Pass mark for new JLPT(N1-N3)

[Sujeesh]

Check the below link for overall pass marks and sectional pass marks of the new Japanese-Language Proficiency Test (N1-N3)

http://www.jlpt.jp/e/topics/201008291283128850.html

[Amarok]

Wow!! This is more lenient than I ever dreamed it would be. Now I'm excited for the test in December because I actually feel like I have some glimmer of hope for passing.

[ichigo_venky]

The link does not seem to be working anymore >_<

[dimitri_can]

Below is the passing mark for N1-N3 paper (July intake).



Total marks for each section is 60 marks, which makes it 180 marks.

However, that does not mean that the December paper might have the same passing mark. This just serves as a gauge for you. I am checking with Japan Foundation on this.

Good luck for your JLPT! =)

If you have questions, please don't hesitate to look for me.

[julianjalapeno]

[Amarok]

Dimitri, you can't just say things like that with nothing to back it up. If you *think* that this is just a temporary scoring process but you can't back it up with actual facts/information from the JLPT people themselves, please don't state it like it's fact.


Oh, yeah, and I'll quote directly from the website, not wild conjecture:

Why has the scoring method of the new test changed?

It is to more accurately show the Japanese-language competency of examinees.
No matter how carefully test questions are designed, their difficulty changes each time and in comparison to previous tests. With the scoring method of the new test, the same competency results in the same score regardless of question difficulty level or test occasion. The same scale is always used to calculate scores for the same-level test. This scoring method can show you the degree of improvement in competency compared to previous test results. It also facilitates setting goals for the next test.

[NeWbY]

I do hope they change it (the overall pass marks and sectional pass marks), actually. Right now it's too easy to pass N1/N2 I think.

[dimitri_can]

[dimitri_can]

[Sujeesh]

More details are updated in this link

http://www.jlpt.jp/e/guideline/results.html

[Osaka]

Well, I'm glad the "every section must be over 70%" rumor is now dead.

[Norry]

So in order to pass N1, out of 180 points, you only need to score 100 total, with a bare minimum of 20 points per section (vocab, reading, listening)? That is so much better than the 70% rumor.

[Fugashi]

I've received my result and the scores are all less than I'd estimated. I notice the passing score is lower too i.e. 56% (=100/180) whereas it used to be 70%.

So I wonder if the new scoring system includes a mathematical adjustment to remove lucky guesses. The choose-one-from-four multiple choice system means you should get about one quarter of the answers you didn't know correct. So if you knew 40% of the answers you would in fact score around 55% (because when guessing the other 60% you should get about 15% right).

Anyway, if this is correct, then to compare marks on the old system with the new one you can do the following simple calculations:

Let P = Previous system test score (pre- 2010)
Let N = New system test score (2010 and later)

N = P - (100 - P) / 3
P = (3N + 100)/4

[dimitri_can]

Got the reply from Japan Foundation. Seems that the passing mark is really not going to be the same.　They didn't directly answer my question. LOL
======================
国際交流基金日本語試験センターです。

2010年第1回（7月）の日本語能力試験の結果（得点）について、2009年までの
試験とは得点の出し方が異なるため、たくさんの質問をいただきました。
みなさんの質問にまとめてお答えします。

*************************

新しい「日本語能力試験」（新試験）では、受験した人の日本語の能力を、
より正確に表すため、得点の出し方を変えました。新試験では、これまでの
ような「いくつ正解したか」に基づく「素点」ではなく、「いまのみなさん
一人一人の能力がどの位置にあるか」を表す「尺度得点」で得点を出します。

試験問題は、どんなに注意して作っても、試験の回によって難しさが少しずつ
変わります。このため、正解の数に基づいて計算する「素点」では、試験が
易しいときと難しいときで、同じ能力の人でも得点が違ってきます。

新試験では、試験問題が全体として難しかったか易しかったかに関係なく、
どの回の試験でも、「同じ能力」の人であれば「同じ得点」になるしくみに
変えました。これを「尺度得点」といいます。
尺度得点では、「言語知識」、「読解」、「聴解」のそれぞれの能力を測り、
0～60の目盛りがある尺度（ものさし）上で得点として示します。「総合得点」
はこの３つを合計したものです。同じレベルの試験の得点は、いつも同じ尺度
（ものさし）で計算します。

新試験では、みなさんのいまの日本語能力を、この共通の尺度（ものさし）で
測るため、みなさん一人一人が、どのような問題にどのように解答しているか
（正解したかまちがったか）を調べて、得点を計算します。ですから、たとえば
聴解試験で正解した数が同じ人がいても、正解した問題が違う場合には、得点
も違う得点になります。このように尺度得点では、いくつ正解したかがそのまま
得点にはなりません。

[piotrek]

That's quite a reply, isn't it? If I understood it correctly, depending on the actual difficulty of each question [defined by the number of people who answered correctly] some questions may be worth 1 point and some may be worth e.g. 3? As they said that the number of correct answers is not equivalent to the number of points you gain, that's the only possibily I can think of.

Am I on the right track?

[Pubeng]

[pj]

The new scoring system for the JLPT is almost certainly based on Rasch IRT:
http://en.wikipedia.org/wiki/Item_response_theory

Essentially, the scoring system aims to give you credit for how good you are, not just how many questions you got right. So, basically, as someone mentioned, you get more credit for correctly answering items that few other people could answer correctly (ie. if you can answer more difficult items that suggest that you've got a relatively high ability in Japanese).

But, people seem to be a little hung up on what the pass percentage is. The scores on the test are not based on percentages. Conceptually, this can be hard to grasp, especially since they've just moved from a system based on percentages. But, I say again, The scores on the test are not based on percentages. How it works is that they statistically figure out the difficulty of the items and the ability of the people taking test. And if they're using IRT, and I don't know of any other up-to-date way of doing this, the scores will typically be between -3 and 3 (this is just how the stats model works) and the average will be 0.

Then, they decide how to transform the scores into something that intuitively makes more sense to people than -.066, or whatever. One way of transforming the scores on a single test (they obviously didn't use this method for the JLPT scores) is thru the formula: X=50+10L (L stands for logits, which is the -.066 score or whatever) and X is your final score. In the example formula, the scores are centred on 50. And the spread of the scores that we get is hugely dependent on the 10 in the formula. So, my point is that the score we get is just an artificial creation. Do not think of it as a percentage. they could just as easily up everyone's score and you would have done exactly as well as you did, and it would be exactly as difficult to pass the test.

Anyway, the new scoring system is much better. And they can use it to keep the difficulty of the test consistent. But, don't get hooked up non-existent percentages.

[Hikouseibishi]

Years ago when I wrote the JPLT level 1 and failed the Dec 1997 and Dec 1998 tests with 64% and 65.5% I got consecutively, there were Chinese students that could pass because they did great on the Reading sections. They'd get a good score on the Vocabulary and Reading sections but the Listening sections they'd do poorly on. Also many of them could barely speak well at all which was apparent based on their listening score. However they got over the 70% pass mark and attained the Level 1 Certificate. I could speak much better than them in 1998 and didn't attain the certificate because I was under the 70% mark. My listening was around 49% at the time too as my listening practice was more Standard Japanese than what the listening actually was. Contracted Standard Japanese that I wasn't used to. I only knew non-contracted Standard Japanese and that killed me with the low listening score. Of course vocabulary also was part of that. I only needed a few more points to get my 70% but couldn't get it despite doing well in the Grammar section.
So it's due to the imbalance that the Chinese Students presented that they went to this system finally.
You must get a minimum of 100 points out of 180 but no individual section can be under 19 points to ensure there is some competency in all areas.
So: 20 + 20 + 20 = 60. NO PASS.
But: 19 + 40 + 41 = 100. PASS. 33+33+34=100. PASS.
How many questions on the test I don't know but some questions could be determined to be easy and only worth 1 point and some could be considered difficult so worth 2 or 3 points for example. So if you get the more difficult ones wrong you won't gain as many points. This seems fair to me as from my experience the test is pretty difficult. Level 1 (N1 now) has a lot of Japanese used in Literature and not the spoken kind. If you aim to speak better do Level 2 first.
It was my study for Level 2 that actually expanded my ability to speak at a normal conversational level. It was a big jump from 3 back in the mid 90s but I returned to Japan in Feb 96 and in Dec 96 passed level 2 with lots of study. I'd really recommend doing level 2 first if you haven't done it. It would be more beneficial to you and you'd probably learn to speak better with the grammar presented in that test.
For my study I used a book called 日本語能力試験に出る文法２級 and 日本語能力試験１．２級直前対策ポイント整理日本語文法 by アスク講談社 ISBN 4-87217-079-2. Probably not in print today. Also a good series to use that is really detailed is: 日本語総まとめ問題集2級文法編
Good Luck everyone. Keep up the study.