At a glance
Headline figures for the current selection.Distribution
The two dashed lines mark the median and the average; the key below the chart names each one.
Distribution and average by group
Each group is shown twice: first a response-frequency chart (how many people gave each answer — the survey question is quoted at the top, blanks and N/A excluded), then the average within each group. Every group is shown. A group with fewer than 10 answers is not tested for significance, and its tooltip says so. Beneath each average chart, any differences between groups that are statistically significant at the 95% confidence level are called out in plain language. Where a grouped dimension supports it, you can switch between grouped and fully detailed categories — the charts, averages, and significance recompute together.
Multi-select responses
Single-choice questions show the share of respondents giving each answer; multi-select questions show the share of respondents picking each option (denominator = respondents who answered, so totals can exceed 100%); categorized write-in answers are interleaved with the question they belong to.
What’s linked to the main measure
Read each row as a pattern, not proof. A bar reaching right means people with more of that factor tend to score the main measure higher; a bar reaching left means they tend to score it lower. The longer the bar, the stronger the pattern. "Linked" just means the two tend to move together — it does not prove one causes the other.
What the labels mean. Every link is rank-based and tie-corrected. Clear pattern = the link is statistically significant at the 95% confidence level and at least moderately strong (a standardized strength of 0.30 or more — |Spearman ρ| for ordinal factors, η² for group differences, or a normalized gap for yes/no flags). Likely pattern = significant at 95% confidence but weaker (strength below 0.30) — real, but a smaller effect. No clear pattern = not statistically significant at 95% confidence (it could be chance). No p-values or means are shown; colour reinforces direction.
Key relationships in the data
An automatic scan of the current selection for the strongest, statistically clear relationships. Each is written in plain language; "median" is the typical (middle) value. Adjust the filters above and this updates with them.
Attitudes
Each bar is centred on the middle of the "Neither" response, so a bar sitting further left leans toward disagreement and further right toward agreement — readable at a glance, not just by colour. The share agreeing (somewhat or strongly) is printed at the right. Everything reflects the current filter selection.
More from the survey
The remaining survey questions, so coverage is complete. Each shows the response distribution, and banded questions show an average beneath the title (averages use the midpoint of each band). Hover any bar for the count and percentage.
Additional questions
Single-choice questions show the share of respondents giving each answer; multi-select questions show the share picking each option; hover any bar for the count and percentage.
Categorized open-ended responses
Free-text questions, with every written answer sorted into a recurring theme. Categories were derived from reading all responses; each bar shows how many answers fall in that theme and its share of those who answered. Hover any bar for the count and percentage.