ORIGINAL PAPER Emotional faces in neutral crowds: Detecting displays of anger, happiness, and sadness on schematic and photographic images of faces Ottmar V. Lipp Sarah M. Price Cassandra L. Tellegen Published online: 7 August 2009 Ó Springer Science+Business Media, LLC 2009 Abstract Detection of angry, happy and sad faces among neutral backgrounds was investigated in three single emotion tasks and an emotion comparison task using schematic (Experiment 1) and photographic faces (Exper- iment 2). Both experiments provided evidence for the preferential detection of anger displays over displays of other negative or positive emotions in tasks that employed all three target emotions. Evidence for preferential detec- tion of negative emotion in general was found only with schematic faces. The present results are consistent with the notion that the detection of displays of anger, and to some extent sadness, does not reflect on a pre-attentive mecha- nism, but is the result of a more efficient visual search than is the detection of positive emotion. Keywords Facial expression of emotion Á Visual search Á Anger superiority effect The ‘face-in-the-crowd effect’, first reported by Hansen and Hansen (1988) is one of the most persuasive and best known findings in psychology that remains highly cited (273 citations since publication in an ISI based search conducted July 14th, 2009). This is noteworthy given that subsequent research has not only challenged the strongest claim forwarded by Hansen and Hansen, namely that detection of an angry face in a crowd of happy faces is automatic and not affected by crowd size (angry face pop- out), but also the weaker claim that angry targets are found faster among happy backgrounds than vice versa (Purcell et al. 1996). Whereas the resilience of the ‘face-in-the- crowd’ effect against inconsistent evidence is an interesting psychological phenomenon in itself, the question remains as to whether there is preferential detection for facial expressions of threat or negative emotion. A satisfactory answer to this question is made difficult by the heteroge- neity of the stimulus materials (schematic vs. photographic faces; black and white vs. grey scale images; number of different emotions used), the stimulus arrangements (within vs. between task comparison of different emotions; emotional vs. neutral control backgrounds; circular, ran- dom, or stimulus grid arrangements) and the tasks used in previous research (variable mapping of target and back- grounds vs. fixed target search; see Horstmann 2007; and Frischen et al. 2008; for recent reviews). The present research was designed to extend this work by using pho- tographic images as well as schematic faces and by com- paring the search for angry and sad facial expressions to that for happy faces. These comparisons were made in a set of two experiments that employed three variable target mapping search tasks in which a single emotion was pre- sented together with neutral expressions (angry, happy, and sad tasks) and one fixed target mapping search task in which all three emotional expressions were used as targets (emotion comparison task). Hansen and Hansen (1988, Experiment 3) presented participants with matrices of four or nine pictures of either a male or a female model, drawn from the Ekman and Friesen (1976) database, who displayed either a happy or an angry facial expression. For each model separately, the displays were either all of angry or happy faces (No target trials) or of one angry/happy target among three or eight happy/angry backgrounds (Target trials). Participants were faster to search happy crowds than angry crowds on No target trials and found angry targets faster than happy O. V. Lipp (&) Á S. M. Price Á C. L. Tellegen School of Psychology, University of Queensland, Brisbane, QLD 4072, Australia e-mail: o.lipp@psy.uq.edu.au 123 Motiv Emot (2009) 33:249–260 DOI 10.1007/s11031-009-9136-2