spell out numbers that are unhyphenated single words, or some such, though I can’t vouch for what the most popular standard is at various journals.

I wonder what people think about this: Is there some clear merit in spelling out all small numbers, including percentages and prices and other things that one might want to calculate or compute with? Is it especially necessary to have a sharp rule that distinguishes numbers by their size rather than by their likely use (and, if so, shouldn’t the sharp rule at least sharply exclude things like prices and percentages)? I’d love to hear what people think.

By the way, I should note that none of this strikes me as a matter of “grammar” or “correctness” under any linguistic standard; the question is what looks better, and therefore what authors should do, and what publications should prefer.