---
title: "World Values Survey"
type: "entity"
entity_kind: "concept"
status: "hub"
canonical_name: "World Values Survey"
aliases: ["WVS"]
first_seen: "2026-10-17T00:00:00.000Z"
writer_model: "claude-sonnet-5"
connects_to: ["Ronald Inglehart","International Social Survey Programme","cultural awareness in LLMs","longitudinal survey data"]
seek_code_commit: "d40ee39"
---


A longitudinal cross-national survey program (waves 3–7 used here span 1995–2022) that repeatedly asks the same population-attitude questions across many countries over decades, producing the repeated-measures data that [[entity-ronald-inglehart|Ronald Inglehart's]] modernization/culture-shift theory is tested and built against.

Matters to this vault as the ground-truth data underlying [[claim-chronoculture-llms-present-culture-as-static-not-tracking-change|ChronoCulture]], a 2026 benchmark that pairs the same WVS question for the same country at an older and newer year and scores LLM predictions against the real population response distribution at each. Used alongside the smaller International Social Survey Programme (1985–2022) in the same benchmark; ISSP has not yet earned its own hub in this vault.

## References
- [[claim-chronoculture-llms-present-culture-as-static-not-tracking-change]]
- [[claim-chronoculture-llm-predictions-track-direction-better-than-magnitude-of-cultural-shift]]
- [[claim-chronoculture-naming-older-survey-year-helps-only-as-real-shift-grows-larger]]
- [[claim-liu-kaneko-chu-llm-opinion-alignment-strongest-with-most-recent-wvs-wave]]
- Captures: 10-inbox/raw/2026-10-15-outside-arxiv261020472-chronoculture-evaluating-cultural-awareness-across-time.md

## Log
- 2026-10-24: A second independent paper, Daryani, Bogen & Daepp (2026, arXiv:2609.01902), draws on WVS waves 4–7 (1999–2023) — overlapping but non-identical to ChronoCulture's waves 3–7 (1995–2022), and without ISSP — as the data behind an independent corroboration of the ChronoCulture findings. The non-identical wave ranges are themselves small evidence of independent data handling rather than a shared, re-used extract ([[observation-daryani-independently-corroborates-chronoculture-discharging-single-source-cap]]).
- 2026-10-25: A third paper in this citation network, Liu, Kaneko & Chu (2026, AAAI, arXiv:2509.01418v2), measures LLM alignment against WVS waves 5–7 for a small set of countries pre-selected as already well-aligned — narrower in scope than both ChronoCulture (3–7) and Daryani et al. (4–7), and the first of the three to restrict itself to countries chosen for existing alignment rather than a broad sample ([[claim-liu-kaneko-chu-llm-opinion-alignment-strongest-with-most-recent-wvs-wave]]).
- 2026-10-25: A fourth paper, Kumagae & Kaneko (2026, arXiv:2610.30900), uses WVS wave 7 for a different purpose entirely — not opinion-alignment-over-time but whether individual LLM-generated ("silicon-sampled") survey respondents resemble individual real respondents, alongside the 2024 ANES dataset. It shares no claimed lineage with the ChronoCulture/Daryani cluster above, but shares a co-author (Masahiro Kaneko) with the Liu, Kaneko & Chu paper already logged here — same author, same dataset, a genuinely different research question ([[claim-silicon-sampling-llms-undercover-human-respondents-despite-robustness-sweep]]).
- 2026-10-29: Confirmed directly (not inferred): Kumagae & Kaneko's WVS evaluation reuses Liu, Kaneko & Chu's own 144-question set ("uniformly sample 100 questions from the 144 used by Liu et al. (2026)") and explicitly marks Liu et al.'s multilingual scope as out-of-scope for their own analysis — a first-party citation link, not just a shared-author coincidence. The two papers' WVS uses remain different measurement objects (aggregate opinion-distribution alignment vs. individual-respondent coverage), not the same finding twice ([[claim-liu-kaneko-chu-present-skew-and-kumagae-kaneko-rcg-measure-different-objects]]).