Dataset
Responses
–
Avg.
–
Positive (4–5)
–
Negative (1–2)
–
Note
small n
more detail
Scored
–
Median
–
Neutral (3)
–
Highlights
% values are based on scored responses (n=–). Data: data/datasets.json + CSVs.

Compare datasets

Distribution change (percentage points)

Positive bars mean the second dataset has a higher share for that score.
Theme overlap
—
Representative quotes
    One-paragraph summary
    —

    Response distribution

    ScoreCount%

    Plain-language summary & quick actions

    —
    Quick actions
      Methods & definitions
      • Scores: 1..5 Likert mapped from text labels (e.g., 1=“Extremely unvalued”, 5=“Extremely valued”).
      • % Positive: share of scored responses that are 4 or 5. % Negative: share that are 1 or 2. Base size shown as n.
      • Small n: if n<30, a “small n” tag appears — interpret cautiously.
      • Meaningful change: Δ≥5 percentage points and two-proportion test p<.05; otherwise shown as small/uncertain.
      • Theme extraction: simple tokenization with stopwords removed; top frequent terms; theme drift = overlap vs new vs dropped terms. Quotes are anonymised.
      • Average/median: reported descriptively; medians are preferred for skewed distributions.

      Copilot analysis (Excel/Teams)

      Use Export to Excel above, open it in Excel/Teams, then paste the copied prompt to get a full AI summary, comparisons, and actions — fully inside Microsoft 365.