CoChat
Sign Up
Cited by 21Evolutionary Psychology and Human BehaviorFace Recognition and PerceptionAesthetic Perception and Analysis

What is beautiful is still good: the attractiveness halo effect in the era of beauty filters

Open access
Gulati, Aditya, Martínez-Garcia, Marina, Fernández, Daniel, Lozano, Miguel Angel, Lepri, Bruno, Oliver, Nuria·2024·Royal Society Open Science
DOI 10.1098/rsos.240882PMC PMC11597472

Published version of a preprint

Open this paper in CoChat — free account

Abstract

The impact of cognitive biases on decision-making in the digital world remains under-explored despite its well-documented effects in physical contexts. This paper addresses this gap by investigating the attractiveness halo effect using AI-based beauty filters. We conduct a large-scale online user study involving 2748 participants who rated facial images from a diverse set of 462 distinct individuals in two conditions: original and attractive after applying a beauty filter. Our study reveals that the same individuals receive statistically significantly higher ratings of attractiveness and other traits, such as intelligence and trustworthiness, in the attractive condition. We also study the impact of age, gender and ethnicity and identify a weakening of the halo effect in the beautified condition, resolving conflicting findings from the literature and suggesting that filters could mitigate this cognitive bias. Finally, our findings raise ethical concerns regarding the use of beauty filters.

AI snapshot

Written from the paper’s full text.

TL;DR

A large-scale online study with 2748 raters and 462 faces finds that AI beauty filters increase perceived attractiveness and also raise ratings of intelligence, trustworthiness, sociability, and happiness for the same individuals, supporting the attractiveness halo effect. The halo effect weakens after beautification for intelligence and trustworthiness, suggesting filters may mitigate some bias while raising ethical concerns.

Key findings

  • Beauty filters significantly increased perceived attractiveness: median increase of 1 point on a 7-point scale; 96.1% of individuals increased, 3.9% (18/462) unchanged, and none decreased. Mean attractiveness rose from 3.57 (PRI) to 5.01 (POST), and variance dropped from 0.83 to 0.60.
  • The same individuals received significantly higher ratings on intelligence, trustworthiness, sociability, and happiness after beautification (p<0.001). Attractiveness remained a significant predictor of all four attributes before and after beautification, but the effect weakened for intelligence (β1 0.30 to 0.12; R² 0.327 to 0.036) and trustworthiness (β1 0.20 to 0.17; R² 0.181 to 0.069).
  • Age and gender of stimuli mattered: younger and female faces received higher attractiveness ratings (p<0.001), while ethnicity showed no statistically significant effect. Female faces were rated more attractive but less intelligent than male faces, especially after beautification (p<0.001).
  • A saturation effect was identified: the relationship between attractiveness and dependent attributes became nonlinear, strongest for intelligence and trustworthiness, which the authors propose may explain inconsistent prior findings and suggest filters could mitigate the halo effect.

Methods

  • Pre-registered online user study via Prolific: n=2748 participants (1375 men, 1373 women; age M=46.47, s.d.=15.09), mostly White (2291) and living in the UK, USA, or Canada (94%). Each participant rated 10 face images; each image received at least 25 ratings.
  • Stimuli: 462 unique neutral-expression faces from the Chicago Face Database and FACES dataset, each shown in original (PRI) and beautified (POST) versions created with a popular selfie-editing beauty filter. No participant rated the same individual in both conditions.
  • Analysis: centralized median ratings compared with Wilcoxon, Kruskal–Wallis, and Levene tests; individual ratings transformed using ordered stereotype models and analyzed with linear mixed models including rater and stimulus age/gender and interactions, with raters as random effects.

Limitations

  • Stated in the paper: participants lacked geographic and ethnic diversity, being predominantly white native English speakers from the USA and UK; analyses are aggregate-level because each participant saw a different image set; only one beauty filter type was tested; negative traits and the halo update effect were not studied; static 2D photos may not generalize to real-world interactions.
  • Evident from the text: gender is treated as binary and ethnicity uses six dataset labels, limiting finer-grained identity analysis; race was not included in the individual-level models because centralized results showed no significant race effect and raters were predominantly white; the study cannot establish a causal link between stereotype formation and the observed saturation effect.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1098/rsos.240882

Cites

126 of 126 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

24 of 90 cited sources checked · 12 supported · 0 flagged · accuracy not yet measured

  • Findingown data

    The rater's age does not exhibit a significant correlation with perceived attractiveness.

    “Conversely, the rater’s age (β5) does not exhibit a significant correlation with perceived attractiveness, which is aligned with previous research [67,68] and in contradiction with what other authors have reported [69].”
    Quote in the paperNot a substantial claim1 of 3 sources checked
    • cites [67] Cross JF, Cross J, Daly J. 1971 Sex, race, age, and beauty as factors in recognition of faces. Percept. Psychophys. 10, 393– 396. (10.3758/BF03210319) source

      The Results report that the rater’s age was not a statistically significant main effect for face recognition, not for perceived attractiveness. The Stimulus Material section says beauty ratings provided data for a separate study; the passages on perceived beauty address its effect on recognition. Our copy does not report whether rater age correlates with attractiveness ratings. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
    • cites [68] Perlini AH, Marcello A, Hansen SD, Pudney W. 2001 The effects of male age and physical appearance on evaluations of attractivenes, social desirability and resourcefulness. Soc. Behav. Pers. 29, 277– 287. (10.2224/sbp.2001.29.3.277) source
    • cites [69] Foos PW, Clark MC. 2011 Adult age and gender differences in perceptions of facial attractiveness: beauty is in the eye of the older beholder. J. Genet. Psychol. 172, 162– 175. (10.1080/00221325.2010.526154) 21675545 source
  • Premisecitation

    Faces with make-up applied are consistently rated as more attractive than those without make-up.

    “faces with make-up applied to them were consistently rated as more attractive than those without make-up.”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [34] Tagai K, Ohtaka H, Nittono H. Faces with light makeup are better recognized than faces with heavy makeup. Front. Psychol. 7, 226. (10.3389/fpsyg.2016.00226) 26973553 PMC4771839 source
    • cites [31] Batres C, Porcheron A, Courrèges S, Russell R. 2021 Professional versus self-applied makeup: do makeup artists add value? Perception 50, 709– 719. (10.1177/03010066211029218) 34210220 source
  • Findingcitation

    Youth is positively correlated with sociability, especially after beautification.

    “Conversely, youth is positively correlated with sociability, especially after beautification, which is supportive of previous research [85].”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [85] Hess U, Adams RB, Simard A, Stevenson MT, Kleck RE. 2012 Smiling and sad wrinkles: age-related changes in the face and the perception of emotions and intentions. J. Exp. Soc. Psychol. 48, 1377– 1380. (10.1016/j.jesp.2012.05.018) 23144501 PMC3491992 source

      The Results (Behavioral intention ratings; page unavailable) report that younger faces were generally rated as more affiliative, the measure participants rated as sociability, except for angry faces. The Method (Stimuli; page unavailable) describes applying identical expressions to young and old faces, not beautifying them; the paper does not support the claim that the association was stronger after beautification. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
  • Findingown data

    There is a gender bias in perceptions of intelligence, with female stimuli rated lower on intelligence.

    “This finding suggests the existence of a gender bias in perceptions of intelligence [3,6].”
    • Also stated in: § Discussion · hedged
      “As discussed below, we find evidence of the existence of a gender bias when judging the intelligence of female stimuli, which could correspond to the application of a different stereotype for highly attractive females.”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [3] Kleisner K, Chvátalová V, Flegr J. 2014 Perceived intelligence is associated with measured intelligence in men but not women. PLoS One (ed. B Fink), 9, e81237. (10.1371/journal.pone.0081237) 24651120 PMC3961208 source

      The Discussion (Correlation between IQ and perceived intelligence; page unavailable) suggests that observers may judge women’s faces using criteria that work for men’s faces, or judge women more strongly by attractiveness. This supports a possible gender difference in how intelligence is perceived, but the cited passages do not show that female stimuli received lower intelligence ratings than male stimuli. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites [6] Talamas SN. 2016 Perceptions of intelligence and the attractiveness halo. PhD thesis, University of St Andrews, UK. https://hdl.handle.net/10023/10851.

      Not checked in the free run.

  • Premisecitation

    The use of cosmetics has been shown to increase perceptions of attractiveness.

    “Cosmetics are a popular tool to alter appearance and their use has been shown to increase perceptions of attractiveness [29–36].”
    Quote in the paperNot a substantial claim2 of 8 sources checked
    • cites [29] Russell R. 2009 A sex difference in facial contrast and its exaggeration by cosmetics. Perception 38, 1211– 1219. (10.1068/p6331) 19817153 source
    • cites [30] Batres C, Robinson H. 2022 Makeup increases attractiveness in male faces. PLoS One (ed. AB Mahmoud), 17, e0275662. (10.1371/journal.pone.0275662) 36327222 PMC9632771 source
    • cites [31] Batres C, Porcheron A, Courrèges S, Russell R. 2021 Professional versus self-applied makeup: do makeup artists add value? Perception 50, 709– 719. (10.1177/03010066211029218) 34210220 source
    • cites [32] Batres C, Porcheron A, Latreille J, Roche M, Morizot F, Russell R. 2019 Cosmetics increase skin evenness: evidence from perceptual and physical measures. Skin Res. Technol. 25, 672– 676. (10.1111/srt.12700) 31111541 source
    • cites [33] Mulhern R, Fieldman G, Hussey T, Lévêque J ‐L., Pineau P. 2003 Do cosmetics enhance female Caucasian facial attractiveness? Int. J. Cosmet. Sci. 25, 199– 205. (10.1046/j.1467-2494.2003.00188.x) 18494902 source
    • cites [34] Tagai K, Ohtaka H, Nittono H. Faces with light makeup are better recognized than faces with heavy makeup. Front. Psychol. 7, 226. (10.3389/fpsyg.2016.00226) 26973553 PMC4771839 source
    • cites [35] Batres C, Russell R, Simpson JA, Campbell L, Hansen AM, Cronk L. 2018 Evidence that makeup is a false signal of sociosexuality. Pers. Individ. Dif. 122, 148– 154. (10.1016/j.paid.2017.10.023) source
    • cites [36] Graham JA, Jouhar AJ. 1981 The effects of cosmetics on person perception. Int. J. Cosmet. Sci. 3, 199– 210. (10.1111/j.1467-2494.1981.tb00283.x) 19469939 source
  • May combine two claims. 30
  • Classified as probably not a substantial claim. 42
  • The quote only partly states this claim. 2

  • May combine two claims.

    Study design and conductcitation

    Beauty filters enable a controlled and consistent adjustment of attractiveness at a large scale while keeping the identity of the face constant, which is crucial for isolating the effects of perceived attractiveness from other confounding variables.

    “The manipulation of beauty by means of beauty filters enables a controlled and consistent adjustment of attractiveness at a large scale while keeping the identity of the face constant [45], which is crucial for isolating the effects of perceived attractiveness from other confounding variables, such as facial identity or expression.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [45] Riccio P, Psomas B, Galati F, Escolano F, Hofmann T, Oliver N. 2022 Openfilter: a framework to democratize research access to social media AR filters. Adv. Neural Inf. Process. Syst. 35, 12 491– 12 503. (10.48550/arXiv.2207.12319) source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The evidence on the interaction between gender of stimuli and gender of evaluators regarding the attractiveness halo effect is mixed.

    “Again, there is mixed evidence in this regard.”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites [4] Dion K, Berscheid E, Walster E. 1972 What is beautiful is good. J. Pers. Soc. Psychol. 24, 285– 290. (10.1037/h0033731) 4655540 source
    • cites [25] Agthe M, Strobel M, Spörrle M, Pfundmair M, Maner JK. 2016 On the borders of harmful and helpful beauty biases. Evol. Psychol. 14. (10.1177/1474704916653968) source
    • cites [24] Kunst JR, Kirkøen J, Mohamdain O. 2023 Hacking attractiveness biases in hiring? The role of beautifying photo-filters. Manage. Decis. 61, 924– 943. (10.1108/MD-06-2021-0747) source

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    The choice of using intelligence, trustworthiness, sociability and happiness as dependent attributes for the halo effect was driven by existing literature.

    “The choice of using intelligence, trustworthiness, sociability and happiness as dependent attributes for the halo effect was driven by existing literature on this cognitive bias, such as [4,6,8,9,22,53,60–63].”
    Quote in the paperSubstantial claim2 of 10 sources checked
    • cites [4] Dion K, Berscheid E, Walster E. 1972 What is beautiful is good. J. Pers. Soc. Psychol. 24, 285– 290. (10.1037/h0033731) 4655540 source
    • cites [6] Talamas SN. 2016 Perceptions of intelligence and the attractiveness halo. PhD thesis, University of St Andrews, UK. https://hdl.handle.net/10023/10851.

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites [8] Golle J, Mast FW, Lobmaier JS. 2014 Something to smile about: the interrelationship between attractiveness and emotional expression. Cogn. Emot. 28, 298– 310. (10.1080/02699931.2013.817383) 23875865 source
    • cites [9] Todorov A. 2008 Evaluating faces on trustworthiness: an extension of systems for recognition of emotions signaling approach/avoidance behaviors. Ann. NY. Acad. Sci. 1124, 208– 224. (10.1196/annals.1440.012) 18400932 source
    • cites [22] Batres C, Shiramizu V. 2023 Examining the 'attractiveness halo effect' across cultures. Curr. Psychol. 42, 25 515– 25 519. (10.1007/s12144-022-03575-0) source
    • cites [53] Ma DS, Correll J, Wittenbrink B. 2015 The Chicago face database: a free stimulus set of faces and norming data. Behav. Res. Methods 47, 1122– 1135. (10.3758/s13428-014-0532-5) 25582810 source

      The Contributions of the CFD section lists trustworthiness and happy expression among its face ratings, and the Results relate both ratings to experts’ judgments of face suitability (page numbers unavailable). The paper does not establish intelligence or sociability as attributes for a halo-effect study, or explain the citing authors’ choice of dependent attributes. (LLM, conf=0.82)

      Experimental — accuracy not yet measured
    • cites [60] Oosterhof NN, Todorov A. 2008 The functional basis of face evaluation. Proc. Natl Acad. Sci. USA 105, 11 087– 11 092. (10.1073/pnas.0805664105) 18685089 PMC2516255 source
    • cites [61] Zebrowitz LA, Bronstad PM, Lee HK. 2007 The contribution of face familiarity to ingroup favoritism and stereotyping. Soc. Cogn. 25, 306– 338. (10.1521/soco.2007.25.2.306) source
    • cites [62] Peterson JC, Uddenberg S, Griffiths TL, Todorov A, Suchow JW. 2022 Deep models of superficial face judgments. Proc. Natl Acad. Sci. USA 119, e2115228119. (10.1073/pnas.2115228119) 35446619 PMC9169911 source

      In Results, “The Structure of Attribute Inferences” (page unavailable), the paper analyzes ratings of smartness, trustworthiness, happiness, and outgoingness, which is close to sociability, and compares some findings with prior work. It does not describe these four attributes as dependent measures of a halo effect or say that literature on that bias motivated their selection. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
    • cites [63] Todorov A, Dotsch R, Porter JM, Oosterhof NN, Falvello VB. 2013 Validation of data-driven computational models of social perception of faces. Emotion 13, 724– 738. (10.1037/a0032335) 23627724 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisenone given

    Categories with overlapping standard deviation intervals are merged into the same category in the OSM.

    “Furthermore, categories with overlapping standard deviation intervals are merged into the same category.”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites [57] Anderson JA. 1984 Regression and ordered categorical variables. J. R. Stat. Soc. B 46, 1– 22. (10.1111/j.2517-6161.1984.tb01270.x) source
    • cites [58] Fernández D, Arnold R, Pledger S. 2016 Mixture-based clustering for the ordered stereotype model. Comput. Stat. Data Anal. 93, 46– 75. (10.1016/j.csda.2014.11.004) source

      Section 5.2.2 describes combining two adjacent categories whose estimated score parameters were very close, and the simulation study says categories with equal score parameters could be merged for detecting clustering. The paper does not give overlapping standard-deviation intervals as the criterion, nor say that such categories are automatically merged. Page numbers are unavailable in our copy. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites [59] Fernandez D, Liu I, Costilla R. 2019 A method for ordinal outcomes: the ordered stereotype model. Int. J. Methods Psychiatr. Res. 28, e1801. (10.1002/mpr.1801) 31568635 PMC7027430 source

      The Methods section (‘Estimation of non‐equal space among ordinal categories’; page unavailable) says adjacent categories could be collapsed if their score parameters are equal. It says overlapping confidence intervals may provide evidence that categories are not distinguishable; it does not specify standard deviation intervals or say that overlap automatically triggers merging. (LLM, conf=0.87)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisereasoning

    The OSM estimates the distance between different categories based on the actual data instead of assuming that all categories are equidistant.

    “The ϕk′s are interpreted as scores and help estimate the distance between different categories based on the actual data (ratings in our case) instead of assuming that all categories are equidistant.”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites [57] Anderson JA. 1984 Regression and ordered categorical variables. J. R. Stat. Soc. B 46, 1– 22. (10.1111/j.2517-6161.1984.tb01270.x) source
    • cites [58] Fernández D, Arnold R, Pledger S. 2016 Mixture-based clustering for the ordered stereotype model. Comput. Stat. Data Anal. 93, 46– 75. (10.1016/j.csda.2014.11.004) source
    • cites [59] Fernandez D, Liu I, Costilla R. 2019 A method for ordinal outcomes: the ordered stereotype model. Int. J. Methods Psychiatr. Res. 28, e1801. (10.1002/mpr.1801) 31568635 PMC7027430 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationcitation

    Our results are supportive of previously reported examples of gender-based discrimination and the challenges faced by women in various spheres of life, including education and professional opportunities.

    “Moreover, our results are supportive of previously reported examples of gender-based discrimination and the challenges faced by women in various spheres of life, including education and professional opportunities [90–93].”
    Quote in the paperNot a substantial claim1 of 4 sources checked
    • cites [90] González MJ, Cortina C, Rodríguez J. 2019 The role of gender stereotypes in hiring: a field experiment. Eur. Sociol. Rev. 35, 187– 204. (10.1093/esr/jcy055) source
    • cites [91] Marlowe CM, Schneider SL, Nelson CE. 1996 Gender and attractiveness biases in hiring decisions: are more experienced managers less biased? J. Appl. Psychol. 81, 11– 21. (10.1037//0021-9010.81.1.11) source
    • cites [92] Eagly AH, Sczesny S. 2009 Stereotypes about women, men, and leaders: have times changed?, pp. 21– 47. Washington, DC: American Psychological Association. (10.1037/11863-002) source
    • cites [93] Heilman ME. 2001 Description and prescription: how gender stereotypes prevent women’s ascent up the organizational ladder. J. Soc. Issues 57, 657– 674. (10.1111/0022-4537.00234) source

  • Classified as probably not a substantial claim.

    Study design and conductown data

    The appropriateness of using OSMs with linear models (and linear mixed models) was evaluated by computing AIC and BIC of different models for attractiveness and each of the four dependent variables.

    “In addition to the theoretical grounding, we further evaluated the appropriateness of using the OSMs with linear models (and linear mixed models) by computing the Akaike information criterion (AIC) [120] and the Bayesian information criterion (BIC) [121] of different models for attractiveness and each of the four dependent variables.”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [120] Akaike H. 1974 A new look at the statistical model identification. IEEE Trans. Automat. Contr. 19, 716– 723. (10.1109/TAC.1974.1100705) source
    • cites [121] Schwarz G. 1978 Estimating the dimension of a model. Ann. Statist. 6, 461– 464. (10.1214/aos/1176344136) source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisereasoning

    The OSM estimates the underlying scale by computing the expected probabilities of the categories based on potential covariates in the data.

    “Thus, the OSMs estimate the underlying scale by computing the expected probabilities of the categories based on potential covariates in the data.”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites [57] Anderson JA. 1984 Regression and ordered categorical variables. J. R. Stat. Soc. B 46, 1– 22. (10.1111/j.2517-6161.1984.tb01270.x) source
    • cites [58] Fernández D, Arnold R, Pledger S. 2016 Mixture-based clustering for the ordered stereotype model. Comput. Stat. Data Anal. 93, 46– 75. (10.1016/j.csda.2014.11.004) source
    • cites [59] Fernandez D, Liu I, Costilla R. 2019 A method for ordinal outcomes: the ordered stereotype model. Int. J. Methods Psychiatr. Res. 28, e1801. (10.1002/mpr.1801) 31568635 PMC7027430 source

  • May combine two claims.

    Premisecitation

    Frequent use of beauty filters has already been found to lead to anxiety and depression, reduced self-esteem, body dysmorphia, an increase of plastic surgery, feelings of inadequacy and increased pressure to conform to unrealistic beauty standards.

    “Frequent use of beauty filters has already been found to lead to anxiety and depression, reduced self-esteem, body dysmorphia, an increase of plastic surgery, feelings of inadequacy and increased pressure to conform to unrealistic beauty standards [47,50,52,96,99–102].”
    Quote in the paperNot a substantial claim2 of 8 sources checked
    • cites [47] Lavrence C, Cambre C. 2020 'Do I look like my selfie?': filters and the digital-forensic gaze. Soc. Media + Soc. 6. (10.1177/2056305120955182) source
    • cites [50] Isakowitsch C. 2023 How augmented reality beauty filters can affect self-perception. In Artificial Intelligence and Cognitive Science. AICS 2022. Communications in Computer and Information Science (eds L Longo, R O’Reilly), pp. 239– 250, vol. 1662. Cham, Switzerland: Springer. (10.1007/978-3-031-26438-2_19) source

      Section 2.2 reports lower self-acceptance from using beauty filters to idealize oneself, while section 2.1 describes a possible kind of body dysmorphia and concern about the promotion of plastic surgery. Sections 4.1–4.2 report anxiety in one participant and mixed emotional responses after filter use. The paper does not establish that frequent use leads to depression, increased plastic surgery, or all the listed outcomes; its findings and cautions are narrower than the citing sentence. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites [52] Eshiet J. 2020 'Real me versus social media me': filters, Snapchat dysmorphia, and beauty perceptions among young women. Master’s thesis, California State University, Long Beach, CA. https://scholarworks.lib.csusb.edu/cgi/viewcontent.cgi?article=2243&context=etd.

      Not checked in the free run.

    • cites [96] Gill R. 2021 Changing the perfect picture: smartphones, social media and appearance pressures. See https://www.city.ac.uk/__data/assets/pdf_file/0005/597209/Parliament-Report-web.pdf.

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites [99] Bakker M. 2022 Nofilter: how beauty filters affect the internalization of beauty ideals. Master’s thesis, Utrecht University, Utrecht, The Netherlands. https://studenttheses.uu.nl/bitstream/handle/20.500.12932/41835/Thesis_SHOP_JMC_Bakker__6224415_22_06-2022.pdf?sequence=1.

      Not checked in the free run.

    • cites [100] Felisberti FM, Musholt K. 2014 Self-face perception: individual differences and discrepancies associated with mental self-face representation, attractiveness and self-esteem. Psychol. Neurosci. 7, 65– 72. (10.3922/j.psns.2014.013) source
    • cites [101] Verrastro V, Fontanesi L, Liga F, Cuzzocrea F, Gugliandolo MC. 2020 Fear the Instagram: beauty stereotypes, body image and Instagram use in a sample of male and female adolescents. Qwerty 15, 31– 49. (10.30557/QW000021) source

      The Discussion (page unavailable) reports that adolescents who post edited pictures show more body-related anxiety and dissatisfaction and feel more pressure to meet Instagram beauty standards. The paper does not establish that frequent beauty-filter use causes these outcomes, and it does not link such use to plastic surgery; its mentions of depression, self-esteem, and dysmorphia concern broader Instagram use or beauty standards rather than that specific behavior. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites [102] Ryan-Mosley T. 2021 Beauty filters are changing the way young girls see themselves. MIT Technology Review. See https://www.technologyreview.com/2021/04/02/1021635/beauty-filters-young-girls-augmented-reality-social-media. source

  • May combine two claims.

    Findingown data

    Linear models in table 1 show a decrease in β1 and R2 after beautification, particularly for intelligence and trustworthiness.

    “The linear models in table 1 reflect a decrease in the value of β1 and R² of the linear models after beautification, particularly for intelligence and trustworthiness, supporting the hypothesis of a mitigation of the halo effect for these attributes.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    After beautification, images of females are given significantly lower trustworthiness scores by male raters, even though male and female raters provided similar trustworthiness scores to females before beautification.

    “Additionally, images of females are given significantly lower trustworthiness scores by male raters, even though male and female raters provided similar trustworthiness scores to females in the PRI dataset.”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    Gender differences in sociability and happiness ratings change significantly more after beautification for female raters than for male raters.

    “Moreover, the gender differences in attractiveness, intelligence and trustworthiness ratings change significantly more after beautification for male raters, whereas a similar effect is observed for sociability and happiness for female raters.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Perceived attractiveness had a significant positive effect on all dependent variables (intelligence, trustworthiness, sociability, happiness) both before and after beautification.

    “Linear models, depicted in table 1, of the centralized score for each dependent variable (ω) as a function of the centralized score of perceived attractiveness for each image (ω=β0+β1Attrac+ϵ) reveal a significant effect (p<0.001)1 of perceived attractiveness on all dependent variables both before (PRI) and after (POST) beautification.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    We find strong evidence of the existence of the halo effect both before and after beautification for the four dependent variables of interest (see table 1).

    “Contrary to previous works [38,39] and supportive of others [26,27,37], we find strong evidence of the existence of the halo effect both before and after beautification for the four dependent variables of interest (see table 1).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The centralized perceived attractiveness score increased for 96.1% of the individuals after beautification and remained unchanged for the rest.

    “The centralized perceived attractiveness score increased for 96.1% of the individuals after beautification and remained unchanged for the rest.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Before beautification, images of females receive significantly higher sociability scores from male raters than from female raters, and significantly higher attractiveness scores from female raters than from male raters.

    “Images of females receive significantly higher: (i) sociability (p<0.001) scores from male raters than from female raters; and (ii) attractiveness (p<0.001) scores from female raters than from male raters before beautification.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Beauty filters increase perceived attractiveness scores for almost all individuals, shifting the distribution of perceived attractiveness to the right on the 7-point Likert scale.

    “Beauty filters increase the perceived attractiveness scores for almost all individuals indicating that they shift the distribution of perceived attractiveness to the right on the 7-point Likert scale.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Both female and male raters provided higher ratings of attractiveness to images of females before and after beautification, with a widening gap between genders after beautification, especially for male raters.

    “Both female and male raters provided higher ratings of attractiveness to images of females before (p<0.001) and after (p<0.001) beautification, with a widening gap between genders after beautification, especially for male raters (figure 4a).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Perceived attractiveness is the strongest predictor of the dependent variables, but its predictive power decreases after beautification.

    “Perceived attractiveness (β1) is the strongest predictor of the dependent variables, yet its predictive power decreases in the models built with data after beautification (see appendix M for a detailed analysis).”
    Quote in the paperSubstantial claim

  • May combine two claims. The quote only partly states this claim.

    Findingown data

    Images of females are given significantly lower attractiveness scores by male raters than by female raters, both before and after beautification.

    “Images of females are also given significantly lower attractiveness (p<0.001) scores by male raters than by female raters, as observed in the PRI dataset.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    After beautification, female raters provide significantly different ratings to images of males and females on all attributes, whereas male raters provide significantly different scores only on perceived attractiveness, intelligence, and sociability but not on trustworthiness and happiness.

    “After beautification, female raters provide significantly different (p<0.001) ratings to images of males and females on all attributes, whereas male raters provide significantly different (p<0.001) scores to images of males and females only on perceived attractiveness, intelligence and sociability but not on trustworthiness and happiness.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The opposite impact of the filters is seen regarding sociability, with no significant difference observed in the ratings received by images of females after beautification despite there being a significant difference before beautification.

    “The opposite impact of the filters is seen regarding sociability, with no significant difference observed in the ratings received by images of females despite there being a significant difference before beautification.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    There is a negative and significant correlation between perceived intelligence, trustworthiness and age (particularly after beautification), suggesting that the older the stimulus, the more intelligent and trustworthy it is perceived.

    “The negative and significant correlation between perceived intelligence, trustworthiness and age (particularly after beautification) suggests that the older the stimulus, the more intelligent and trustworthy it is perceived.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Approximately 17% of the faces in the PRI dataset were rated as having an attractiveness level greater than or equal to 5, and this percentage increased to approximately 75% after beautification.

    “As a result, while only approximately 17% of the faces in the PRI dataset were rated as having an attractiveness level greater than or equal to 5 (with 4 being the neutral point on the scale), this percentage increased to approximately 75% after beautification.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    No image decreased in centralized perceived attractiveness after beautification, and it remained the same in 3.9% (18 out of 462) of cases.

    “There were no images where the centralized perceived attractiveness score decreased after beautification and it remained the same before/after beautification only in 3.9% (18 out of 462 images) of the cases.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Findingown data

    The participants' self-reported race was predominantly white.

    “Additionally, the participants’ self-reported race was predominantly white (see appendix D) and hence it was also not considered as a variable in the models.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Male stimuli are perceived as more intelligent than female stimuli both by male and female raters, with a widening gap between genders after beautification such that female stimuli are perceived as less intelligent on average by male raters after beautification than before applying the filter.

    “As described in the previous paragraph, male stimuli are perceived as more intelligent than female stimuli both by male and female raters, with a widening gap between genders after beautification such that female stimuli are perceived as less intelligent on average by male raters after beautification than before applying the filter.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Findingcitation

    Previously reported results with the centralized ratings revealed no significant impact of race, neither on attractiveness nor on the dependent attributes.

    “Race was not included as a variable in the analysis because the previously reported results with the centralized ratings revealed no significant impact of race, neither on attractiveness nor on the dependent attributes.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Participants considered males to be more intelligent than females, particularly after beautification (p<0.001), also with a widening gap between genders.

    “Conversely, participants considered males to be more intelligent than females, particularly after beautification (p<0.001), also with a widening gap between genders (figure 4b).”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductown data

    No participant rated an image corresponding to the same individual in both conditions, and participants were not told that half of the images corresponded to beautified versions.

    “Furthermore, no participant rated an image corresponding to the same individual in both conditions (with and without the filter applied) and participants were not told that half of the images that they evaluated corresponded to the beautified versions of the original face images.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    The same individuals were perceived not only as more attractive, but also as more intelligent, trustworthy, sociable, and happy after applying a beauty filter, providing evidence that supports the existence of the attractiveness halo effect.

    “Thus, the same individuals were perceived not only as more attractive, but also as more intelligent, trustworthy, sociable and happy after applying a beauty filter, providing evidence that supports the existence of the attractiveness halo effect.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    Our study design does not enable the establishment of a causal link between stereotype formation and the observed saturation effect.

    “However, our study design does not enable the establishment of a causal link between stereotype formation and the observed saturation effect.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Applying make-up requires technical skill, and perceptions of attractiveness differ depending upon the skill of the person applying the make-up.

    “applying make-up requires technical skill, and perceptions of attractiveness differ depending upon the skill of the person applying the make-up [31].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [31] Batres C, Porcheron A, Courrèges S, Russell R. 2021 Professional versus self-applied makeup: do makeup artists add value? Perception 50, 709– 719. (10.1177/03010066211029218) 34210220 source

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    There exists a saturation effect: beyond a certain level of perceived attractiveness, there is a significant reduction in the impact that attractiveness has on the dependent variables.

    “We postulate the existence of a saturation effect, i.e. beyond a certain level of perceived attractiveness, there is a significant reduction in the impact that attractiveness has on the dependent variables.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    We report findings at an aggregate level.

    “Second, we report findings at an aggregate level.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Each image received at least 25 ratings.

    “Furthermore, the images were presented such that each image received at least 25 ratings.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Half the images were from the POST dataset and the other half were from the PRI dataset.

    “Half the images were from the POST dataset, i.e. had a beauty filter applied on them, and the other half were from the PRI dataset.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Half the images corresponded to male and the other half to female subjects.

    “Half the images corresponded to male and the other half to female subjects.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductreasoning

    The images were balanced across ethnicity (for participants in the CFD condition) or across age groups (for participants in the FACES condition).

    “The images were also balanced across ethnicity (for participants in the CFD condition) or across age groups (for participants in the FACES condition).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The order of the questions was randomized for each participant.

    “The order of the questions was randomized for each participant as per the algorithm described below, but remained the same across all the images rated by the same participant.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Regarding race, 2291 participants reported being White, 181 Asian, 178 Black, 63 Mixed and 33 from other ethnic groups.

    “Regarding race, 2291 participants reported being White, 181 Asian, 178 Black, 63 Mixed and 33 participants reported being from other ethnic groups.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Participants in our study were not presented with images of the same individual before and after beautification.

    “Since participants in our study were not presented with images of the same individual before and after beautification, we leave the investigation of the halo update effect for future research.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Participants were randomly assigned to see faces either from the FACES dataset or the CFD.

    “Participants were randomly assigned to see faces either from the FACES dataset or the CFD.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductreasoning

    Each image received ratings from a different subset of participants, so pairwise comparisons are only possible at an aggregate level using centralized scores.

    “As a result of our study methodology, each image received ratings from a different subset of participants such that pairwise comparisons are only possible at an aggregate level by means of the centralized scores.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    The presentation order of the images was randomized.

    “The presentation order of the images was randomized and participants were not told that some of the images were beautified.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Using ordinal response models such as the cumulative link model is more appropriate for ordinal response variables.

    “Using ordinal response models such as the cumulative link model (CLM) [116,117] is more appropriate for ordinal response variables [118]”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [118] Agresti A. 2010 Analysis of ordinal categorical data, online edn. Hoboken, NJ: Wiley. (10.1002/9780470594001) source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    Participants who failed two or more attention checks were rejected and additional participants were recruited to replace them.

    “Participants who failed two or more attention checks were rejected and additional participants were recruited to replace them.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The gender of the stimulus plays a stronger role in impacting the perceptions of intelligence than perceived attractiveness given that images of females were rated as more attractive than those of males.

    “Therefore, we conclude that the gender of the stimulus plays a stronger role in impacting the perceptions of intelligence than perceived attractiveness given that images of females were rated as more attractive than those of males.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The user study was pre-registered in the Open Science Foundation registry.

    “The user study was pre-registered in the Open Science Foundation registry2 and was approved by the Ethics Board of the University of Alicante.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    No image received ratings from the same subset of participants.

    “Note that no image received ratings from the same subset of participants.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    We do not explore the impact of different beauty filters on the halo effect.

    “Third, we do not explore the impact of different beauty filters on the halo effect, but previous work has reported that popular beauty filters perform similar transformations to the faces [45].”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    The majority of participants (94%) reported living in the United Kingdom, United States or Canada.

    “The majority of participants (94%) reported living in the United Kingdom (1817), United States (686) or Canada (72).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    We opted to use a 7-point Likert scale because they have been reported to be the most accurate and reliable.

    “While some work collected these ratings on a 9-point Likert scale [22], we opted to use a 7-point Likert scale because they have been reported to be the most accurate and reliable [113–115], despite the popularity of 5-point scales.”
    Quote in the paperNot a substantial claim0 of 4 sources checked
    • cites [22] Batres C, Shiramizu V. 2023 Examining the 'attractiveness halo effect' across cultures. Curr. Psychol. 42, 25 515– 25 519. (10.1007/s12144-022-03575-0) source
    • cites [113] Johns R. 2010 Likert items and scales. Survey Q. Bank: Methods Fact Sheet 1, 11– 28.

      Not checked in the free run.

    • cites [114] Symonds PM. 1924 On the loss of reliability in ratings due to coarseness of the scale. J. Exp. Psychol. 7, 456– 461. (10.1037/h0074469) source
    • cites [115] M GA. 1956 The magical number seven plus or minus two: some limits on our capacity for processing information. Psychol. Rev. 63, 81– 97. (10.1037/h0043158) 13310704

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    Our findings might not generalize to real-world scenarios where attractiveness perceptions interact with other factors, such as situational dynamics, personality and social context.

    “Hence, our findings might not generalize to real-world scenarios where attractiveness perceptions interact with other factors, such as situational dynamics, personality and social context.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    A per-rater level analysis is not possible on our collected dataset because each participant was exposed to a different set of images to maximize diversity.

    “A per-rater level analysis, while interesting, is not possible on our collected dataset because each participant was exposed to a different set of images to maximize diversity.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Participants were not told that some of the images were beautified.

    “The presentation order of the images was randomized and participants were not told that some of the images were beautified.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The data collected in the study has been deposited in a public online repository.

    “The data collected in the study has been deposited in a public online repository [112].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [112] Gulati A, Oliver N, Lepri B. 2024 The beauty survey. Zenodo. (10.5281/zenodo.13836854) source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    The sample was n=2748, gender balanced: 1375 men and 1373 women.

    “The sample (n = 2748) was gender balanced: 1375 men and 1373 women, with ages ranging between 18 and 88 years old (age M = 46.47, s.d. = 15.09).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    We do not study the relationship between attractiveness and socially undesirable attributes, such as vanity or materialism.

    “Fourth, we do not study the relationship between attractiveness and socially undesirable attributes, such as vanity or materialism.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    We observe several and novel significant interactions between the gender of the stimulus and the gender of the human evaluators, contradicting early work that reported a lack of such an interaction.

    “In sum, we observe several and novel significant interactions between the gender of the stimulus and the gender of the human evaluators, contradicting early work that reported a lack of such an interaction [4].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [4] Dion K, Berscheid E, Walster E. 1972 What is beautiful is good. J. Pers. Soc. Psychol. 24, 285– 290. (10.1037/h0033731) 4655540 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Ages ranged between 18 and 88 years old (M=46.47, s.d.=15.09).

    “The sample (n = 2748) was gender balanced: 1375 men and 1373 women, with ages ranging between 18 and 88 years old (age M = 46.47, s.d. = 15.09).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    We do not study the potential perceived differences between real and beautified faces for the same level of attractiveness.

    “We also do not study the potential perceived differences between real and beautified faces for the same level of attractiveness.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Participants always rated images corresponding to 10 different individuals such that they never had to judge the same person in both the beautified and non-beautified conditions.

    “Furthermore, participants always rated images corresponding to 10 different individuals such that they never had to judge the same person in both the beautified and non-beautified conditions.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Seventeen participants failed at least two of the four attention checks and were removed from the analysis and replaced by new participants.

    “Seventeen participants failed at least two of the four attention checks and hence were removed from the analysis and replaced by new participants, yielding a total sample of 2748 participants.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The study was approved by the Ethics Board of the University of Alicante.

    “The user study was pre-registered in the Open Science Foundation registry2 and was approved by the Ethics Board of the University of Alicante.”
    Quote in the paperNot a substantial claim
65 of 65

  • The paper relies on this without support.

    The impact of cognitive biases on decision-making in the digital world remains under-explored.

    “The impact of cognitive biases on decision-making in the digital world remains under-explored despite its well-documented effects in physical contexts.”

  • The paper relies on this without support.

    Prior findings on the attractiveness halo effect have generally been obtained by small user studies with typically small samples of face images with limited diversity.

    “However, these findings have been generally obtained by means of small user studies where study participants provided judgements of a typically small sample of face images with limited diversity.”

  • The paper relies on this without support.

    There is a lack of research on how beauty filters impact perceptions of attractiveness and associated cognitive biases on the same individuals.

    “there is a lack of research on how they impact perceptions of attractiveness and associated cognitive biases on the same individuals.”

  • The paper relies on this without support.

    The improvements in attractiveness that can be achieved as a result of applying make-up are limited.

    “the improvements in attractiveness that can be achieved as a result of applying make-up are limited.”

  • The paper relies on this without support.

    The attention checks reliably identify inattentive participants.

    “Seventeen participants failed at least two of the four attention checks and hence were removed from the analysis and replaced by new participants, yielding a total sample of 2748 participants.”

  • The paper relies on this without support.

    The OSMs rescaling is appropriate for analyzing the relationship between attractiveness and the dependent variables.

    “Figure 5 depicts the relationship between perceived attractiveness and the dependent variables after rescaling the data according to the OSMs both before (a) and after (b) beautification.”

  • The paper relies on this without support.

    The participants' ratings reflect their genuine first impressions of the faces.

    “After providing informed consent, each participant was presented a page with instructions: they were told that they would be shown 10 face images and would be asked to provide their assessment of different aspects of the faces, based on their first impression.”

  • The paper relies on this without support.

    The beauty filter application did not introduce artifacts that affect perceptions.

    “The filters were applied by running the selfie editing app on an Android emulator. An automated clicker loaded the pictures onto the application, applied the filter and then stored the transformed version.”
8 of 8

Statistics recomputed

Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.

  • Consistentz = -12.395reported p < .001 · recomputed p < .001

    “z=−12.395, p < .001”

1 of 1

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Male raters' trustworthiness results conflict after beautification

    Minor

    The results say male raters did not rate male and female images significantly differently on trustworthiness after beautification. The discussion instead says both male and female raters showed significant differences on trustworthiness both before and after beautification.

  • Reported age correlation has the opposite direction to its interpretation

    Minor

    A negative correlation with stimulus age means perceived intelligence and trustworthiness decrease as age increases, but the paper interprets it as older stimuli being perceived as more intelligent and trustworthy.

2 of 2

Cites

Flagged first, then by citations

Research works printed but not linked: 7

  1. Eshiet J. 2020 'Real me versus social media me': filters Snapchat dysmorphia and beauty perceptions among young women. Master’s thesis California State University Long Beach CA. https://scholarworks.lib.csusb.edu/cgi/viewcontent.cgi?article=2243&context=etd.
  2. Bakker M. 2022 Nofilter: how beauty filters affect the internalization of beauty ideals. Master’s thesis Utrecht University Utrecht The Netherlands. https://studenttheses.uu.nl/bitstream/handle/20.500.12932/41835/Thesis_SHOP_JMC_Bakker__6224415_22_06-2022.pdf?sequence=1.
  3. Johns R. 2010 Likert items and scales. Survey Q. Bank: Methods Fact Sheet 1, 11–28.
  4. Christensen RHB. 2018 Cumulative link models for ordinal regression with the R Package ordinal. Submitted in J. Stat. Software 35.https://cran.uni-muenster.de/web/packages/ordinal/vignettes/clm_article.pdf
  5. Mangiafico SS. 2016 Summary and analysis of extension program evaluation in R. New Brunswick, NJ: Rutgers Cooperative Extension.
  6. 10.1037/0022-3514.82.1.19 · DOI
  7. Cotter L. 2011 Self-perceived attractiveness and its influence on the halo effect and the similar-to me effect. PhD thesis Bucknell University Lewisburg PA. https://digitalcommons.bucknell.edu/honors_theses/18.

Cited by

Sorted by citations

Related

Authors
Gulati, A., Martínez-Garcia, M., Fernández, D., Lozano, M. A., Lepri, B., & Oliver, N.
Affiliations
  • University of Alicante (ES)
  • Universitat Jaume I (ES)
  • Universitat Politècnica de Catalunya (ES)
  • University of Alicante, Alicante, Spain
  • Fondazione Bruno Kessler (IT)
  • ELLIS Alicante (ES)
ORCID
Venue
Royal Society Open Science
Publisher
The Royal Society
Published
2024-11
Volume
11
Issue
11
Pages
240882-240882
ISSN
2054-5703 (ISSN-L)
DOI
10.1098/rsos.240882
PMC
PMC11597472
OpenAlex
W4404733393
Semantic Scholar
254aeb9ae4b1b0440fdb079e0b752e47542d6cd8
Cited by
21
References
122
Citations by year
2026: 14 · 2025: 6 · 2024: 1
FWCI
8.63
Type
journal article
Language
en
Open access
Licence
Locations
  • Royal Society Open Science (publishedVersion, cc-by)
  • Research Square (acceptedVersion, cc-by)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion, cc-by)
  • UPCommons institutional repository (Universitat Politècnica de Catalunya) (submittedVersion, cc-by-nc-nd)
  • PubMed Central (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Repositorio Institucional de la Universidad de Alicante (Universidad de Alicante) (submittedVersion, cc-by)
  • arXiv (Cornell University) (cc-by)
  • Zenodo (CERN European Organization for Nuclear Research) (cc-by)
Funders
  • Intel Corporation
  • European Commission (101120237, 13039/501100011033, 501100011033, HORIZON2020, 951847)
  • Generalitat de Catalunya (501100011033, 2021 SGR 01421, PID2019)
  • Generalitat Valenciana (501100011033, CIGE/2022/066, Ministerio de educación y formación profesional, PID2020-118071GB-I00, PID2020-118071GB-I00)
  • Ministerio de Ciencia e Innovación (.13039/501100011033, PID2019, PID2019-104830RB-I00/ DOI, PID2019-104830RB-I00, 501100011033, PID2020118071GB-I00, PID2020)
  • Horizon 2020 Framework Programme (ELIAS: 101120237, ELISE: 951847)
  • Agencia Estatal de Investigación (PID2020-118071GB-I00, 501100011033, 13039, .13039/501100011033, PID2019)
Topics
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Evolutionary Psychology and Human Behavior
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Face Recognition and Perception
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Aesthetic Perception and Analysis
Keywords
beauty filters, facial attractiveness, cognitive biases, AI-generated images, perceived intelligence, perceived trustworthiness, digital decision-making

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-28.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-28.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review