NEW Repetition Mode for Dialog Cards?

Submitted by etherscreen on Tue, 02/05/2019 - 16:02
Forums: 
Content creation

I discovered a new "repetition" mode for dialog cards.

Is based on spaced repetition as a learning method?

Is there any documentation about this functionality?

If yes, then do the dialog cards send completion statements for xAPI?

Thanks so much!

 

Attachments: 
Image icon repetition.png
BV52's picture

BV52

Wed, 02/06/2019 - 06:31

Hi etherscreen,

The repetition mode is inspired by the Leitner system but I'm not sure how similar or different it is with the space repetition.

There is no documentation yet for this feature and currently this content does not support xAPI.

-BV52

etherscreen

Wed, 02/06/2019 - 14:56

Thanks so much for this info.

Is xAPI on the roadmap?

otacke's picture

otacke

Wed, 02/06/2019 - 18:22

Hi etherscreen!

As BV mentioned, repetition is a "spaced repetition" mode (not storing the state to be able to continue later yet - unfortunately).

Instead of presenting the whole pool of cards to the learner, each round he/she is presented with cards based on her/his proficiency level (PL, imagine this as a pile of cards) for a card/on the results of the previous round according to the rules:

  • Each round consists of all cards currently on proficiency level #1, and "1 / number of proficiency level" cards for the next proficiency levels (except for the last one), so ½ cards from pile #2, ⅓ from pile #3 and so forth (the number of cards picked will be rounded upwards)
  • Cards on the last proficiency level are considered to be completed. No card will be picked from this proficiency level.
  • The LIFO (last-in-first-out) method is used to pick cards from different piles. The cards that have been in the pile the longest will be preferred over cards that have been added later.
  • If pile #1 is empty when a new round starts, pile #2 will become the new pile #1, pile #3 the new pile #2 and so forth.

The algorithm doesn’t reflect the original Leitner system, but will also make sure that cards that have been answered correctly are less likely to be presented to the learner in the next round than cards that have been answered incorrectly.

Best,
Oliver

papi Jo's picture

papi Jo

Wed, 02/13/2019 - 18:50

This new "repetition" feature is certainly more elaborate than my experimental "got it" feature proposed some time ago.

The only thing that seems to be missing from it is the "keep state" feature, which rather defeats the purpose. I hope you will be able to implement "keep state" in the future.

otacke's picture

otacke

Thu, 02/14/2019 - 10:33

Hi papi Jo!

We know that this feature is missing, but the repetition mode was contract work and it didn't cover "save content state". Shouldn't be a huge effort to add it, but someone has to have resources to do it.

Best,

Oliver 

etherscreen

Sun, 02/17/2019 - 05:35

What are the steps and estimated resources required to add "save content state"?

Thanks!

otacke's picture

otacke

Sun, 02/17/2019 - 10:58

Hi etherscreen!

You'd have to implement a function called "getCurrentState" that will be called by the core of H5P. In that function, you need to gather all the data that's necessary to identify the state (e.g. card piles, round number) and return it. These state data will be passed to the content type when it's loaded, and then the previous state needs to be restored. Also, you'd probably want to offer a solution to reset the task which would probably require a little thinking about UX at least and implementing a button or something similar. And of course, don't forget testing :-)

Cheers,

Oliver

etherscreen

Tue, 02/19/2019 - 22:37

Thanks for this. Know of any highly competent developers up to the task?

otacke's picture

otacke

Wed, 02/20/2019 - 01:21

Hi etherscreen!

Sure! The H5P core team at Joubel is probably the first address to go to, of course. This way you'd get the quickest (and guaranteed) integration into the official repository as everything can be decided immediately.

I could do it myself as well, but it would require me to at least discuss the UX part for resetting the task, and I could not tell you an exact date when the changes would be merged into the official repository (and I wouldn't want to create a fork). Even if I complete the code, someone from the core team at Joubel would have to give it a code review first.

Best,
Oliver

otacke's picture

otacke

Fri, 03/08/2019 - 13:37

Hi etherscreen!

I implemented the feature: Unlock 'spaced repetition' for Dialog Cards

Cheers,
Oliver 

otacke's picture

otacke

Fri, 03/08/2019 - 13:36

Hi Papi Jo!

There you go: Unlock 'spaced repetition' for Dialog Cards

Cheers,
Oliver 

GrantSWilson

Tue, 02/19/2019 - 20:11

Hi there,

I have a question about the "Repetition" mode of the Dialog Cards. Maybe it's a bug or maybe not; it could be I'm just not understanding how the "Maximum Proficiency Level" setting works.

In Repetition mode, the Behavioural Settings let me select the "Maximum Proficiency Level". The default is 5. I take this to mean that the learner needs to self-report a card as being "right" 5 times in a row for it to considered as "completed learning".

Here's the issue: if I have "Maximum Proficiency Level" = 5, the rounds stop and I get the "Well done..." message after getting all cards "right" 4 times in a row, not 5 times in a row.

If I set "Maximum Proficiency Level"  = 4, the rounds stop when I get cards correct 3 times in a row. The number of times in a row that is considered "completed learning" is always one fewer than the number in "Maximum Proficiency Level".

Is this a feature or a bug? :-) Or am I just not understanding how the "Maximum Proficiency Level" is supposed to work?

I'm using H5P on WordPress, plugin version 1.12.0; the Dialog Card Library version is 1.7.7.

I've attached the Dialog Card interaction I've been testing with.

Thanks and take care!

H5P file: 
Binary Data Dialog Card Example
otacke's picture

otacke

Thu, 02/21/2019 - 03:34

Hi Grant!

Nothing wrong here. The learner simply starts at "proficiency level" 1, not 0. In consequence, the maximum proficiency level is the "number of boxes" the cards are put into in terms of the classic analogue approach.

Best,

Oliver 

GrantSWilson

Thu, 02/21/2019 - 06:52

Hi Oliver,

Thanks for the reply. I'm glad to hear all is well. I'll still need to wrap my head around the underlying system being used and your Feb 6 post will help a lot with that. 

Thanks again and take care!

jimmyjames

Tue, 03/12/2019 - 06:15

Hi Oliver @otacke , I checked out the Indigogo page and the features look GREAT!  I didn't see anything though on "submit answers" or xAPI output for those of us using LRS.  Is there any output that can be used for this to track ANY interactions?

otacke's picture

otacke

Tue, 03/12/2019 - 07:27

Hi Jimmy!

DialogCards doesn't have any xAPI support so far, but I was actually planning to put that into the campaign as a bonus :-) However, I still need to figure out what xAPI verbs should be used, what the object definition should look like in order to make sense, and when the statements should be triggered.

Best,

Oliver

doanschool

Tue, 04/21/2020 - 12:01

Hi Oliver. My nam is Mao. Thank you so much for the Dialog Cards. I've been looking for a spaced repetition tool to help my students learn vocabulary and found your great apps.

I have a question about the algorithm behind the repetition mode. Even if the learners have "mastered" a word after 5 correct answers in a row, they may forget it sometime in the future. I've been following my students and found that the answers that they ticked "got right" and considered mastered nearly a week ago have not come back so that they have to do some more rounds. The only option is to try all again.

I'm I mistaken about it or won't the app predict the point in the future when the learners should review a particular card? Is an algorithm like SM2 in Anki or SuperMemo functioning or if not, would you consider adding it in the future? 

Thank you very much.

Mao 

otacke's picture

otacke

Wed, 04/22/2020 - 01:59

Hi Mao!

It's not me who created DialogCards, I was merely the one to implement the repetition mode when I was working for Joubel.

I described the algorithm above in a previous post. The specs were designed to suit the sponsor's needs in particular.

I have a long list of things that I plan to contribute to H5P in my free time, and I'll gladly put some other algorithm options for DialogCards onto the list as well - but I don't know when I'll come around to it.

Best,

Oliver 

 

doanschool

Sat, 04/25/2020 - 04:30

Hi Oliver,

Thank you for your reply. I will be very happy to see the new options. Not only for Dialog Cards, but I think it will also be very helpful if added to and across other types of H5P activities, flashcards, quizzes and so on. A rigorous spaced repetition algorithm like this can improve students' retention rate tremendously. I know that a lot of teachers and schools are interested in this.

So if you have such a plan to add this new feature in the future, please let us know, we'll be very thankful and glad to make contributions to our best ability. 

Thank you and take care.

Mao

otacke's picture

otacke

Sat, 04/25/2020 - 19:21

Hi Mao!

I think I'll have to dig into that topic a little bit, but if you are suggesting that "all" contents should get some repetition features, it would be more sensible to either write a wrapper content type (like question set is a wrapper for other content types) or to add such a feature to the H5P plugins.

Best,

Oliver