**Mental_Health_Index varies only slightly by Primary_Platform, with a spread of under 2 points.** I chose both on judgement, not by a statistical test: - **Metric: Mental_Health_Index.** It is the outcome the dataset's decisions revolve around: which platforms to flag as higher-risk, whether to limit use, and whom to prioritise for outreach. - **Category: Primary_Platform.** It is the lever that most directly supports "flag a platform as higher-risk", and it has 7 groups. Several other splits (age, gender, academic level, device) could also be argued for. **Findings (from the Primary_Platform summary table I queried):** - **Highest average:** X (Twitter) at 80.81 (206 students) and YouTube at 80.16 (818 students). - **Lowest average:** Snapchat at 78.99 (472 students), then Reddit at 79.28 and LinkedIn at 79.39. - **Spread:** about 1.8 points between the best and worst platform, on an index averaging roughly 80. That is a small difference. - **Usage hours match the ranking.** Snapchat has the most daily use (5.46 hours) and the lowest mental-health score. X has the least daily use (5.14 hours) and the highest score. That fits a "more time, slightly worse wellbeing" reading. - **Stress points the same way.** Perceived stress is lowest on X (12.46) and highest on LinkedIn (14.06), with Snapchat close behind (13.97). - **Sleep and GPA barely move.** Average sleep ranges from 6.64 to 6.81 hours, and GPA is about 3.43 on every platform. **Caveats:** - The platform differences are modest, so platform alone is a weak signal for flagging risk. - LinkedIn (96 students) and X (206) are small groups, so their averages are less stable than those for Instagram (1,453) and TikTok (1,226). - The data is observational, so the pattern shows association, not that a platform causes lower mental health.