Why isn’t this dynamic number formatting approach used more often for regression outputs?

External question — Cross Validated Stack Exchange Author: CrisisStudent Original post: https://stats.stackexchange.com/questions/659425 License: CC BY-SA 4.0 — https://creativecommons.org/licenses/by-sa/4.0/ Adaptation: HTML converted to plain text; contact email addresses removed. I’ve come up with a dynamic approach to formatting numbers that I find much clearer than the traditional fixed-decimal formats, especially when displaying regression outputs. Here’s the system I use: For values above 100: I use 0 decimal places (e.g., 256.435 → 256). For values between 10 and 100: I use 1 decimal place (e.g. 56.421 → 56.4). For values between 1 and 10: I use 2 decimal places (e.g., 5.232 → 5.23). For values below 1: I use 2 significant digits (e.g., 0.03423 → 0.034). This approach balances readability and precision, especially when regression coefficients or p-values vary widely in magnitude. It makes large numbers easy to interpret while maintaining meaningful precision for smaller values. However, I don’t see this kind of dynamic formatting used in regression outputs, where it seems fixed-decimal formats (e.g., always showing 2 or 3 decimal places) are the standard. My questions are: Why isn’t this dynamic formatting approach more widely adopted for regression outputs? Is it due to software defaults, tradition, or complexity? What is the preferred approach for formatting numbers in regression results, and what are the key trade-offs when deciding between fixed decimals and dynamic formatting? I’d appreciate any insights into why fixed decimal formatting is dominant and whether this dynamic system could be more useful in practical settings. Thanks! P.S.: Here is example (ignore the color coding, that is a separate adjustment) P.P.S.: The closest question I found is this , but it is somewhat separate.

External answer — Cross Validated Stack Exchange Author: Michael Lew Original post: https://stats.stackexchange.com/a/659426 License: CC BY-SA 4.0 — https://creativecommons.org/licenses/by-sa/4.0/ Adaptation: HTML converted to plain text; contact email addresses removed. I think that there are likely two things in play. The first is that relatively few people seem to be aware that one might choose to round to significant figures rather than decimal places. Certainly my students (science, biomedicine) would only slowly realise what I was on about when I told them of empty precision and significant figures. That problem is probably exacerbated by the rarity of seeing (and doing) such rounding. The second issue is that every programming language that I've worked with has built in functions for rounding to decimal places but not for rounding to significant figures. Programmers are lazy just as often as other people, and just as often they do not know about rounding to significant figures.