Priming Cooperation in Social Dilemma Games

Priming Cooperation in Social Dilemma Games Drouvelis Metcalfe Powdthavee

Tag layer of the DHSS Meta-Tagging Project. The article text is reproduced from the staged copy named in sourceDesc and is not altered; the tags are held separately in standOff so that overlapping and nested spans are preserved.

Priming Cooperation in Social Dilemma Games 2010 10.2139/ssrn.1631098 drouvelis2010-priming-cooperation-in-social-dile drouvelis2010-priming-cooperation-in-social-dile.txt, the project's staged copy 1396213eca2a27155303f6e39150627373aa5cf2 open-access experimental-economics

the paper's own words

the tag layer; a sentence written by the tagger about the paper, always anchored to a quotation but not itself a quotation

0 spans located by searching the stored text for the quote's exact words, allowing any whitespace between them existence Quotation withheld because this record is not cleared for redistribution. priming of cooperation in a public goods game Quotation withheld because this record is not cleared for redistribution. behavioural-experiment behavioural-experiment Quotation withheld because this record is not cleared for redistribution. mechanisms Priming works in two steps: the cooperation words first raise women's beliefs about how much the other group members will give, and those raised beliefs then drive their own contribution decisions Quotation withheld because this record is not cleared for redistribution. mechanisms The primes serve as anchors or focal points on which subjects can coordinate, which lowers the strategic risk of contributing and so releases more generous behaviour, particularly among women Quotation withheld because this record is not cleared for redistribution. mechanisms The word-search puzzle acts as a subconscious reminder of cooperation, raising the accessibility of the concept at the moment of choice Quotation withheld because this record is not cleared for redistribution. The paper says nothing about long-term or repeated-play effects of priming, or about applicability beyond public goods games. Quotation withheld because this record is not cleared for redistribution. Because a one-shot design was used, whether the priming effect persists with repetition is left open. Quotation withheld because this record is not cleared for redistribution. Beliefs were elicited without incentives, so belief reports are not incentive-compatible. Quotation withheld because this record is not cleared for redistribution. the word-search priming procedure itself, lifted from psychology experiments and applied inside an incentivised economic game where money is at stake Quotation withheld because this record is not cleared for redistribution. random allocation of participants to the two treatments before the experiment was run eliminates: that a difference between treatments is a difference between the people who chose them · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. priming priming Quotation withheld because this record is not cleared for redistribution. framing framing Quotation withheld because this record is not cleared for redistribution. contribution levels contribution levels Quotation withheld because this record is not cleared for redistribution. beliefs beliefs Quotation withheld because this record is not cleared for redistribution. gender gender Quotation withheld because this record is not cleared for redistribution. emotional responses emotional responses Quotation withheld because this record is not cleared for redistribution. revealed preferences revealed preferences Quotation withheld because this record is not cleared for redistribution. contribution decisions contribution decisions Quotation withheld because this record is not cleared for redistribution. priming works through a shift in females’ beliefs (this paper, IZA DP 4963, May 2010) → priming does not affect beliefs and works by changing preferences (the same authors' published version, Theory and Decision 2015, in this corpus as drouvelis2015-cooperation-priming) SAME AUTHORS, SAME EXPERIMENT, OPPOSITE MECHANISM - AND THE SCHEMA HAS NO TYPE FOR IT. The 2015 paper reports two waves, the first run in 2009 with 51 and 54 subjects, which is this paper's N. So this working paper is wave one and the publication is wave one plus wave two. In wave one the raw contribution difference is NOT significant (Wilcoxon two-sided p = 0.249) and the effect appears only with a gender control at the 10% level, carried by a four-cell split of 105 subjects (19, 25, 32, 29); the mechanism is stated as two steps through female beliefs. In the pooled publication the raw difference IS significant (p = 0.044 one-sided), there is `insufficient evidence to suggest that the effect of priming varies significantly across ... gender`, and `priming does not affect subjects' beliefs`, which is precisely what licenses its claim that priming works through PREFERENCES. So the whole two-step belief mechanism, and the gender effect that motivated it, are absent from the published version of the same experiment. Neither paper is being called wrong here: more data arrived and a subgroup result did not survive it. What the corpus can record is that both statements exist, that they are by the same authors about the same data plus more, and where each one is. `superseded-by` / `revised-in` is the type this needs and the vocabulary does not have it - on the schema's missing relation types. Quotation withheld because this record is not cleared for redistribution. priming - the same formal definition this group uses in the 2015 publication, taken from Hertel & Fiedler: a PROCEDURAL feature, previously activated information acting on the processing of subsequent information. The definition is what makes the import from psychology into economics arguable at all Quotation withheld because this record is not cleared for redistribution. priming as against framing - separated in a FOOTNOTE, and the distinction is the same one the 2015 paper puts in its main text: framing re-describes a logically equivalent decision problem; priming activates stored knowledge and leaves the problem alone Quotation withheld because this record is not cleared for redistribution. revealed preferences, as used here - the actual tokens handed over in a one-shot public good game, offered as the behavioural counterpart to the self-reported emotions Quotation withheld because this record is not cleared for redistribution. how priming works, in THIS paper's own words - a two-step chain through beliefs, stated as a finding about the female subsample: primes move what females expect others to give, and that expectation moves what they give. The 2015 publication of the same experiment reports no belief effect and attributes the result to preferences instead Quotation withheld because this record is not cleared for redistribution. builds-on → Bargh et al. (2001) word-search priming methodology Quotation withheld because this record is not cleared for redistribution. builds-on → Croson and Gneezy (2009) on gender differences in social preferences Quotation withheld because this record is not cleared for redistribution. priming —distinct-from→ framing priming framing differs in: WHETHER THE DECISION PROBLEM IS RE-DESCRIBED - framing restates a logically equivalent problem in a different light, priming activates stored knowledge and leaves the problem as it was · strength of the relation: 0.95 · note: Drawn in a footnote here and in the main text of the same authors' 2015 paper, in almost the same words. A distinction they carried across five years and two venues. Quotation withheld because this record is not cleared for redistribution. priming —causes→ contribution levels priming contribution levels depends on: a gender control - the sentence states the result only under it · direction on the measure: increases · strength of the relation: 0.7 · note: WEIGHTED WELL BELOW THE 2015 EDGE FOR THE SAME CLAIM, and the qualifier is why: in this wave the RAW difference is not significant (Wilcoxon two-sided p = 0.249, Kolmogorov-Smirnov p = 0.552), and, elsewhere in the paper, the effect reaches only the 10% level once gender is added. The published version with a second wave finds it in the raw comparison. Quotation withheld because this record is not cleared for redistribution. priming —causes→ beliefs priming beliefs holds when: in the female subsample - the sentence names females as the ones whose beliefs shift · strength of the relation: 0.6 — the record marks this weak · note: THE FIRST STEP OF THIS PAPER'S MECHANISM, AND THE ONE THE 2015 PUBLICATION CONTRADICTS: there, priming does not affect beliefs at all, and that null is what licenses the preferences claim. See the unhoused-relations row on this record. Quotation withheld because this record is not cleared for redistribution. beliefs —causes→ contribution decisions beliefs contribution decisions strength of the relation: 0.6 — the record marks this weak · note: The second step. Both steps come from one sentence, and the paper presents them as a sequence rather than a correlation. Quotation withheld because this record is not cleared for redistribution. priming —causes→ emotional responses priming emotional responses direction on the measure: increases · strength of the relation: 0.8 · note: The paper's cleanest positive result, and the one the 2015 publication drops entirely - there the emotions data are collected and then set aside unused. Quotation withheld because this record is not cleared for redistribution. ordinary least squares regression ordinary least squares regression Quotation withheld because this record is not cleared for redistribution. ordered probit regression ordered probit regression Quotation withheld because this record is not cleared for redistribution. Wilcoxon rank-sum test Wilcoxon rank-sum test Quotation withheld because this record is not cleared for redistribution. Kolmogorov-Smirnov test Kolmogorov-Smirnov test Quotation withheld because this record is not cleared for redistribution. Spearman rank correlation Spearman rank correlation Quotation withheld because this record is not cleared for redistribution. robust standard errors robust standard errors Quotation withheld because this record is not cleared for redistribution. locally weighted regression (lowess) locally weighted regression (lowess) Quotation withheld because this record is not cleared for redistribution. Solving a word-search puzzle seeded with cooperation words raised voluntary contributions to a one-shot public good once subjects' gender was controlled for. Quotation withheld because this record is not cleared for redistribution. Mean contributions did not differ significantly between primed and unprimed treatments (p = 0.249). Quotation withheld because this record is not cleared for redistribution. The effect was concentrated in women: positively primed females contributed about 3.25 tokens more than neutrally primed females, whereas priming did not move male contributions. Quotation withheld because this record is not cleared for redistribution. Primed women expected more from their partners than primed men did, believing the other two group members would contribute 9.52 tokens against 6.44 tokens. Quotation withheld because this record is not cleared for redistribution. Within each treatment a subject's own contribution rose with what that subject believed others would contribute, the association being slightly tighter under priming. Quotation withheld because this record is not cleared for redistribution. Priming shifted self-reported positive emotions (significantly so when partners contributed at least as much as the subject) but left negative emotions unchanged. Quotation withheld because this record is not cleared for redistribution. Full cooperators contributing the whole 20-token endowment appeared only in the primed treatment, at 13 per cent, and were absent from the neutral treatment. Quotation withheld because this record is not cleared for redistribution. In both treatments the most common contribution choices were the extremes of giving nothing or giving half the endowment. Quotation withheld because this record is not cleared for redistribution. priming priming Quotation withheld because this record is not cleared for redistribution. framing framing Quotation withheld because this record is not cleared for redistribution. NP-treatment and P-treatment NP-treatment and P-treatment Quotation withheld because this record is not cleared for redistribution. priming manipulation priming manipulation Quotation withheld because this record is not cleared for redistribution. Level of contributions Level of contributions Quotation withheld because this record is not cleared for redistribution. Beliefs about contributions Beliefs about contributions Quotation withheld because this record is not cleared for redistribution. gender gender Quotation withheld because this record is not cleared for redistribution. field of study and nationality field of study and nationality Quotation withheld because this record is not cleared for redistribution. 51 ... 54 — subjects in the neutrally primed treatment and in the positively primed treatment Quotation withheld because this record is not cleared for redistribution. N = 19 ... N = 25 ... N = 32 ... N = 29 — cell sizes of the four gender-by-treatment groups used in the regression Quotation withheld because this record is not cleared for redistribution. 5 tokens — mean contribution out of a 20-token endowment in the neutrally primed treatment Quotation withheld because this record is not cleared for redistribution. 6.89 tokens — mean contribution in the positively primed treatment Quotation withheld because this record is not cleared for redistribution. 1.889 — raw uncontrolled contribution difference between primed and neutral groups, not significant Quotation withheld because this record is not cleared for redistribution. 2.104 — priming coefficient once gender is added as a control, significant at the 10 per cent level Quotation withheld because this record is not cleared for redistribution. 4.165 — extra tokens contributed by positively primed females relative to neutrally primed males, significant at 5 per cent Quotation withheld because this record is not cleared for redistribution. ρ = 0.572 and ρ = 0.6734 — correlation between own contribution and beliefs about others' contributions, neutral and primed treatments Quotation withheld because this record is not cleared for redistribution. p-value = 0.000 — difference in mean positive emotions between treatments when partners' deviations were non-negative Quotation withheld because this record is not cleared for redistribution. 0.5 money units — per-member return on each token contributed, against a private return of 1 for keeping it Quotation withheld because this record is not cleared for redistribution. £0.40 per guilder — exchange rate used to convert experimental earnings into cash payments Quotation withheld because this record is not cleared for redistribution.

The full text of this article is not redistributable, so it is not included here. The evidence sentences are withheld with it; the tag layer and its grounding status remain. The article itself is at https://doi.org/10.2139/ssrn.1631098.