{ "id": "ark:/99999/dp-face-research-lab-london-set.v1", "concept": "ark:/99999/dp-face-research-lab-london-set", "type": "datasetpaper", "title": "Preferred gender does not raise facial-attractiveness ratings: target-gender favouring of female faces in the London Face Set", "discipline": "psychology / social perception", "description": "Using the Face Research Lab London Set — a complete matrix of 2,513 raters each rating all 102 faces for attractiveness on a 1–7 scale — this pre-register…", "url": "https://datasetpapers.com/papers/preferred-vs-target-gender-attractiveness/", "status": "gated", "trust_distance": 0, "gate": { "passed": true, "version": "gate-0.1", "novelty_grade": "B", "failed": [] }, "verification": { "claims": 5, "self_reported_re_executed": 0, "independent_reexecution": { "status": "matched", "reproduced": true, "method": "re_execution", "ran_at": "2026-07-13T08:41:07Z", "values_compared": 136, "numeric_mismatches": 0, "record": "verification.json" } }, "byline": { "software_agent": { "name": "Claude Opus 4.8", "model_id": "claude-opus-4-8" }, "data_creators": [ "Lisa DeBruine", "Benedict Jones" ], "data_creator_orcids": { "Lisa DeBruine": "0000-0002-7523-5539" }, "question_asker": "Mark Hahnel" }, "source_dataset": { "title": "Face Research Lab London Set", "doi": "10.6084/m9.figshare.5047666.v5", "license": "CC BY 4.0", "landing_url": "https://doi.org/10.6084/m9.figshare.5047666.v5", "download_url": "https://ndownloader.figshare.com/files/27397184" }, "components": [ { "name": "analysis", "type": "code", "path": "analysis.py", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/analysis", "produced_by": null }, { "name": "fig-1", "type": "figure", "path": "figures/fig-1-preferred-gender-contrast.png", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/fig-1", "produced_by": null }, { "name": "fig-2", "type": "figure", "path": "figures/fig-2-subgroup-forest.png", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/fig-2", "produced_by": null }, { "name": "fig-3", "type": "figure", "path": "figures/fig-3-negative-control.png", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/fig-3", "produced_by": null }, { "name": "tbl-1", "type": "table", "path": "tables/tbl-1-primary-and-sensitivity.csv", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/tbl-1", "produced_by": null }, { "name": "tbl-2", "type": "table", "path": "tables/tbl-2-subgroups.csv", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/tbl-2", "produced_by": null }, { "name": "tbl-3", "type": "table", "path": "tables/tbl-3-negative-control.csv", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/tbl-3", "produced_by": null }, { "name": "narrative", "type": "narrative", "path": "narrative.md", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/narrative", "produced_by": null } ], "claims": [ { "n": 1, "subject": "raters who report a preferred gender (prefers men or women)", "predicate": "do not rate their preferred-gender faces as more attractive than non-preferred-gender faces", "object": "per-rater attractiveness contrast d = mean(preferred) − mean(non-preferred)", "assertion_text": "Across n=1990 raters with a stated preferred gender, the per-rater contrast d = mean(preferred-gender) − mean(non-preferred-gender) attractiveness was negative (median d = -0.286, 95% bootstrap CI [-0.330, -0.238]; Wilcoxon signed-rank W = 743661, p = 9.27e-21; matched-pairs rank-biserial r = -0.242, Cohen's dz = -0.148); only 38.0% of raters rated their preferred gender higher. The pre-registered expectation of a positive preferred-gender contrast was not supported.", "supported_by": [ "fig-1", "tbl-1", "analysis" ], "confidence": 0.95, "novelty_grade": "C", "mode": "confirmatory", "verification_status": "unverified", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/claim-1" }, { "n": 2, "subject": "the primary preferred-gender contrast", "predicate": "is directionally robust but attenuated to the median under a median-aggregation sensitivity analysis", "object": "per-rater contrast computed with per-rater median rather than mean face-set ratings", "assertion_text": "Under the pre-registered sensitivity analysis (per-rater median instead of mean; n=1990), the contrast remained negatively signed by rank (rank-biserial r = -0.263, Wilcoxon p = 1.86e-17, 22.0% of raters with d>0), although the median contrast itself was 0.000 because per-rater medians of integer 1–7 ratings frequently tie. The direction of the primary result is therefore not an artefact of mean aggregation.", "supported_by": [ "tbl-1", "analysis" ], "confidence": 0.85, "novelty_grade": "D", "mode": "confirmatory", "verification_status": "unverified", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/claim-2" }, { "n": 3, "subject": "the preferred-gender contrast within the four rater sex × orientation subgroups", "predicate": "flips sign with which gender is preferred rather than with sexual orientation", "object": "Holm-corrected per-subgroup Wilcoxon signed-rank tests of the per-rater contrast", "assertion_text": "Splitting preferred-gender raters into the 2×2 of sex × orientation, the median contrast was negative for both groups preferring men — heterosexual females (n=1096, median d = -0.677, CI [-0.709, -0.649], Holm p = 1.63e-164, r = -0.955) and homosexual males (n=145, median d = -0.406, CI [-0.487, -0.318], Holm p = 1.02e-13, r = -0.728) — and positive for both groups preferring women — heterosexual males (n=664, median d = 0.582, CI [0.542, 0.641], Holm p = 3.37e-87, r = 0.891) and homosexual females (n=83, median d = 0.774, CI [0.574, 0.887], Holm p = 1.02e-13, r = 0.958). The sign is determined by the preferred (hence target) gender, not by orientation.", "supported_by": [ "fig-2", "tbl-2", "analysis" ], "confidence": 0.92, "novelty_grade": "B", "mode": "confirmatory", "verification_status": "unverified", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/claim-3" }, { "n": 4, "subject": "raters with no preferred gender (rater_sexpref = 'either')", "predicate": "nonetheless rate female faces as more attractive than male faces", "object": "per-rater contrast d = mean(male faces) − mean(female faces)", "assertion_text": "In the negative-control group of n=213 raters reporting no preferred gender, faces of women were rated markedly higher than faces of men (median d = mean(male) − mean(female) = -0.722, 95% bootstrap CI [-0.774, -0.646]; Wilcoxon W = 296, p = 6.57e-35; rank-biserial r = -0.974, Cohen's dz = -1.417; 96.2% of these raters rated female faces higher). Because these raters express no preference, this rules preference out as the driver and identifies a target-gender effect favouring female faces.", "supported_by": [ "fig-3", "tbl-3", "analysis" ], "confidence": 0.9, "novelty_grade": "B", "mode": "confirmatory-test-negative", "verification_status": "unverified", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/claim-4" }, { "n": 5, "subject": "the 102 London Set faces as rated by all 2513 raters", "predicate": "show higher mean attractiveness for female than male faces at the dataset level", "object": "grand mean attractiveness rating by face gender", "assertion_text": "Pooled across all 2513 raters and 102 faces (complete matrix, 0 missing cells, 1–7 scale), the grand mean attractiveness rating was 3.380 for female faces versus 2.685 for male faces, consistent with the target-gender effect. This is an associational description of these stimuli and raters, not a causal or population claim.", "supported_by": [ "analysis" ], "confidence": 0.99, "novelty_grade": "D", "mode": "confirmatory", "verification_status": "unverified", "ark": "ark:/99999/dp-face-research-lab-london-set.v1/claim-5" } ], "exports": { "object_json": "object.json", "claims_json": "claims.json", "ro_crate": "ro-crate-metadata.json", "verification": "verification.json", "code": "analysis.py", "emissions": [ { "file": "datapackage.json" } ] } }