RE: How to do bad statistics, and how NOT to do it, Part II (DaVinci)
There are lots of visuals to relate here. I can see you put a lot of effort in practice, and I appreciate that.
I missed the first post of the series. However, I glance through to get more info.
Your view about the questionnaire isn't a bad one. The questionnaire isn't perfect, and that is why we always try to improve it whenever we can. As you have said, most of those questions are centered around quality, but I think some of those questions are different.
Don't forget, suggestions for improvement are always welcome!
There are a few analysis/visuals I don't understand in the post. I think it would have been clearer, if you have added more texts to give a more detailed explanation.
With the information in the post, I think this post is also about the translation post by silviu95 and the review given to it. The mistake does not look nice, and such shouldn't persist.
As you have said, 'we all are doing a good job''. Thank you!
Please note that while the CM hasn't changed the footer, I am not scoring #iamutopian posts based on the questionnaire. They have their own metric, and that will be the case until we go live with the new guidelines and new questionnaire, which will be comprehensive enough to reflect these types of posts.
To view those questions and the relevant answers related to your post, click here.
Chat with us on Discord.
[utopian-moderator]
Concerning the questionnaire, I had a very radical solution - but it wasn't accepted.
The basic premise is that translation must be as good as possible.
Clients must get the top quality, no discussion about that.
As all the translations will be perfect when finished - all the translations are equal. Scores will be equal.
As the consequence... Moderators are forced to get the best translators, in order to facilitate their own work.
The level of quality is going up, there are no sparks between translators and moderators as they are collaborators and not rivals. There is no bad blood between the teams concerning the scores *(as all scores are equal). Of course, if translators are bad - they will be replaced.
In the current system, there are several things which are not very logical:
Or... We can see that all the parameters concerning quality are telling us that the translation is "so-so", but there is only one error. Ho-how... If it's bad - make it better. Imagine having 200-300 strings, 1 mistake and the quality is "meh... so-so". It's not very logical to me.
I also can't understand how some people are constantly making 10 mistakes per 1000 words. My answer is already 300 words long. Imagine 3-4 errors, made while writing in the native language. How is it possible - I don't understand.
Your idea is good, regarding the quality of translations. However, I believe that, within a group of professional, there are people that will be superior. Thus, there is no how the quality of work would be equal.
Some people have years of experience in translation, while some aren't. Yet, anyone can get better if the chance is given. As I have mentioned in my comment above, things are not really perfect now, but it would be better to look for a less strict way.
Also, anyone that provides way less quality could be replaced since we want good translations. We shouldn't nurture poor works.
Thank you for your review, @tykee! Keep up the good work!