City-Type Sampling / 城市类型抽样

Definition

City-type sampling uses quantitative rankings and administrative knowledge to select urban cases, then interprets each case within a bounded regional, historical, industrial, and geographic type instead of treating one middle-ranked city as a national average.

Current Synthesis

城市就是这样22 develops the method through a failed premise. David Fishman initially chose 临汾 because it sat near the middle of a 337-city list, but fieldwork showed that rank position could not erase north-south variation, resource history, city-county differences, or rapid change. The revised method keeps rankings as a way to narrow the search while moving representativeness claims to plural types such as northern post-industrial cities, Central Plains cities, and southwestern minority mountain cities.

Key Claims

• A list median is a property of selected indicators and weights, not proof that the selected city represents a national population.
• Regional geography, industrial history, administrative structure, and migration patterns define more useful comparison strata.
• The relevant observation scale may be below the prefecture-level city because central districts and outer counties can follow different trajectories.
• Static comparisons need longitudinal checks when infrastructure, population settlement, environment, and consumer services change quickly.
• Quantitative selection becomes more informative when paired with interviews about concrete household choices.

Evidence

• Rank-to-type correction: 城市就是这样22 shows Fishman revising Linfen from “China’s median city” to a northern resource-dependent, post-industrial case.
• Within-city heterogeneity: 城市就是这样22 contrasts more pessimistic interviews in Linfen’s old urban area with more optimistic accounts in some outer counties.
• Temporal qualification: 城市就是这样22 uses a Jishou new-district revisit and Kangbashi comparison to show why a single empty-looking moment cannot settle whether a development is a permanent “ghost city.”

Counterevidence & Qualifications

The proposed city types are analytical categories, not demonstrated statistical strata, and the episode does not specify a reproducible classification model. Small, convenience-based interview samples cannot estimate prevalence within a type. Rankings may still be useful for explicit indicator comparisons, while field impressions can introduce their own selection and observer effects.

What Changed

• Replaced a single national median-city proxy with plural regional and historical comparison types.
• Added district and county heterogeneity to the unit-of-analysis decision.
• Added repeated observation as a requirement for claims about fast-changing urban areas.

Related Concepts

• Urban Field Interviewing / 城市田野访谈 - supplies qualitative tests of categories selected by city-level data.
• Fieldwork As Knowledge Method / 田野作为知识方法 - broader practice of testing abstractions through contact with places and people.
• Observation Before Inference - guards against converting a rank, empty image, or short visit directly into a general conclusion.
• City Commercial Observation - adjacent practice of using grounded city cases to explain commercial and industrial systems.