After using the flash cards for the past couple months, I've come to realize that the spaced repetition works on a per-word basis. This means that one can, for example, be able to understand a spoken word when hearing it, but have no ability to read or recall the word. In this situation, you end up waiting too long to be shown the card again.
I think each 'mode' of a given flash card needs to be spaced on its own, because it is in effect it's own flash card. So for a given word, there should be separately spaced and repeated cards, for:
* Reading (kanji, kana, or romaji given. User responds with meaning and pronunciation)
* Listening (audio spoken. User responds with meaning and written form)
* Recall (English given. User responds with pronunciation and written form)
In the current method, where they're all jumbled up, there is no way effectively ensure that all three skills will be appropriately tested.