Haven’t seen many well-designed ones in my experience. 
You’re absolutely right, though – I didn’t mean to infer that it’s ONLY a Chinese test problem. But there are more resources for ESL tests and so forth, and just the nature of things dictates that the more resources, the better your chance of getting someone who knows what they’re doing, as those people usually command a fairly good salary, after all.
It’s tough writing monolingual tests for a second language. If you have the luxury to write the test partly in English, for example, then you can truly test reading or listening. The student reads the passage in Chinese, then answers English questions on the content. You know if the student fails to pick the right answer (assuming your choices are solidly written) the problem most likely is that he or she didn’t understand the Chinese passage. At least you have eliminated the idea that he DID understand the passage but did not understand the question or one or more of the choices.
Even “zao juzi” is testing multiple things. What if the target word is used correctly but something else in the sentence is wrong? Does the teacher take off points for characters that are written wrong? How “wrong” is “wrong” when it comes to writing a character? When I taught 2nd year Chinese at Texas, I used to put a section on the tests where the students could write any way they could make me understand what words they were using. Characters, Pinyin, whatever. I was looking ONLY at whether the sentence made sense. In one way that made sense as I couldn’t do individual oral testing because of the number of students in the class, so this was the nearest thing.
Of course language is such a complex thing that it’s almost impossible to reduce it to a neat formula that will yield a number. Unfortunately, that’s what many programs want or need (for grades, rankings, etc.) I much prefer the testing method that goes with TPRS (the method I teach) – the tests and quizzes are intended only for teacher feedback. If the teacher is doing his or her job, you get at least 80% of students scoring 80% or higher. Else, you need to go back and find out what you didn’t present enough or well enough. The idea that someone can’t learn a language is astounding to me when our brains are wired to do so. Some students may have problems with written forms, but spoken forms are within the reach of nearly everyone. I had special-needs kids and a Deaf student in my classes in the US and they did just fine.
Personally I only test vocab, as grammar will be acquired in a certain order anyway, no matter what you teach in which order. If you try to pound late-acquired stuff into a learner’s head at an early stage, you get overanalysis. IMHO better to be patient and focus on things that let the learner get out and about and use the language, which will give him more input and more motivation to keep on. (Okay, I digress…) 
