Sheila Wilson wrote:
Tony M wrote:
What rather horrifies me is that, on a typical job of assessing 20 translators, there was NOT A SINGLE ONE whom I considered to be absolutely faultless, and only 3 who were "passable". And of the 20, 11 who were complete non-starters! I find this utterly horrific, and although to some extent it shakes my faith in test translations, it also to some extent highlights why agencies consider them necessary.
I evaluate translation tests for TWB, just the "general" ones in the French to English pair. I've checked about 300 and I rarely come across one that's both accurate and pleasant to read. Many applicants are clearly translating in the wrong direction as their texts sound more French than my own French writing (which I accept sometimes sounds very English in its structure
). And I sometimes get to see the medical or engineering ones, some of which really are laughable: raw Google Translate! I know that TWB accepts only around 50% of applicants. I'm actually surprised it's that high.
I've assessed something like twenty translators in Korean-English so far, and it's clear that the vast majority are native speakers of Korean who are in far over their heads. Two of the others have seemed to be native English speakers who lacked an adequate understanding of the source and wrote nonsense. Only one has actually written grammatically - not even especially well, just grammatically - and shown a clear understanding of the (not very complex) source. I can hardly blame agencies for wanting test translations at this point...
As for accepting test translations, I would say that the most critical thing is that the agency agrees to your terms. If the agency also has a satisfactory Blue Board rating and the translation isn't more than 200-300 words at the absolute maximum, you will do better in the long run by accepting, even if a few turn out to be a waste of time, rather than rejecting them out of hand. That said, I recently had a $4000 project assigned to me and paid promptly by a direct client who found me on ProZ, had no Blue Board record, and wanted about 450 words as a test. I wasn't very busy at the time, so I did the whole test anyway, and they liked it. If I'd been very rigid about test translation standards in that case, I would have missed out on that $4000. For that reason, I think you have to be a bit flexible at times and be willing to take a potential waste of a small amount of your time in return for a potential new, profitable client.