Knowledge and work — Examine the “mismatch” between fields studied and fields important at work (by prefecture)
Compare the share of people whose fields studied and fields important at work do not coincide, by prefecture. An attempt to understand diversity through this “mismatch”.
| Prefecture | Share with different fields | Diversity effective fields | Respondents (different fields) | Unpin |
|---|
Compare distributions in the two pinned prefecturesPin exactly two prefectures to compare.
How to read this chart and its data
The X-axis is the share of people whose field studied and first-ranked work field differ. The denominator includes respondents naming specific fields for both; responses including Other or None are excluded. Identical fields require an exact match of the original field codes.
The Y-axis is a Hill number calculated from the composition of first-ranked work fields among people whose fields differ. Work responses are pooled across fields studied. It expresses the equivalent number of equally common fields, not the number of occupations, vacancies or educational choices.
- q = 0: number of observed fields. Rare fields receive equal weight, making this measure sensitive to sample size.
- q = 1: Shannon effective number of fields, calculated as exp(−Σp log p), where p is each field’s share.
- q = 2: inverse Simpson index, 1 / Σp².
- q = 3 or 4: (Σp^q)^(1/(1−q)). Higher q gives more weight to common work fields.
The Y-axis adjusts automatically for each q by default. To compare absolute differences across q, enable the shared Y-axis scale. Pinning prefectures does not change that scale.
Pin up to four prefectures. To examine two distributions in detail, pin exactly two and follow the comparison link. Settings are recorded in the URL and can be bookmarked.
The 2025 all-age sample across 47 prefectures contains 51,192 plotted respondents, including 38,254 with different fields. Another 13,335 responses are excluded. No adjustment for fields studied or age composition, population expansion, or sample-size correction is applied. Regional differences should not be interpreted as causal effects of opportunities offered by a region.