Mostly, that is. While it’s a bold statement that goes against the industry standard, I think quite a few people will agree.
1. First and foremost, the level of quality a translator demonstrates in a sample translation doesn’t necessarily represent their typical level of commitment to quality. Even an average translator can find motivation to output a perfect sample translation once in a while by investing a lot of time into proper research and using help from more experienced colleagues. But this doesn’t mean, of course, that this is the level of quality a client will consistently get, say, over a period of six months.
2. Because the level of quality a translator delivers isn’t constant, evaluating an individual on a basis of just one test doesn’t make sense. Consider this example: recently, we received a translation from a freelance translator with a huge amount of spelling and grammar errors. This wasn’t typical of her at all since all her previous translations were virtually flawless in this regard. When we sent her the error report, she explained that she had made those errors because she had caught cold and hadn’t been feeling well during the translation. If the quality of translation depends on things like translator’s mood, time pressure, or distractions, can you really evaluate performance using just one sample translation that is one page long?
3. This is not to say that all translation tests are meaningless. A certain percentage of sample translations might serve a purpose, especially in specific projects where an end client is selecting among several translators the one they like best. But I just don’t understand why sample translation should be an industry standard for evaluating translators because it’s often isn’t meaningful enough to justify endless hours candidates spend on tests.
For me, the two solutions are to optimize the way I search for the best translators and to ask for existing samples rather than tests.
1. Optimizing the search for translators is important because by selecting your candidates more carefully, you eliminate those candidates who have a little chance of passing the test. This can dramatically reduce the amount of tests you distribute. One reason to take this approach is that there are not that many good translators. When a client tests a group of candidates, most of them are usually doomed from the outset. Better selection of candidates reduces the amount of unnecessary tests, thereby saving time downstream for both the candidates and the client who has to review all those tests.
2. Asking a translator to send an existing sample is my favorite approach. I believe it’s much more effective and fair. When a translation agency tests candidates using a sample translation, it’s not uncommon for them to reject about 95% of candidates. While this is fair because an agency doesn’t force anyone to take those sample translations, I don’t think that this is a sustainable way of testing translation talent. It means that 95% of people will end up spending hours of their time in vain. If an agency tells me upfront that I just have a 5% chance of passing their test, I’d rather not take it. When a client asks for existing samples instead, everyone benefits. A candidate doesn’t need to spend time on sample translation. And an existing sample is just as good for evaluation purposes as a sample translation.