---
id: "20260809-0219-does-heath-larrick-wu"
title: "Does Heath, Larrick & Wu (1999) 'Goals as Reference Points' itself state the below-goal-works-harder mechanism the vault recorded?"
type: "capture"
status: "promoted"
origin: "batch"
writer_model: "claude-sonnet-5"
date_created: "2026-08-09T00:00:00.000Z"
provenance: "batch web-research run, 2026-08-09, resolving [[question-verify-heath-larrick-wu-1999-goals-as-reference-points-primary]]"
derived_from: []
tags: ["behavioral-economics","loss-aversion","reference-points","motivation","prospect-theory","primary-source-verification"]
source_urls: ["https://faculty.chicagobooth.edu/george.wu/research/abstracts.html","https://faculty.chicagobooth.edu/george.wu/teaching/public/38913/class7.htm","https://sites.duke.edu/larrick/","https://api.semanticscholar.org/graph/v1/paper/787648e203428f1e3839e6656677eaf371f78d17?fields=title,openAccessPdf,externalIds"]
source_authors: ["George Wu","George Wu (course co-taught with Reid Hastie)","Richard P. Larrick","Semantic Scholar (Allen Institute for AI)"]
source_dates: ["undated faculty page, accessed 2026-08-09","Fall 2016 course page, accessed 2026-08-09","undated faculty page (footer copyright 2026), accessed 2026-08-09","accessed 2026-08-09"]
source_titles: ["George Wu abstracts","BA 38913 Class 7","Rick Larrick – research and teaching","Semantic Scholar Graph API — paper record 787648e203428f1e3839e6656677eaf371f78d17"]
source_venues: ["George Wu's faculty site, University of Chicago Booth School of Business","Business 38913: Foundations of Judgment and Decision Making (course site), University of Chicago Booth School of Business","Rick Larrick's faculty site, Duke University Fuqua School of Business","Semantic Scholar API (semanticscholar.org)"]
source_tiers: ["Tier 1 — author's own faculty site, reproducing the primary paper's own abstract.","Tier 1 — co-author's own teaching site, direct scare-quoted restatement of the paper's own phrase.","Tier 1 — co-author's own site; used for bibliographic confirmation only, no full text.","Tier 4 — aggregator/metadata index; used only for a negative finding (no open-access copy indexed), not as evidence for a positive claim."]
source_shas: ["0c562b114fd78f9a103cacba71ef3acc4c06bb26d86939bba381350f5696a70a","559872da7db18805d16783e3eacc2b097a25c2b6eb32194334abe4d3fdf8cf1a","d93b86f8672251b2455e5045f7c74c3da53781e5da0efb55960b687d958edc9d","fb323e8efb8e34767796f81aeb1589efa5df6626663dd5979a43e0d04f7a95ee"]
promoted_to: ["30-notes/claim-heath-larrick-wu-1999-abstract-states-reference-point-architecture.md","30-notes/claim-goals-as-reference-points-below-goal-work-harder.md (revised in place — audit_status updated, status unchanged at seedling)","40-entities/entity-george-wu.md","40-entities/entity-prospect-theory.md (updated — new dated line + reference added, not created)"]
not_promoted: ["The 'exact below-goal-works-harder formulation could not be verified' finding — not an atomic claim in the vault's sense (it's an absence of confirmation, not a fact); folded into the progress line appended to [[question-verify-heath-larrick-wu-1999-goals-as-reference-points-primary]] and into the revised existing claim-note's audit_status, rather than written as a standalone note.","The second corroborating quote (George Wu's course-page comprehension question dividing loss-aversion vs. diminishing-sensitivity work across effort/attention/persistence) — not atomic on its own (same underlying architecture claim, second document); folded into the new claim-note as corroborating detail rather than split into its own note.","Daniel Kahneman (entity candidate) — declined as a hub. Real, foundational to the domain, but named here only via citation inside HLW1999's abstract, not from any Tier-1 source about Kahneman's own words fetched this session; recurs across 8+ existing vault notes with no hub yet, which is a real gap, but promoting it opportunistically off a citation-of-a-citation would set a low bar. Consistent with the 2026-07-24 Berger-Pope capture, which named Kahneman/Tversky in the same way and also declined to promote them. Left as a mention; flagged below for a dedicated future pass.","Amos Tversky (entity candidate) — declined, same reasoning as Kahneman.","Chip Heath (entity candidate) — declined as a hub. Real co-author and first-named author of the foundational paper, but no Tier-1 source in his own words was fetched this session (his site was not in the source list); 'first-named co-author of a paper we read the abstract of' is thinner than the bar George Wu cleared. Left as a mention on the new claim-note and on entity-george-wu.","Richard P. Larrick (entity candidate) — declined as a hub, for the same reason as Heath: his faculty site was fetched (Tier 1) but only for bibliographic confirmation, no quotable text in his own words. Left as a mention.","Wu, Heath & Larrick (2004) unpublished manuscript; Larrick, Heath & Wu (2009) Social Cognition; Jeffrey, Onay & Larrick (2010) JBDM; Markle, Wu, White & Sackett (2018) JRU; Allen, Dechow, Pope & Wu (2017) Management Science — all 'further leads,' none read this session; not load-bearing doubts under a kept claim, so not routed to 50-questions/ per intake discipline. Left as leads in this capture's own body for a future session to chase."]
seek_code_commit: "46c6736"
---


Raised by [[question-verify-heath-larrick-wu-1999-goals-as-reference-points-primary]], which flagged that [[claim-goals-as-reference-points-below-goal-work-harder]] quotes the "people slightly below their goal work harder" mechanism from a 2023 replication paper's introduction *citing* Heath, Larrick & Wu (1999), not from the 1999 primary itself. This capture attempts a direct read of the primary.

**Access outcome, stated up front:** Heath, C., Larrick, R. P. & Wu, G. (1999), "Goals as Reference Points," *Cognitive Psychology* 38(1):79–109, DOI 10.1006/cogp.1998.0708, is paywalled. ScienceDirect (the publisher's own venue, https://www.sciencedirect.com/science/article/abs/pii/S0010028598907083) returned HTTP 403 Forbidden. PsycNet (https://psycnet.apa.org/record/1999-10901-003) returned HTTP 403 Forbidden. A CiteSeerX mirror link returned HTTP 404 Not Found. Neither co-author's faculty site (George Wu at Chicago Booth, Richard Larrick at Duke Fuqua, both checked directly) hosts a full-text PDF. The Semantic Scholar API record for this paper reports `"openAccessPdf": {"url": "", "status": "CLOSED", "license": null}` — confirming no open-access copy is indexed anywhere Semantic Scholar crawls. No pirate/mirror site was used to work around this, per source-venue discipline. The full primary text was not read in this pass.

What follows is therefore built from the *author's own* secondary restatements of the paper — Tier 1 by the rubric (the person speaking, on their own venue, about their own work), but not the primary document itself, and not a substitute for reading its effort/persistence results section directly.

## Claim: HLW1999's own abstract states the reference-point-plus-loss-aversion architecture, in the paper's own words

**Claim type:** technical-mechanism (how the paper's proposed mechanism works) → floor requires Tier 1–2. Source is Tier 1 (co-author's own site, reproducing the paper's own abstract). Clears floor.

George Wu's faculty page at Chicago Booth reproduces the abstract of the 1999 paper under his own "Abstracts of Publications" listing:

> "We argue that goals serve as reference points and alter outcomes in a manner consistent with the value function of Prospect Theory (Kahneman & Tversky, 1979; Tversky & Kahneman, 1992). We present new evidence that goals inherit the properties of the value function-not only a reference point, but also loss aversion and diminishing sensitivity. We also use the value function to explain previous empirical results in the goal literature on affect, effort, persistence, and performance."

This confirms, in language traceable to the paper itself rather than a downstream restatement, that (a) the paper explicitly models goals as prospect-theory reference points; (b) it claims the goal-as-reference-point inherits *both* loss aversion and diminishing sensitivity from the [[entity-prospect-theory|prospect-theory]] value function; and (c) it explicitly claims to use this account to explain prior findings on "affect, effort, persistence, and performance" in the goal-setting literature.

## Claim: the paper explicitly names effort as one of the phenomena its reference-point account explains

**Claim type:** definitional/historical (what the paper claims to explain, not a specific number or a worked mechanism detail) → floor Tier 3–4 acceptable, but source here is Tier 1, exceeding the floor.

The same abstract sentence above states the value function is used "to explain previous empirical results in the goal literature on affect, effort, persistence, and performance." This establishes that effort is, in the paper's own stated scope, one of the outcomes the goals-as-reference-points account is built to explain — consistent with the vault's existing claim that effort is the mechanism's target phenomenon. It does not, by itself, establish the specific *direction and magnitude* claim (harder work specifically below the goal, easing off above it) in the paper's own words — see next claim.

## Claim: the exact "below-goal-works-harder" formulation could not be verified against the primary's own text

**Claim type:** technical-mechanism, specific (the precise shape of the effort-vs-goal-position relationship) → floor requires Tier 1–2, and no Tier 1–2 source with the primary's own wording on this specific point was obtained in this pass.

The vault's [[claim-goals-as-reference-points-below-goal-work-harder]] currently rests the specific sentence — "people who are slightly below their goal work harder than those who have reached or exceeded it, in a manner consistent with loss aversion" — on Klein Teeselink, van den Assem & van Dolder (2023), a Tier 1 source that is itself *citing* Heath, Larrick & Wu (1999) rather than being the primary. Despite genuine search across the publisher's site, PsycNet, CiteSeerX, both living co-authors' own faculty sites, and a Semantic Scholar open-access check, the 1999 paper's own effort/persistence results section was not accessible in this session. `[unverified — could not confirm or deny the exact below-goal-works-harder wording against the primary after search]`. The architecture the claim rests on (reference point + loss aversion, explicitly linked to effort as an explained phenomenon) is independently confirmed at Tier 1 by the claim above; what remains unconfirmed is only whether HLW1999 phrases the effort prediction in the specific below/above-goal comparative form the vault currently quotes secondhand.

## Claim: a second author-site document independently corroborates the loss-aversion-explains-effort framing, in language close to the primary's own

**Claim type:** technical-mechanism, corroborating → floor Tier 1–2. Source: Tier 1, a different page on a co-author's own site than the abstract above, giving mild independent triangulation (same author, same institutional venue, different specific document).

George Wu's own graduate-course page for "Foundations of Judgment and Decision Making" (Fall 2016, co-taught with Reid Hastie) lists the paper in a reading list and glosses it in his own words, using a scare-quoted phrase that matches the abstract almost verbatim:

> "This paper interprets goal setting literature in terms of the value function. The goal acts as a reference point, and 'goals "inherit" the properties of the value function.' More generally, this paper suggests the role non-status quo reference points may have on behavior."

The same course page's comprehension question 6 frames the paper's explanatory division explicitly: "The major stylized facts of goal setting can be summed up as goals increase performance by increasing effort, attention, and persistence. How does the 'goals as reference points' approach organize these findings? What part of the work does loss aversion do? What part of the work does diminishing sensitivity do?" This is Wu's own pedagogical framing (not a quote from the 1999 paper), but it corroborates — from the same author, independently of the abstract — that the paper assigns specific explanatory work to loss aversion versus diminishing sensitivity across effort, attention, and persistence, rather than treating them as interchangeable.

## Overall verdict on the core question

Partially confirmed, with a specific residual gap named rather than papered over. Heath, Larrick & Wu (1999) does, by the paper's own abstract (Tier 1, author-site copy), explicitly build a goals-as-reference-points account on the prospect-theory value function, explicitly claim loss aversion and diminishing sensitivity as inherited properties, and explicitly claim to explain effort (among other outcomes) using that account. What could not be confirmed in this pass, after genuine search, is whether the *specific comparative sentence* the vault currently quotes — below-goal effort exceeding at/above-goal effort — appears in those words in the primary itself, versus being the 2023 replication paper's own restatement of a result the 1999 paper reports differently (e.g., via a figure, a regression coefficient, or different phrasing). The primary remains paywalled with no open-access route found. `[unverified — could not confirm or deny the exact below-goal wording after search; the mechanism architecture is confirmed at Tier 1, the specific comparative phrasing is not]`

> [!note] Seek's commentary:
> This is a case where "the source cited by the source" turns out to be more available than the source itself — both co-authors keep abstracts and syllabi on their own sites, but not the paper. The abstract alone is enough to confirm the vault isn't hallucinating a mechanism that isn't there — the architecture is real, in the authors' own words. But it's not enough to confirm the exact sentence, and treating "author's own abstract" as equivalent to "read the results section" would be exactly the kind of silent upgrade the sourcing floor exists to prevent. The honest outcome is: closer than before, not closed.

## Further leads

- Wu, Heath & Larrick (2004), "A prospect theory model of goal behavior" — unpublished manuscript listed on George Wu's abstracts page, described as a formal model of the same mechanism; worth chasing as a working-paper PDF if one surfaces, since unpublished manuscripts are sometimes hosted even when the published version isn't.
- Larrick, Heath & Wu (2009), "Goal-Induced Risk Taking in Negotiation and Decision Making," *Social Cognition* 27:342-364 — same three authors, same reference-point mechanism applied to risk-taking rather than effort; abstract on George Wu's site states goals create "a region of perceived losses for outcomes below a goal," which is closer to the vault's phrasing than the 1999 abstract itself, though it's a 2009 companion paper, not the 1999 primary.
- Jeffrey, Onay & Larrick (2010), "Goal attainment as a resource: the cushion effect in risky choice above a goal," *Journal of Behavioral Decision Making* 23:191-202 — listed on Larrick's site, extends the same value-function account.
- Markle, Wu, White & Sackett (2018), "Goals as Reference Points in Marathon Running: A Novel Test of Reference-Dependence," *Journal of Risk and Uncertainty* 56:1-32 — a large-field-study test of the same theory, abstract on George Wu's site.
- Allen, Dechow, Pope & Wu (2017), "Reference-Dependent Preferences: Evidence from Marathon Runners," *Management Science* 63:1657-1672 — related field evidence for round-number reference points, not goals specifically.
- If institutional access to ScienceDirect (DOI 10.1006/cogp.1998.0708) becomes available, that closes this question directly — the abstract and results section would give the exact primary wording this capture couldn't reach.

## Entity candidates

- Daniel Kahneman — person — with Tversky, originated the prospect-theory value function (1979) that HLW1999 explicitly imports; this is the foundational ancestry the whole paper is built on, named first in its own abstract.
- Amos Tversky — person — co-originator of prospect theory (1979) and cumulative prospect theory (1992), jointly cited alongside Kahneman as the value-function source HLW1999 "inherits" from.
- George Wu — person — co-author; the only one of the three whose own site yielded quotable Tier 1 text in this pass (abstracts page + course page).
- Chip Heath — person — co-author (Stanford GSB); first-named author of the 1999 paper.
- Richard P. Larrick — person — co-author (Duke Fuqua); maintains a full publication list with this entry but no hosted PDF.
- [[entity-prospect-theory]] — concept — already in the vault; the theoretical framework this paper claims goals "inherit" properties from.

## Safety flags

None. All sources fetched this session were ordinary academic faculty pages, course syllabi, and API metadata — no addressed-to-AI language, override language, claimed authority, tier self-assignment, file-system instructions, credential requests, or urgency framing encountered.