I agree with Marcus

Kyoto/Melbourne ngrams
迷惑	281001
ご迷惑	58981

I misunderstood Johan's number to come from the 
Kyoto/Melbourne N-gram Corpus, so I assumed this form 
was more common than it actually is. Still, 59k in the 
k/m ngrams is nothing to scoff at.
I don't really think there's an issue with including ご/
お compounds merely based on them being very common? 
"... then there are probably dozens of words we could 
add." 
Let's add them, then.

For 〜的 and ～性 compounds, I agree but the unwritten policy on ご〜/お〜 entries that we've been following up till now seems to work quite well.

Typically, we don't accept ご〜/お〜 entries (that aren't listed in dictionaries) unless they satisfy at least one of the following two conditions.

1. The word is more commonly used with ご/お than without (e.g. お湯).
2. The ご/お version has a different or more restricted meaning (e.g. お会計).

If a ご/お compound merely has to be common to warrant inclusion then there are probably dozens of words we could add.

OK. I'll retag them.

I agree with Marcus that
ご～, お～, ～的, ～中, ～性 etc. compounds
that are obvious and not in the refs
should still be included if they are very common.