![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3705461%2F574a1b5a93f132703017147a3b390ce3%2Fq_l_q.jpg?generation=1604961542797271&alt=media\")
I am running some data analysis and I am trying to calculate some metrics for questions by shifting the prior question elapsed time and explanation per user. And by aggregation, it turns out the same question may have an explanation sometimes and other times it won't. I also found the below when a user answered the question incorrectly, then a lecture and directly again the same question (which answered correctly this time)
\n
My questions:
\nI am running some data analysis and I am trying to calculate some metrics for questions by shifting the prior question elapsed time and explanation per user. And by aggregation, it turns out the same question may have an explanation sometimes and other times it won't. I also found the below when a user answered the question incorrectly, then a lecture and directly again the same question (which answered correctly this time)
\n
My questions:
\n1. is there a rule when a question has an explanation?
My observations:
\nFrom these facts, a simply rule saying so and so questions get an explanation can be safely ruled out. My assumption therefore, is that most questions come with explanations by default and the user may choose not to view the explanations. The pq_had_explanation might actually mean pq_explanation_viewed. This is also consistent with the fact that users can choose the content they engage with (check out this paper).
2. is there a rule for questions re-occurrence within a short time frame?
From my understanding, certain users may mark a question for review. And after watching a lecture to understand the concepts better, she may try to attempt the questions again. So there isn't any rule to determine when a question might occur again. Take user 15632472 for example, this person attempts the same question over and over again (max of 30 times for a single question) - A rule based system would never ask questions that frequently.
\nAlso the RIIID github repo says this:
The student can re-do the questions that he already solved before using review system, which is recorded as in_review.\nHope this helps :)
", "rawMarkdown": "```1. is there a rule when a question has an explanation?```\n\nMy observations:\n- Certain questions explained for certain users, were not explained for a different set of users. \n- A very few set of questions were *never explained*. \n- The first 30 or so questions are *always without an explanation* as these are part of the app's diagnostics tests. \n\nFrom these facts, a simply rule saying so and so questions get an explanation can be safely ruled out. My assumption therefore, is that **most questions come with explanations by default and the user may choose not to view the explanations**. The `pq_had_explanation` might actually mean `pq_explanation_viewed`. This is also consistent with the fact that users can choose the content they engage with (check out this [paper](https://arxiv.org/abs/2005.03818)).\n\n***\n\n```2. is there a rule for questions re-occurrence within a short time frame?```\n\nFrom my understanding, certain users may mark a question for review. And after watching a lecture to understand the concepts better, she may try to attempt the questions again. So there isn't any rule to determine when a question might occur again. Take user *15632472* for example, this person attempts the same question over and over again (max of 30 times for a single question) - A rule based system would never ask questions that frequently.\nAlso the RIIID github repo says this:\n```\nThe student can re-do the questions that he already solved before using review system, which is recorded as in_review.\n```\n\nHope this helps :)", "votes": null }, { "id": "1074564", "postDate": "11/10/2020 19:53:52", "content": "Very helpful. Thank you for sharing your observations
", "rawMarkdown": "Very helpful. Thank you for sharing your observations", "votes": null } ], "comments": [ { "id": 1074302, "author_name": "doctorkael", "author_url": "", "post_date": "11/10/2020 13:54:26", "content": "1. is there a rule when a question has an explanation?
My observations:
\nFrom these facts, a simply rule saying so and so questions get an explanation can be safely ruled out. My assumption therefore, is that most questions come with explanations by default and the user may choose not to view the explanations. The pq_had_explanation might actually mean pq_explanation_viewed. This is also consistent with the fact that users can choose the content they engage with (check out this paper).
2. is there a rule for questions re-occurrence within a short time frame?
From my understanding, certain users may mark a question for review. And after watching a lecture to understand the concepts better, she may try to attempt the questions again. So there isn't any rule to determine when a question might occur again. Take user 15632472 for example, this person attempts the same question over and over again (max of 30 times for a single question) - A rule based system would never ask questions that frequently.
\nAlso the RIIID github repo says this:
The student can re-do the questions that he already solved before using review system, which is recorded as in_review.\nHope this helps :)
", "votes": null, "replies": [ { "id": 1074564, "author_name": "wisamnaji", "author_url": "", "post_date": "11/10/2020 19:53:52", "content": "Very helpful. Thank you for sharing your observations
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1073773": "I am running some data analysis and I am trying to calculate some metrics for questions by shifting the prior question elapsed time and explanation per user. And by aggregation, it turns out the same question may have an explanation sometimes and other times it won't. I also found the below when a user answered the question incorrectly, then a lecture and directly again the same question (which answered correctly this time)\n\n\n\nMy questions: \n\n1. is there a rule when a question has an explanation?\n2. is there a rule for questions re-occurrence within a short time frame?", "1074302": "```1. is there a rule when a question has an explanation?```\n\nMy observations:\n- Certain questions explained for certain users, were not explained for a different set of users. \n- A very few set of questions were *never explained*. \n- The first 30 or so questions are *always without an explanation* as these are part of the app's diagnostics tests. \n\nFrom these facts, a simply rule saying so and so questions get an explanation can be safely ruled out. My assumption therefore, is that **most questions come with explanations by default and the user may choose not to view the explanations**. The `pq_had_explanation` might actually mean `pq_explanation_viewed`. This is also consistent with the fact that users can choose the content they engage with (check out this [paper](https://arxiv.org/abs/2005.03818)).\n\n***\n\n```2. is there a rule for questions re-occurrence within a short time frame?```\n\nFrom my understanding, certain users may mark a question for review. And after watching a lecture to understand the concepts better, she may try to attempt the questions again. So there isn't any rule to determine when a question might occur again. Take user *15632472* for example, this person attempts the same question over and over again (max of 30 times for a single question) - A rule based system would never ask questions that frequently.\nAlso the RIIID github repo says this:\n```\nThe student can re-do the questions that he already solved before using review system, which is recorded as in_review.\n```\n\nHope this helps :)", "1074564": "Very helpful. Thank you for sharing your observations" }, "source": "meta" }