I have noticed over time that the (multiple choice) answers the Memrise presents are not always random distributed. I have had the feeling for 2 months, but didn’t have time to check it. To be sure it was not a feeling, I decided to consistently make screen shots of every type-the-answer question in one single review session. Here is the result.
The right answers are:
감동적 impressive
꼭 certainly
외식 eating out
병 illness
나이 age
데이트 date
울다 to cry
병에 걸리다 to fall ill
때문에 because (of)
The answer is nearly always in the last buttons, must more often than we expect by chance.
Issue on Android Kit-Kat, Memrise v2.9_3930_memrise. Issue also present in (a) previous version (not sure which version that was, but I downloaded that version early March).
Could Memrise perhaps confirm the issue and offer a fixed version of the App? I’ve spend time copying and pasting things and reporting the issue, and the issue limits the ability to be randomly tested. Any good solution highly appreciated. Thank you!
Did Memrise promise that the answers would be randomly placed? What if they have collected data that says that placing the correct answers towards the last buttons increased learners’ passive exposure to all of the answers, and that correlates with improved metrics for successful learning (not just testing) of all the choices that are not selected?
Maybe they aren’t doing anything like that, but there is a reason that they build the multiple choice out of items you are learning, and don’t have you select from misspelled answers and grammatical errors. The multiple choice reviews are not just tests, they are learning activities including the other entries.
Okay, I am probably dreaming about Memrise using data mining to improve instruction. But there’s already plenty of research supporting mixing true testing modalities with low-demand exposure activities. Even if this is a bug, it probably supports your language learning.
[quote=“lurajane, post:2, topic:13010”]
Did Memrise promise that the answers would be randomly placed?
[/quote]Well, as with any test, we commonly don’t know what the right answer is. If they have a different idea from what is common, that is fine of course, as long as they tell me! How can I know? I’d like to make an informed decision whether to buy or not!
But more to the point: yes, the fact that it is not very random does impact my learning negatively, because I tend to get the right answer more easily and subsequently fail my test in class where I don’t have this advantage!
It is true tough that starting with low demand improves most people’s learning - it’s called scaffolding, which starting very close to things to learner already masters and move further and further away from that starting point.
