Feedback on word strength system [suggestion]
First of all I want to say that I think Duolingo's best feature is its word strength system [even though I sometimes post constructive criticism regarding it. (read: complain about it) :) ]
Flagging weak words is suggestion that has been posted on numerous occasions. In essence it would give the user ability to manually adjust the word strength.
The thing I'm suggesting is similar but it wouldn't have a direct effect on the word strength system.
It would be used to *mark and rate words in order to *provide feedback.
Why? Because then they could compare our word strength data with how we think we know those words and maybe use those comparisons to improve upon the system.
How would providing feedback work? (look at the picture below) First of all I think it should be optional. I imagine something like a "Word strength" button that would be shown next to our "report a problem" button. When we press that button we are able to click individual words which opens a dropdown menu with list of choices for rating the strength of that word.
explanation of the choices from the dropdown menu
- ++++ I know this word perfectly.
- +++ I could recall this word with ease.
- ++ I knew this word but I am still not confident using it.
- '+' I had trouble recalling this word.
- 0 I didn't know this word at all.
I welcome any constructive criticism regarding the title of this thread, mechanics of the feature (how would it work), choices for the dropdown menu :) and of course the idea itself.
yes please! But I might be completely ignorant to what it entails programming this XD. Might be complicated.
As a general question, I have been wondering whether any of the suggestions that are being made regarding flagging words independently etc might mess up Duolingo's data analyses. I know they put a lot of effort into building models of how we learn, and this would potentially add annoying parameters to the analyses they do.
(On this note I would love to know what tatou's up to on a day to day basis - his jobs sounds dead-cool!!!)
Yes, this is an excellent way of determining if word or concept is difficult. Duolingo can only test whether we know the word perfectly, not if the word is quite hard for most people. So this would help to check their model against what we think, to see if overall users believe words are hard.
Isn't the setup good as is? If you know a word well, you're going to get it correct and that's recorded in the algorithm. If you have trouble with a specific word, you're more likely to get it wrong and likewise that will be tallied as well. How well we know a word is reflected by how we answer exercises associated with that word, no?
And what if you're confident about a word, but you take a break from Duolingo for two weeks (for whatever reason) and aren't as confident as you once were - do you have manually change it again?
What you said is correct but the current system is limited as it is because what if you are learning the same language in school, or using some other language software beside Duolingo or you have previous knowledge of that language. Data the system collects is limited in that regard. The decay rate will be the same as if you are a beginner.
Also what if someone is tired or somewhat distracted and makes a lot of typos when writing words he/she knows perfectly and the system lowers the word strength accordingly.
Because of that and probably some other reasons there have been dozens, scratch that, hundreds of posts suggesting or asking for ability to flag weak words (example 381 likes from year ago http://www.duolingo.com/comment/35498 ).
Even I've requested that feature but nowadays I am not so fond of that idea because user's tinkering with such a complex system could make it worse and more importantly doulingo wouldn't have a unified system for all its users making word strength data analysis substantially more difficult.
I think our data that Duolingo collects is more reliable but the system should as well take in account other factors.
That is why I posted this crude suggestion which is essentially data collection to determine how important other variables are and how our beliefs regarding word strength compare with data collected by Duolingo.
As for your last question I don't know at the moment. I will think about it.
Personally, I think this is subject to gross misinterpretation. It is far too subjective. If I understand what you're proposing correctly, this method would rely completely on a user's own self-assessment of how well he or she knows a word. It's akin to having students finish an exam in school and rating their performance based off of how well they think they did.
Plus, it is really similar to the ubiquitous flashcard systems in existence today. Doesn't duolingo want to distinguish itself from a run-of-the-mill, self-assessing flashcard tool? If this somehow has great appeal among users, I would suggest that it be optional. I think that the length of time it takes a user to answer correctly (or incorrectly) combined with how many times a user gets a word correct combined with the length of time between testings would be far superior tests of whether or not a user had really mastered a word.
I realize you posted this quite some time ago, but I'm just seeing it for the first time now. I like your creativity and interest in improving duolingo and for asking for constructive criticism. I hope my directness and criticism of this idea is not taken as a criticism of you. It isn't. I also hope you continue coming up with ideas for duolingo, sharing them with us, and encouraging us to give you our honest opinion. In the meantime, best wishes with your creative pursuits.