For the previous results I used Google with the site: operator to restrict the results to particular countries based on the server's domain. Using their web API I was able to target these results more specifically, using an option which determines site locations based on the "top level domain name of the server and the geographic location of the server IP address". This would now correctly identify those numerous sites in the .com space but which reside in other countries, presumably giving more accurate results for my survey. In addition, I could now include US sites, which are impossible to distinguish based on domain name alone.
| UK | Australia | Canada | US | |
| aeroplane / airplane | 1:1.6 | 1:1.4 | 1:18 | 1:20 |
| aluminium / aluminum | 2:1 | 5.4:1 | 1:2 | 1:11 |
| burgled / burglarized | 14:1 | 5:1 | 1:1.5 | 1:2 |
| car boot / car trunk | 100:1 | 5:1 | 1:2 | 1:4 |
| car park / parking lot | 14:1 | 14:1 | 1:21 | 1:8 |
| colour / color | 5:1 | 2.5:1 | 1:1.5 | 1:16 |
| estate agent / realtor | 13:1 | 40:1 | 1:1.2 | 1:2 |
| father christmas / santa claus | 1:1.1 | 1:6.5 | 1:13 | 1:12 |
| forthcoming / upcoming | 3:1 | 1:1.3 | 1:7 | 1:7 |
| frying pan / skillet | 2:1 | 4:1 | 1:2 | 1:1.7 |
| fortnight / two weeks* | 1:1.2 | 1:4 | 1:11 | 1:25 |
| flatmate / roommate | 9:1 | 3:1 | 1:34 | 1:44 |
| gaol / jail | 1:5.7 | 1:5.3 | 1:49 | 1:78 |
| infant school / kindergarten | 1.7:1 | 1:51 | 1:173 | 1:300 |
| liquorice / licorice | 2.8:1 | 1:1.6 | 1:9 | 1:9 |
| maths / math | 2.5:1 | 1:1 | 1:30 | 1:71 |
| mobile phone / cell phone | 2:1 | 7:1 | 1:2.2 | 1:2.8 |
| nappy / diaper | 3:1 | 3:1 | 1:8.4 | 1:7.8 |
| pavement / sidewalk | 9:1 | 4:1 | 1:1.3 | 1:1 |
| spelt / spelled | 1.2:1 | 1.3:1 | 1:6 | 1:7 |
| theatre / theater | 7:1 | 6.6:1 | 4:1 | 1:1.7 |
| tv aerial / tv antenna | 7:1 | 1:4 | 1:6 | 1:16 |
| write to us at / write us at | 54:1 | 12:1 | 1.3:1 | 1.6:1 |
* fortnight / two weeks isn't a strict pairing, because two weeks can often be used in UK english where fortnight is not applicable. However, it's still useful as a comparison between countries.
The Australian and UK results (the languages with which I am more familiar) look to be more accurate than the previous table.
I wanted to put up an online tool to enable anyone to determine the ratios for any words, but I don't have the web server facilities at the moment.
I didn't know that cell phone was so popular in the UK, though I have been noticing its increasing use on the BBC's international services.
Note that I had to use "write us as" rather than "write us" so as to exclude some other common usages such as "write us a letter". For example, "write me a letter soon" is normal UK english, but "write me soon" is not.
Updated 20 June: Fixed tv aerial / tv antenna results. Added write to us at / write us at, mobile phone / cell phone pairs.
Updated 30 June: Added aeroplane / airplane, father christmas / santa claus, gaol / jail, infant school / kindergarten, liquorice / licorice, maths / math pairs.
Updated 10 July: Added many more pairs, and tried (unsuccessfully) to make the table layout more readable. I should really colour-code each entry, with shades indicating how great the ratio is.