Surveys and questionnaires

A wellbeing survey: screen time, sleep and stress

Written by Provetale's pipeline from this data, then checked by code.

Source

Synthetic data generated by Provetale for this sample (a fixed random seed, so it can be rebuilt exactly). It describes no real people.

People who answer 'Always' to social comparison spend 5 daily hours on social media, while those who answer 'Never' spend 1.5 hours.

Check
VERIFIED. Summary VERIFIED.
Audience in the file
GENERAL
Rows in the saved run
1,800
Claims kept
23
Claims dropped
1
Model named in the file
mistral-large-latest
Snapshot date
2026-10-28

Summary

  • Fact

    People who answer 'Always' to social comparison spend 5 daily hours on social media, while those who answer 'Never' spend 1.5 hours.

  • Interpretation

    The 'Always' group spends more than three times as much time on social media as the 'Never' group.

  • Fact

    People who say social media has a negative overall impact have a stress score of 24.1, while those who say it is beneficial have a score of 11.4.

  • Interpretation

    The stress score of the 'Negative' group is more than twice that of the 'Beneficial' group.

Key figures

The totals in your file, before any pattern

  • Fact

    The file holds 1,800 rows.

  • Fact

    The largest groups by primary platform are YouTube (26%), Instagram (26%), and TikTok (18%).

Hidden signals

Hidden signals

  • Fact

    61% of TikTok users answer 'Often' or 'Always' to social comparison.

  • Fact

    Only 9% of Facebook users do.

  • Interpretation

    The comparison is based on the rank of answers across platforms, but it has not been tested to rule out chance.

  • Fact

    47% of people aged 18-24 answer 'Often' or 'Always' to social comparison.

  • Fact

    Only 12% of those aged 55 and over do.

  • Interpretation

    The comparison is based on the rank of answers across age groups, but it has not been tested to rule out chance.

  • Fact

    Answers to Social comparison differ by Overall impact: Negative highest, Beneficial lowest. 48% of Negative answer Often / Always, against 16% of Beneficial (compared by rank across 3 groups).

Segments and concentration

Groups and links

  • Fact

    Always averages 3.4× Never on Daily hours. Average Daily hours: Always 5, Never 1.5, across 5 groups.

  • Fact

    Negative averages 2.1× Beneficial on Stress score. Average Stress score: Negative 24.1, Beneficial 11.4, across 3 groups.

  • Fact

    Stress score and Daily hours rise and fall together. A strong relationship across 1,771 rows. It shows only that the two move together.

  • Fact

    Always averages 2.1× Never on Stress score. Average Stress score: Always 24.2, Never 11.4, across 5 groups.

  • Fact

    Sleep quality = 1 averages 2.1× 5 on Stress score. Average Stress score: Sleep quality = 1 23.2, 5 11.2, across 5 groups.

What your file contains

Facts about the file itself, not patterns

  • Fact

    The file has 1 column with data-quality issues: Stress score (29 missing values).

  • Fact

    No clear difference was found in 9 of the 41 tests comparing pairs of columns, including Late night use and Age group, and Late night use and Primary platform.

  • Interpretation

    This does not show that the groups are the same: a smaller difference than this file can detect may still exist.

  • Fact

    34% of rows have Late night use set to yes.

  • Fact

    35% of answers to social comparison are 'Never' or 'Rarely', while 30% are 'Often' or 'Always'. The middle answer is 'Sometimes'.

What was examined

  • 41 tests of relationships between columns were run; 31 met the thresholds to be reported.
  • Change over time: not part of this report (the data has no date column).
  • What moved a change: not part of this report (the data has no date column).
  • The file has 1,800 rows and 0 exact duplicate rows. Columns needing attention: Stress score (29 missing).
  • 1 drafted statement was left out because it did not match its evidence; no required section was lost.
  • 8 of the 16 findings that met the bar are in this report, plus 5 facts about the file itself in "What your file contains".
  • 6 of the findings are described in their own words, written by code from the evidence, because the written report did not describe them.
  • How to read this report: a finding appears only when it is strong and reliable enough, weaker results are left out, and every number comes from your file.
  • What “verified” means here: every number, group and date in a statement was checked by code against the evidence the statement cites. It does not prove that the conclusion holds beyond that evidence.

Download wellbeing-survey.csv

Evidence

This saved file has no trace for that evidence.