{
  "id": 591201,
  "title": "Do Demographics Really Matter?",
  "url": "/competitions/cmi-detect-behavior-with-sensor-data/discussion/591201",
  "author_name": "manoj akondi",
  "post_date": "2025-07-26T05:58:14.324000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Do a subject's demographic details matter? In my experiments, I noticed a decent difference in macro F1 (collapsed classes) across various demographic cuts such as sex, handedness, and height. I was curious if anyone else has observed this. I also tried including these features as inputs, but didn’t see any improvement.</p>",
  "messages": [
    {
      "id": 3254261,
      "postDate": "2025-07-26T05:58:14.323Z",
      "content": "<p>Do a subject's demographic details matter? In my experiments, I noticed a decent difference in macro F1 (collapsed classes) across various demographic cuts such as sex, handedness, and height. I was curious if anyone else has observed this. I also tried including these features as inputs, but didn’t see any improvement.</p>",
      "rawMarkdown": "Do a subject's demographic details matter? In my experiments, I noticed a decent difference in macro F1 (collapsed classes) across various demographic cuts such as sex, handedness, and height. I was curious if anyone else has observed this. I also tried including these features as inputs, but didn’t see any improvement."
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "3254261": "Do a subject's demographic details matter? In my experiments, I noticed a decent difference in macro F1 (collapsed classes) across various demographic cuts such as sex, handedness, and height. I was curious if anyone else has observed this. I also tried including these features as inputs, but didn’t see any improvement."
  }
}