I recently took a look at the "progress" page for the whole year and noticed something interesting.

Added Learned # Reviews
Words 777 764 14738
Characters 1277 1250 51978
Tones 1163 1152 32947

If you calculate the number of reviews per item learned, you get

# Reviews/Learned
Words 19.3
Characters 41.6
Tones 28.6

So you can clearly see that, at least in my particular case, learning words is much quicker than learning characters, and learning tones is right in the middle. I suppose that in my case the words are the easiest because I always add the individual characters in each word to my queue, and add lots of words that are made up of characters I already know. However, the fact that the tones are substantially easier than the characters is a little surprising to me.

Are other people seeing similar numbers? Extremely different numbers? Don't feel compelled to post your complete statistics if you feel like that's too personal. ;)

It seems like it would be interesting to add these numbers to the progress pages and/or the progress reminder emails. Even more interesting would be a breakdown that showed which characters/words were causing us the biggest problems. As a former statistics teacher I would love access to the raw data so we could actually test hypotheses like "character complexity doesn't relate to difficulty in retention".

James

I'd never thought of calculating that statistic, but here's what I got:

W: 34 / 757 (22.56)
C: 93 / 2201 (23.67)
T: 91 / 1239 (13.61)

I imagine that since you have substantially more learned than I do that it's not an apples : apples comparison. But the statistics aren't entirely descriptive, either, I suppose; if there was a breakdown of "learning" reviews vs. "retention" reviews it would probably give a better picture. But even ignoring that (and the difference between added/learned), I'm surprised at the relatively low number - I would have guessed significantly higher.

Here are mine:

Added Learned #Reviews
Words 1922 1921 18414
Characters 1723 1688 55699
Tones 1713 1704 35290

So:

# Reviews/Learned
Words 9.59
Characters 33.0
Tones 20.7

I'm not sure this means that words are easier than characters though. By their nature, characters are reviewed at least as but usally more often than words. Additionally, after you've been studying long enough the number of words learnt will exceed the number of characters. Therefore, it's inevitable that the reviews/learned ratio will be higher for characters.

Ben

Mine, although I haven't added many words yet.

Words 47 42 221
Characters 519 488 4232
Tones 532 514 3438

So that's er.

Words 4.7
Characters 8.2
Tones 6.4

Low as I haven't really been doing this very long...

I've pondered this, too. My numbers are a bit different:
w: 1393/10950 - 7.8
c: 1362/29747 - 21.8
t: 1368/43060 - 31.5
I wondered whether others were as bad about learning tones as I am, apparently not. Part of it is that I feel I have pretty much mastered the character if I can write it and the tone is really more about speaking. Regrettably, I am sloppy about the tones in speech, too.

Knowing that my husband is more careful about tones, I checked his stats:
w: 420/4454 - 10.6
c: 541/11962 - 22.1
t: 536/15390 - 28.7
Not what I expected, maybe there's a family component here, too.

By the way, in the era before Skritter, we used paper flashcards and I tended to ignore tones but Skritter doesn't let that happen :) Recently we have reviewed those paper cards (char on one side, only pinyin on the other) and I was gratified that I was getting the tone right nearly all the time!

On a slightly related note: the shorter delay when first learning a character ( I think I read it was shortened to 30 seconds) is a real help to me. I believe it is taking fewer reviews to learn a new character and I like that.

Don't know if I can really add to the conversation, but thought I'd add to the data collection. I started Skritter at the beginning of September, so I haven't been at it long. Had to do it; I was curious.

Added Learned # Reviews
Words 514 500 5822
Characters 660 630 16192
Tones 496 478 8639

Avgs:
Words: 11.6
Characters: 25.7
Tones: 18.07


I have a question that's kind of related to this exercise: does anyone feel like Skritter isn't reviewing recently learned characters as often on the new practice page as on the old one? Maybe I'm wrong, but it seems like they aren't coming up as often any more...

The data:

Words 804 801 7107
Characters 1021 998 19277
Tones 1025 1021 7108

Yes, words are remembered more easily. But isn't that to be expected. After all, as soon as you have two characters appearing together you have context, and context is what it is all about.
sarac: I have the same experience. I didn't mind the tones too much before. Kept postponing really studying them. I expected the tones to stick as I would be learning mor phrases (context). But now I feel more confident about them. Somehow tones seem to stick now.

Very interesting! Here's my stats:

Type: # Learned, Reviews Per, % Reviews Correct
Words: 3400, 4.8, 88.4%
Chars: 2122, 18.7, 90.9%
Char Tones: 1953, 27.8, 91.5%
Word Tones: 2472, 5.9, 85.8%

Progress page doesn't give you all these, sorry. Retention rates would be roughly 94%, 95%, 96%, 93%.

So for me, I review characters' tones more often than their writings, but I remember the tones better. Maybe it just goes to show that I can always forget tones of words I knew well, bringing them back down to shorter intervals and causing a lot more reviews of them to happen, whereas it's less likely to forget a character once I've learned it. That's one possible explanation. Or I just pay less attention to tone prompts, occasionally gettings ones I wrong that I should have known.

I'm not yet sure what to make of the character / word differential, though.

Here are some juicier stats. I just ran a dirty average of all users who have studied more than 12 hours and learned at least 300 characters.

Chars learned: 874
Words learned: 720
Hours spent: 51
Minutes per char learned: 3.5

Some users are as low as 1 minute per char learned! I'm guessing most of them knew a lot coming in, though.

When you start, you know more single characters than multiple-character words using them. As you get higher, though, you get more and more words.

I tried analyzing the minutes/char for those users who are mainly learning single characters, but I didn't get very good data out of it yet. I think I'd need to put more work into it to identify who's doing something like a Remembering the Hanzi approach.

Some other random statistics I made:
We have about 5000 Chinese characters and their traditional forms now. We've defined 26,000 words. For Japanese, it's 2300 characters and 7000 words so far.

As far as making more data available: we've got a ton of raw data buried in our datastore, but none of it is organized--it'd take some work to make it available either in raw form or on the progress page. It is something we'd like to do, but can't be a super high priority. We've got stuff like rankings of words by importance that combines standard frequency and ordering in textbooks. We've got graphs showing the relative vocabulary levels of all our textbooks. Tons of cool stuff, so little time!

Certainly for me the 'minutes per character learned' is going to be seriously screwed... Often I spend ages thinking of a mneumonic and writing it, or more likely I go on facebook or forums, or I make myself some lunch...

Still, it's quite nice to know how other people are doing to compare your progress with :p.

Do you guys have any particular techniques for remembering tones? I write the character down in colours which I associate with the tones, so I either think of that or go with gut instinct. It's not very efficient...

Words 1082 1069 7015
Characters 1064 1062 20309

Words: 6.5

Characters: 19.1

I decided not to try tones way back in the day when I first tried skritter out, just to see how it was. Now I'm kind of regretting it, but also terrified of what might happen if I start them now....

Nicki: believe me tones are essential. Make a start now.

I would go with what marchey said, Nicki. Tones are essential and quite possibly one of the most difficult things to get the hang of when speaking. Granted, it be discouraging at the beginning but in the long term it's totally worth it (and anyway...the fact that you are learning Chinese proves that you aren't easily discouraged:).

Thanks. I know the tones are essential and I am learning them, I just am not currently practicing them on skritter. I'm just afraid of what will happen if I enable them now, with over 1,000 characters and 1,000 words already "learned" in the system.

It will only add the tones for new words as they're added, so you won't get a sudden influx of tones. You can also turn them off again at any time.

Ok, thanks for letting me know! I'll try it.