Does anyone know of any kind of way to efficiently check to see how common it is for a word to be written in kanji vs hiragana? Due to verb conjugations and the fact that many compounds are the same in Chinese (and thus looking them up in google inflates kanji numbers), it is hard to efficiently find out using google whether or not I should learn a word's kanji, or if they are only used less than 10 percent of the time and with furigana, in which case it is pointless to learn them at this stage.

For example, I wanted to see if I should learn 剃る (bad example since you'd be able to guess the reading from context but whatever), if you look it up in google to find out how often it has kanji vs hiragana you'd have to look it up in all verb conjugations (I don't know of any way to check all at the same time).

So is there any way to do this, or some database that shows exactly how commonly some words are written in hiragana? Jisho's "usually kana" designation is pretty broad (sometimes meaning just written in kana 60% of the time, in which case you should learn the kanji).

I actually do find the "usually kana" tag to be pretty accurate, even if there are some cases where people out there will use the kanji. But given that it seems to be pretty spot on the majority of the time it's easier to learn from the exceptions than it is to try and firmly establish every single word from the rule (assuming of course there is even a way to definitely establish character guidelines).

If you check on James Breen's WWWJDIC site, you will note that the (uk) label is sometimes applied to words for which a majority of example ("Ex" link, from the Tanaka corpus) are written in Kanji. Admittedly those examples are not claimed to be representative in this respect.
A related question is how to decide between hiragana and katakana (まぐろ vs. マグロ). I wonder if the (P) label attached to the corresponding spelling can be regarded as accurate.

I think I mentioned knowing about the UK in my original post. Deciding between a katakana vs hiragana is of course dead easy, as searching it won't bring up any Chinese examples and there are no inflections to worry about.

I never did find an efficient answer to this. The way i handled it was that if I met a word in the wild that I wanted to add to Anki, then if I met it in kana I created a kana card for it, if I met it in kanji I created both kana and kanji cards. If a met the same word later, thanks to anki magic, it's easy to 'turn on' the kanji version of the card. Regardless of how I met the word, I always included the kanji information in the base information. (I forget what it's called now, not a 'fact' but a... a... uhm... y'know, ... thing that cards are derived from. 'fact' was such a good term I dunno why that got changed.)

Because of multiple kanji spellings, I eventually revised my template to allow an alternate kanji spelling. I may eventually revise that to allow 4 or 5 spellings but not today.

One slightly useful hint is that when using a J-E dictionary like http://dictionary.goo.ne.jp/en/ , then just look at how it is spelled in the example sentences. The headword always shows the kanji if there is accepted kanji for it, but the example sentences (if there are any for the given word) follow the editor's notion of normal usage. No guarantees of course. Kanji usage varies wildly across different segments of society. I certainly have added kanji cards for many words that are spelled in kana in the J-E dictionary because i frequently encounter the word in kanji in light novels.

My solution was to learn the kanji for pretty much anything that had a kanji version in the dictionary I was using; depending on what you're reading, some words that are often written in kanji will be written in hiragana, some words will be written in kanji that usually aren't, sometimes things will be written in katakana for emphasis (or because the author feels like it, sometimes I can't tell...)
I remember reading something that used 子ども (or something like that) and 何故, possibly in the same sentence. And I just read something where they consistently wrote 子ども, but would use 供 in other words...
(This all depends on what you're reading, of course)

Point being, I think that one might as well learn the kanji version, since it's easier to recognize a word in hiragana that you learned with kanji than it is the other way around.

As for how to check... Other than the uk tag, it's just a matter of exposure, I think.

I tend to the opposite approach -- if EDICT claims 'usually kana', then ignore the kanji (which I find mostly makes the word easier to learn). If I do happen to find something that writes the word in kanji later it's easy enough to pick that up then (since I already know the word it's a smaller thing to remember).

As I go through core10k I'm dropping all the cards which are prompting me with kanjified versions of a bunch of basic conjunctions and similar 'grammar words'.

(2016-01-04, 11:00 pm)rainmaninjapan Wrote: (…) Due to verb conjugations and the fact that many compounds are the same in Chinese (and thus looking them up in google inflates kanji numbers), it is hard to efficiently find out using google whether or not I should learn a word's kanji,(…)

You should be able to restrict Google's search to Japanese by using the "Advanced search" feature and selecting "Japanese" in the "Language" field.

(2016-01-06, 2:04 am)jmignot Wrote:

(2016-01-04, 11:00 pm)rainmaninjapan Wrote: (…) Due to verb conjugations and the fact that many compounds are the same in Chinese (and thus looking them up in google inflates kanji numbers), it is hard to efficiently find out using google whether or not I should learn a word's kanji,(…)

You should be able to restrict Google's search to Japanese by using the "Advanced search" feature and selecting "Japanese" in the "Language" field.

domo arygahtoh m8ty