In short, I probably wouldn't, at least not in isolation. I may have an inkling of how my unknown was made, or other spectral data. Also, in 1H NMR these are usually integrated, which may also give a clue.
I don't disbelieve you as such, but consider https://en.wikipedia.org/w/index.php?title=Special:CiteThisPage&page=P-Toluic_acid&id=738054995 - is this incorrectly stating the preferred IUPAC name then?
Although recommendations can change over time, see also https://chemistry.stackexchange.com/questions/75668/is-4-chlorotoluene-also-an-iupac-name-for-1-chloro-4-methylbenzene , https://www.quora.com/Is-4-methyl-chlorobenzene-an-accepted-IUPAC-name ...
The "order" of something is an approximation. If given correct data you calculated the time it takes Usain Bolt to run 100m to be 0.1 seconds, 1 second, 10 seconds, 100 seconds, or 1000 seconds which is most likely to be correct?
But if you're taking both figures as diameter, where has the 0.5...
Not necessarily a fallacy. You are asked for the "order" of the size, not the "exact" size. What this means that as long as you have approximately the right figure you're OK and is a useful check. You may have made an error mathematically or otherwise and calculated this as a kilometre, or a...
And 4 sides and 1 diagonal is not sufficient either. Consider a shape with sides 1, 1, 0.8, 0.8 in that order. Now the 0.8 v can be inside, or outside (convex or concave) with the area different.