https://www.duolingo.com/Schengis

Fluency Estimate = Detrimental?

Schengis
  • 13
  • 12
  • 10
  • 9
  • 8
  • 7
  • 7
  • 4
  • 3
  • 3
  • 3
  • 3
  • 3
  • 2

I've been using duolingo for some time now, and I noticed the fluency estimate since the day it was put up. Although I assume that it's a work in progress, I feel that it is ultimately detrimental to the process of language acquisition due to an inaccurate assumption directed at the learner.

I myself have studied several languages in a university setting and have learned much through immersion in foreign countries. I can assure anyone reading this that my fluency in both of my most fluent languages is not even remotely close to that of my estimated fluency as presented by duolingo. So, why does that matter?

Two reasons: First - When I see that here, especially having completed the skill trees in both languages that I consider myself fluent in, I take it as a little slap in the face and/or a poorly constructed system, not as a mode of encouragement to learn more (as the intent states).

Second - For those who use this website daily without conjunction with other materials (e.g. dictionaries, articles, videos), then they certainly do have a limited parameter for their complete vocabulary, although many lessons do include a broad base of subjects.

With all of this being said, and due to my laziness in the wee hours of the morning, I'd like to simply suggest that the fluency estimate be removed. I sincerely feel that it causes more harm than good, and may make many users feel inadequate, or perhaps overly confident in their skills only to get put in their place later on. Given the platform, it just isn't a helpful or beneficial addition to the site or app.

Thanks for reading. Comments welcome, although I may not reply due to the feebleness of my argument in the scheme of things.

3 years ago

23 Comments


https://www.duolingo.com/Rob2042

Completely agree. It is misleading and therefore detrimental.

3 years ago

https://www.duolingo.com/Cambrian0616

He's right, you know. Fluency just doesn't map to an integer.

3 years ago

https://www.duolingo.com/TrioLinguist

Exactly. It's absurd to give fluency a percentage - if someone told me they were 29% fluent in German, I would be taken aback at how misguided they were.

3 years ago

https://www.duolingo.com/Schengis
Schengis
  • 13
  • 12
  • 10
  • 9
  • 8
  • 7
  • 7
  • 4
  • 3
  • 3
  • 3
  • 3
  • 3
  • 2

Yep. I'm fluent in both Spanish and German in the "real world" and speak German most of my day., I've barely started French and my percentage is nearly twice as high as my German. I understand the system enough to comprehend WHY it's unusual, but come on...

3 years ago

https://www.duolingo.com/rspreng

I agree.

3 years ago

https://www.duolingo.com/Angie766657

I agree as well. I'm at Level 7 and it shows I'm 22% fluent. NO WAY. I had a hard time telling my son "I want a big glass of water", in Spanish tonight. I said it, but it took me longer to say than I figured it would. If I was reading or translating on here, it would have been a piece of cake, so I assumed I would easily speak such a simple phrase. Not yet..... Reviewing and actually speaking/understanding are two different things!

3 years ago

https://www.duolingo.com/SquirlRat
SquirlRat
  • 25
  • 9
  • 3
  • 1219

A little more transparency about how the percentage is calculated would help. I had to go searching through threads and found conflicting information. I think it is 'percentage of words (counting each occurance of a word separately) you would understand in an average text of that language' but really not sure. Misleading if people assume it is calculated differently.

3 years ago

https://www.duolingo.com/Ana.salimi
Ana.salimi
  • 21
  • 7
  • 6
  • 3
  • 3
  • 2
  • 2
  • 2
  • 2
  • 2

agree.my Fluency Estimate is the same for a very long time, even though i have probably improved!!!!!

3 years ago

https://www.duolingo.com/SquirlRat
SquirlRat
  • 25
  • 9
  • 3
  • 1219

If the fluency estimate works how I think it does then words earlier in the tree that are more common matter more than less common words later in the tree, therefore the increase from learning several new words would be counteracted by a few of the more common words reducing in strength.

3 years ago

https://www.duolingo.com/KevanSF
KevanSF
  • 25
  • 25
  • 25
  • 25
  • 25
  • 19
  • 16
  • 16
  • 14
  • 13
  • 13
  • 12
  • 10
  • 8
  • 7
  • 7
  • 6
  • 6
  • 6
  • 6
  • 6
  • 6
  • 6
  • 5
  • 5
  • 4
  • 4
  • 4
  • 4
  • 4
  • 3
  • 2
  • 2
  • 2
  • 2
  • 1320

agreed

I've already explained my rationale and opinion quite thoroughly two of the previous times this topic has come up.

Since this topic does tend to come up every week or two, I would think that TPTB at Duo would eventually take notice...

3 years ago

https://www.duolingo.com/gabzerbinatoEng
gabzerbinatoEng
  • 11
  • 11
  • 9
  • 7
  • 6
  • 6
  • 5
  • 5
  • 4
  • 4
  • 3
  • 3
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2

I couldn't have said it with better words.

3 years ago

https://www.duolingo.com/Shimna1

Surely fluency isn't about the number of words you've learnt, but the speed and ease with which you can recall them and use them to construct a natural sounding sentence. I don't think that Duolingo is measuring that (unless it has an algorithm to measure the speed at which you supply correct translations), so the fluency estimate is flawed, of course. However, if you view Duolingo as a game, the fluency aspect is just one more score you can use to compete against your friends, or one more challenge in the learning of a language.

3 years ago

https://www.duolingo.com/HuShifang
HuShifang
  • 22
  • 16
  • 16
  • 15
  • 15
  • 9
  • 8
  • 6
  • 6
  • 5
  • 5
  • 2

Agreed.

Obviously, there are profound epistemological issues with slapping a number on fluency, but even in terms of "engagement", I suspect it's counter-productive (in ways that, from what I can gather from my perspective as an outside observer, will elude Duo's metrics). That is, I think that while it may be at least somewhat motivating to the majority of users (i.e. those who don't progress very far, or very quickly), I suspect that for those who actually hit the 50ish% ceiling, it's demotivating. It has been highly discouraging for me: the site's insistence on quantifying fluency make me suspicious of the soundness of its pedagogy, certainly, but more importantly, the impossibility of raising my "fluency" score discourages me from using Duolingo. All I can do is run in place, so to speak -- no matter how much I practice, I'll never make any significant forward progress. And, if I'm going to grant at least some credibility to the score qua indicator of proficiency rather than simply reject it out of hand, the limitations of Duolingo are really highlighted: I get a big badge reminding me of how this site has nothing more to teach me. I've learned a solid set of vocabulary, and I can drill it, but that's all.

Probably, the real motivation behind the fluency score was replacing the points & levels system without replacing it. That is, introduce a new metric to supersede the old one without doing away with it entirely, and thereby angering people who'd slaved away for their points and levels. There are advantages to the new metric -- "percentage fluency" will make more sense qua objective certification than "Duolingo levels/points", and Duo seems to be moving more towards testing and certification in its business model -- and maybe there are some people with lots of points and levels who'll be briefly motivated by a new bauble to seek. But absent new content they'll inevitably be demotivated, for the reasons above.

Does having parallel metrics make sense as a stopgap measure? That is, is Duo better off gently easing people away from points and levels moving forward rather than having simply done a quick conversion? I think that remains to be seen. The fact that my level 15 French is far more "fluent" than my level 21 German suggests to me that past a certain point, points and levels don't matter. If we were to, again, grant at least some credibility to the fluency percentage, then that means that Duo's points and levels were a bad system, and that someone who's earned multiple level 25s has wasted a lot of time -- even if the practice was good for them, they weren't learning or objectively-improving anywhere near as much as they might have thought. Maybe having parallel metrics in place diminishes the probability that they storm away from the site in anger -- they can be frustrated about their fluency percentage, but they're still level 25.

Really, the solution is more content -- more lessons, more types of lessons, etc. Something that will advance fluency past 50ish%. (And, if Duo's smart, they'll keep the levels around and raise the cap too, in the interest of postponing indefinitely people's potential, if not necessarily legitimate, feeling of having been had.)

3 years ago

https://www.duolingo.com/harryclark17
harryclark17Plus
  • 24
  • 19
  • 15
  • 11
  • 10
  • 7
  • 7
  • 7
  • 4

It's something to get people hooked. Seeing progress towards fluency is exciting. It's all about getting people to keep learning. Downlaod TamperMonkey for your browser and check these out: http://duolingo.wikia.com/wiki/Duolingo_Userscripts

It has a link to an app that will let you turn of the fluency marker, and a bunch of other super useful stuff.

3 years ago

https://www.duolingo.com/Schengis
Schengis
  • 13
  • 12
  • 10
  • 9
  • 8
  • 7
  • 7
  • 4
  • 3
  • 3
  • 3
  • 3
  • 3
  • 2

Getting them hooked is one thing, but there isn't exactly an outstanding business goal for these guys at this point, and it's a free program.

3 years ago

https://www.duolingo.com/harryclark17
harryclark17Plus
  • 24
  • 19
  • 15
  • 11
  • 10
  • 7
  • 7
  • 7
  • 4

Yes, but to keep Duolingo alive, they have to have people using it.

3 years ago

https://www.duolingo.com/Schengis
Schengis
  • 13
  • 12
  • 10
  • 9
  • 8
  • 7
  • 7
  • 4
  • 3
  • 3
  • 3
  • 3
  • 3
  • 2

Indeed. However, I do think that they are succeeding through their use of experience and leveling to accomplish that. Just like any "skill grinding" format, it's an effective hook to a certain point.

The community is also a strong and progressive one intent on maintaining.

3 years ago

https://www.duolingo.com/SquirlRat
SquirlRat
  • 25
  • 9
  • 3
  • 1219

In some ways I think the 'fluency' rating is a more meaningful metric than the level/XP as it is actually a measure of proficiency rather than just time spent on duolingo. I'd like the option to show ones 'fluency' rating publicly to be honest. It is clear though that many users are not happy with this rating being referred to as fluency, maybe Duo should come up with a better name for it?

3 years ago

https://www.duolingo.com/gringorjones

I agree with you. It's really a naming issue. I like that there's a number I can use when talking with my friends who do duo. We can see how each of us is progressing through our languages. I don't think any of my friends are foolish enough to think that it equates to real-world fluency. It's just a good measure of your progress through the duo app. I wouldn't even mind something that just counted your % through the lessons on your tree, if that was simpler for people to understand.

3 years ago

https://www.duolingo.com/Schengis
Schengis
  • 13
  • 12
  • 10
  • 9
  • 8
  • 7
  • 7
  • 4
  • 3
  • 3
  • 3
  • 3
  • 3
  • 2

The problem is how they define fluency. If you're frequently using the program and working your way slowly through the skill tree, then you'll be considered more fluent than someone like me, who just tested out of all the levels on the skill tree. I live and work in my second and third languages, and I have a speaking fluency that tests C1 or higher (enough to get residency in a foreign country). Therefore, my Duolingo fluency is perhaps about 1/4th or less of my true fluency.

The problem is that it is not reflective of true fluency, and there isn't really a method that can do that in this setting, so it should be removed or re-named. Possibly a "vocabulary meter", or in the direction of word recognition.

3 years ago

https://www.duolingo.com/Jessica118867

I did a controlled test of fluency estimates... I am a native Spanish speaker. According to Duolingo, my Spanish level is less than that of my French or Italian. I'm guessing it might be partly because I refuse to use redundant pronouns when the verbs have got it covered? I don't really know. In any case, my French is pretty good, but I've only taken a semester of Italian, and they're at the same level... make of it what you may, but those percentage levels of fluency don't really mean much.

1 year ago

https://www.duolingo.com/AdamWallace33
AdamWallace33
  • 22
  • 13
  • 11
  • 10
  • 475

I definitely find the fluency rating more discouraging than anything. It outright punishes you for not putting fifty XP into a lesson every day. I take 1 strengthening lesson in Italian every day, hardly screwing up, and my rating is STILL dropping. As far as I'm concerned, it should only go down if mistakes are being made.

1 year ago

https://www.duolingo.com/JuniorSniper

I just think Duolingo should do fluency estimation in a different way besides percentage.

1 year ago
Learn a language in just 5 minutes a day. For free.