8 分钟，手机上做完Eight minutes, on a phone
不用报名，不用先加微信。做完立刻出一张等级卡，可以截图留底。 No sign-up, no contact needed first. You get a level card at the end — screenshot it and keep it.
一 · 听音选字1 · Hear it, pick it
听一个词，在四个字里选一个。选完自动跳下一题。 六个字频段各抽十题，答得好往上走，某一段错太多就停。 Hear a word, choose one of four characters; it advances automatically. Ten items from each of six frequency bands, climbing while the child does well and stopping when they don't.
二 · 读一篇短文2 · Read a short text
按上一部分测出的级别，给一篇难度相称的短文，四道题。 答对不到两道，级别往下调一级——认得字不等于读得懂。 A passage matched to the level just measured, with four questions. Fewer than two correct drops the placement a level — recognising characters isn't reading.
三 · 听两段话3 · Listen to two clips
只记分，不影响级别。用来看出「听得懂但不认字」的孩子—— 海外家庭里这样的孩子占大多数。 Scored but not used for placement. It identifies the child who understands everything and reads almost nothing — which describes most kids in overseas families.
够用来定级，不够用来比高低Good enough to place a child. Not good enough to rank one.
测评给的是一个区间，不是一个准数。六十道题估一个上千字的识字量， 单次测量的误差在 ±100 以上；两次相减，误差约 ±140。 The test gives a range, not a precise figure. Estimating a vocabulary of a thousand-plus characters from sixty items carries an error of more than ±100 on a single sitting — and about ±140 when you subtract two sittings.
所以：定级够用（差一百字不影响他该读哪一级）， 比高低不够用（测出 1050 的孩子和测出 1180 的孩子，很可能一样）。 我们不会把两个孩子的数字并排放，也请你不要。 So: fine for placement — a hundred characters either way doesn't change which level a child belongs in. Useless for comparison — a child scoring 1050 and one scoring 1180 are probably the same. We never put two children's numbers side by side, and we'd ask you not to either.
这是我们自己算出来的，写在这里不是客套。一个肯把自己工具的误差告诉你的机构， 才值得你把孩子交给他。 This is our own arithmetic, stated plainly. An outfit willing to tell you the error bars on its own instrument is one worth trusting with your child.
四件小事Four small things
让孩子自己做Let the child do it
家长在旁边提示，测出来的就不是他的水平，而且看不出来是提示过的。 If you prompt, the result isn't your child's — and nothing in the data shows that you did.
不用复习No preparation
测的是他现在认得多少，不是临时能记住多少。 It measures what they know now, not what they can cram.
要有声音Sound on
第一部分要听词发音。请把手机声音打开，或者戴上耳机。 The first section plays audio. Turn the volume up or use headphones.
我们会看到什么What we see
孩子的名字、年龄、城市、年级、家庭语言，你留的微信号，以及逐题作答。 只用来给你出结果和定级，不做别的。 Your child's name, age, city, school year, home language, the WeChat ID you leave, and the item-by-item answers. Used only to produce the result and place them — nothing else.