| Dim | Dimension | Score | Weight | Flag |
|---|---|---|---|---|
| D1 | Definitional Precision | 5 | 12% | Caste and ethnicity conflated in survey instrument |
| D2 | Classification Rigor | 3 | 18% | Self-administered — no quality control documented |
| D3 | Case Capture & Sampling | 2 | 15% | ⚑ Score ceiling applies — sampling cannot support population claims |
| D4 | Coverage Symmetry | 5 | 15% | — |
| D5 | Source Independence | 6 | 10% | — |
| D6 | Verification Standards | 3 | 18% | Data under organizational control — no external access process |
| D7 | Transparency & Governance | 5 | 5% | — |
| D8 | Counter-Evidence | 2 | 7% | No limitations section — critics characterized as motivated |
| Composite Score | 3.8 | Advocacy-Grade | ||
Scope: Respondents were recruited through an advocacy-controlled website and social media, not through an independent sampling frame. This means the pool of respondents is not representative of South Asian Americans — it overrepresents people who have experienced caste discrimination, which is the population most likely to engage with advocacy platforms. A scoring rule prevents reports with this sampling structure from exceeding 5.9.
Scope: Report contains no limitations section. Methodological critiques in peer literature and from Carnegie (fn. 29) are not engaged substantively in the report or in published responses.
Scope: No interviewer training, pilot testing, or quality control protocol documented for self-administered online instrument.
Scope: Caste discrimination not operationally distinguished from ethnic or racial discrimination in the survey items. This conflation makes the 67% figure difficult to interpret even within the self-selected sample.
Scope: Raw microdata is not available and no formal research data request process is documented — in contrast to Pew, which maintains an archive request process, or the Roper Center, which provides structured access.
Caste and ethnicity conflated in survey instrument
'Caste discrimination' not operationally distinguished from ethnic or racial discrimination in question design.
Self-administered — no quality control documented
Survey delivered via advocacy website with no documented interviewer training, pilot testing, or quality control protocols.
⚑ Score ceiling applies — sampling cannot support population claims
No external sampling frame. Distribution via castesurvey.org and advocacy social media. Respondent pool systematically overrepresents discrimination-affected individuals. A scoring rule prevents non-representative sampling from reaching Deficient or above.
Asks about personal experiences symmetrically but does not survey non-Dalit respondents about their behavior.
Equality Labs is the sole producer. No co-authors or independent review documented.
Data under organizational control — no external access process
Raw data not released. No replication possible. No formal data request process documented.
Organizational mission disclosed. Funding sources not fully enumerated in report.
No limitations section — critics characterized as motivated
Report contains no limitations section. Published responses to methodological critiques characterize them as motivated by caste privilege rather than engaging them substantively.
How this report's findings have been cited or applied after publication. Severity reflects the gap between what the report establishes and how it was represented.
Claimed scope: Population prevalence of caste discrimination among South Asian Americans in California and nationally
Established scope: Self-selected respondents who engaged with advocacy networks; findings cannot be generalized to population
Legislative testimony cited the 67% figure as establishing prevalence of caste discrimination in the South Asian American population. The figure describes the self-selected sample, not the population. The distinction was not disclosed in testimony.
Claimed scope: Documented workplace caste discrimination prevalence in South Asian American professional communities
Established scope: Survey of advocacy-network-recruited respondents; workplace subsample not reported separately
Multiple corporations adopted anti-caste HR policies citing this survey as evidentiary basis. The methodological limitations were not disclosed in the policy rationales reviewed.
Scope: Explicitly characterized as limited by sampling methodology — the most accurate citation in the corpus
Carnegie's footnote 29 explicitly documents the sampling limitations. This is the only major citation that accurately represents what the survey can and cannot establish. It is the methodological landmark citation in the corpus.