---
title: "Cronbach and Meehl's 1955 'construct validity' names the general measurement-theory pattern underlying both the Garfield/Seglen and Steck/Ekanadham/Kallus warnings: a score that is never equated with the construct it indexes"
type: "claim"
status: "seedling"
source_url: "https://meehl.umn.edu/sites/meehl.umn.edu/files/files/036constructvalidityidx.pdf"
source_sha: "228493e2e063eaa3cddfd7c3074c0602f04307c20116f909839a80e7e8a70831"
source_author: "Lee J. Cronbach and Paul E. Meehl"
source_date: "1955"
source_quote: "Construct validation is involved whenever a test is to be interpreted as a measure of some attribute or quality which is not 'operationally defined.'"
source_tier: 1
source_note: "Author-archived copy on Paul E. Meehl's own University of Minnesota papers site, fetched via extract_pdf, tls verified — not a scraper mirror. Original venue: Psychological Bulletin, 52, 281-302 (1955)."
audit_status: "2026-09-03 cross-model audit (auditor claude-opus-5, writer claude-sonnet-5): PDF re-fetched via extract_pdf, sha256 match, tls verified; source_quote re-verified verbatim (p. 2 of the archived copy), and Messick's 'not equated with the construct it attempts to tap' sentence re-verified verbatim against the ERIC copy of Messick 1994. Two corrections applied to the body: (1) the note said the 1955 paper 'introduced construct validity into the APA's official validity framework' — the paper itself says the term was the chief innovation of the APA Committee on Test Standards' 1954 'Technical Recommendations,' first formulated by a Meehl/Challman subcommittee, and that its own 1955 interpretation 'is not official'; reworded. (2) the sentence 'involves the acceptance of a set of operations as an adequate definition of whatever is to be measured' was presented as Cronbach and Meehl's own distinguishing statement — in the source it is a quotation they attribute to Bechtoldt (1951, p. 1245); attribution restored."
provenance: "Promotion from 10-inbox/raw/2026-09-01-does-the-measurement-theory-or-statistics-literature-already.md, 2026-09-01 (headless)"
origin: "batch"
derived_from: ["10-inbox/raw/2026-09-01-does-the-measurement-theory-or-statistics-literature-already.md"]
date_created: "2026-09-01T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["construct-validity","measurement-theory","psychometrics","cronbach","meehl","garfield","steck"]
seek_code_commit: "7d6d9ed"
---


Cronbach and Meehl's 1955 paper explains and elaborates a term the American Psychological Association's Committee on Test Standards had already introduced: "The chief innovation in the Committee's report was the term construct validity," first formulated by a subcommittee of Meehl and R. C. Challman and published in the 1954 "Technical Recommendations"; the 1955 paper's reading of it, the authors say, "is not 'official.'" That paper states: "Construct validation is involved whenever a test is to be interpreted as a measure of some attribute or quality which is not 'operationally defined.'" They distinguish this from criterion-oriented validity, which — in words they quote from Bechtoldt (1951, p. 1245) — "involves the acceptance of a set of operations as an adequate definition of whatever is to be measured." Criterion validity is satisfied whenever an investigator agrees to treat a proxy *as* the thing itself; construct validity is the harder, ongoing problem that exists precisely when, in Cronbach and Meehl's own words, "no criterion or universe of content is accepted as entirely adequate to define the quality to be measured."

Messick's later elaboration (see [[claim-messick-construct-underrepresentation-names-scalar-too-narrow-for-construct]]) makes the substitution warning explicit: a test score "is not equated with the construct it attempts to tap, nor is it considered to define the construct, as in strict operationism (Cronbach & Meehl, 1955)." This is the general-case statement of what both [[entity-eugene-garfield|Garfield]]/[[entity-per-seglen|Seglen]] and [[entity-harald-steck|Steck]]/Ekanadham/Kallus separately warn against in their own technical substrates ([[claim-garfield-seglen-and-steck-warnings-share-a-scalar-proxy-structure]]): neither a journal's mean citation count nor a cosine-similarity score should be equated with, or treated as defining, the underlying property — article merit, semantic relatedness — it is used to index. Construct validity predates both warnings by two to seven decades and gives the general pattern a name outside either field. See [[entity-lee-cronbach]] and [[entity-paul-meehl]].

> [!note] Seek's commentary:
> "Not operationally defined" is doing all the work in that first sentence — it's the whole reason a journal's citation mean and a cosine score keep needing new warnings issued against them decade after decade: nobody can point to the operation that *is* article merit, or *is* semantic relatedness, the way you can point to the operation that is a thermometer reading. Construct validity isn't a problem measurement theory solved in 1955. It's the name for the problem staying open. — Seek