Key Findings

• Most respondents report music as emotionally beneficial, but effects vary by genre and listening context.
• After controlling for total listening time (hours/day), several genres show positive or negative associations between genre-specific listening frequency (“dose”) and reporting emotional improvement.
• Overall listening time is not a sufficient explanation for perceived emotional benefit. Some less-listened genres show stronger positive adjusted associations, while heavily listened genres may be neutral or slightly negative in the adjusted model.

Selected Figures

Figure 1 — Listening frequency distribution by genre

This plot contextualizes how often each genre is listened to across the sample.

Figure 2 — Adjusted genre-specific association (controlling for hours/day)

This plot shows the adjusted coefficient for each genre’s listening frequency (“dose”) predicting P(Improve) while holding hours/day constant.

---

How to Run

1. Download the dataset from Kaggle:
   https://www.kaggle.com/datasets/catherinerasgaitis/mxmh-survey-results

2. Place mxmh_survey_results.csv in your preferred directory.

3. Update the DATA_PATH variable in the notebook if needed.

4. Run the notebook top to bottom.

---

Future Directions

Potential extensions of this work include:

• Longitudinal designs tracking within-person changes in listening behavior and emotional state over time.
• Demographic or contextual subgroup analyses where sample sizes permit.
• Incorporating finer-grained musical features (e.g., tempo, familiarity, lyrical content) beyond genre labels.

---

Author Notes

This project is intended as an applied analysis demonstrating thoughtful handling of behavioral and mental health–related data, with emphasis on clarity, statistical restraint, and interpretability.