Does red enhance human BONUS Thomas V. Pollet, • As expected, no suggestion of Leonard S. Peperkoorn, performance in contests? publication bias. Robert A. Barton & • No evidence for (strong) Russell A. Hill A meta-analysis heterogeneity. • We also conducted analyses with Proportion Weight Weight tertiles. Study Events Total won by red 95%−CI (fixed) (random) Introduction • These show an effect for close Olympics 1996: Boxing 158 340 0.4647 [0.4124; 0.5178] 6.4% 6.4% • Colour effects argued to exert Olympics 2000: Boxing 139 293 0.4744 [0.4179; 0.5315] 5.6% 5.6% competitions but likely driven by Olympics 2004: Boxing 146 266 0.5489 [0.4888; 0.6076] 5.0% 5.1% outlier (see graph below). (1) Olympics 2008: Boxing 130 266 0.4887 [0.4292; 0.5485] 5.1% 5.2% effects on psychology and Olympics 2012: Boxing 126 236 0.5339 [0.4702; 0.5965] 4.5% 4.7% (2) Olympics 2016: Boxing 132 235 0.5617 [0.4978; 0.6237] 4.4% 4.6% behaviour. Olympics 2000: Taekwondo 38 68 0.5588 [0.4408; 0.6706] 1.3% 1.5% Alternative explanations / Questions / Olympics 2004: Taekwondo 43 75 0.5733 [0.4605; 0.6790] 1.4% 1.6% • Hill & Barton found a ‘winning Olympics 2008: Taekwondo 38 73 0.5205 [0.4078; 0.6312] 1.4% 1.6% moving forward. Olympics 2012: Taekwondo 37 75 0.4933 [0.3833; 0.6040] 1.4% 1.6% red’ effect at the Athens Olympics 2016: Taekwondo 37 73 0.5068 [0.3947; 0.6183] 1.4% 1.6% Olympics 1996: Greco−Roman Wrestling 37 63 0.5873 [0.4641; 0.7005] 1.2% 1.3% • Was Athens (2004), a statistical (2004) Olympics. Olympics 2000: Greco−Roman Wrestling 34 55 0.6182 [0.4859; 0.7351] 1.0% 1.1% Olympics 2004: Greco−Roman Wrestling 24 46 0.5217 [0.3814; 0.6588] 0.9% 1.0% outlier? (Perhaps for close contests but Olympics 2008: Greco−Roman Wrestling 37 81 0.4568 [0.3527; 0.5648] 1.5% 1.7% • Effect more pronounced when Olympics 2012: Greco−Roman Wrestling 41 80 0.5125 [0.4049; 0.6189] 1.5% 1.7% not for overall effect) Olympics 1996: Free−Style Wrestling 36 70 0.5143 [0.3995; 0.6275] 1.3% 1.5% competition was ‘close’. Olympics 2000: Free−Style Wrestling 34 50 0.6800 [0.5413; 0.7930] 0.8% 1.0% • Potential non-independence issues? Olympics 2004: Free−Style Wrestling 27 51 0.5294 [0.3952; 0.6595] 1.0% 1.1% • What is the effect of rule changes? Olympics 2008: Free−Style Wrestling 35 82 0.4268 [0.3254; 0.5348] 1.5% 1.7% Olympics 2012: Free−Style Wrestling 38 79 0.4810 [0.3743; 0.5895] 1.5% 1.7% • Are we underpowered to detect a • Objective: Does red affect the World Boxing Championship 2005 194 374 0.5187 [0.4681; 0.5689] 7.1% 6.9% World Boxing Championship 2007 255 523 0.4876 [0.4450; 0.5303] 9.9% 9.0% ‘noisy’ effect? likelihood of winning in a World Boxing Championship 2009 258 526 0.4905 [0.4480; 0.5331] 10.0% 9.1% World Boxing Championship 2011 277 536 0.5168 [0.4745; 0.5588] 10.2% 9.2% • Should we move toward more studies World Boxing Championship 2013 216 440 0.4909 [0.4445; 0.5375] 8.4% 7.9% large sample of Olympic World Boxing Championship 2015 130 235 0.5532 [0.4893; 0.6154] 4.4% 4.6% which experimentally manipulate colour? For example (5) combat sports / World Boxing Fixed effect model 5291 0.5095 [0.4960; 0.5230] 100.0% −− Random effects model 0.5106 [0.4959; 0.5253] −− 100.0% Championships? Prediction interval [0.4812; 0.5400] Heterogeneity: I 2 = 10%, τ2 = 0.0024, p = 0.31 0.4 0.5 0.6 0.7 Olympics 2004: Boxing Methods Forest plot for all contests Proportion Weight Weight Study Events Total won by red 95%−CI (fixed) (random)

• We included data from 6 World Boxing Championship 2009 Olympics 1996: Boxing 38 72 0.5278 [ 0.4140; 0.6387] 6.2% 6.2% Olympic events (1996-2016) Olympics 2000: Boxing 28 53 0.5283 [ 0.3966; 0.6562] 4.6% 4.6%

Olympics 2004: Boxing 37 58 0.6379 [ 0.5090; 0.7497] 4.6% 4.6% World Boxing Championship 2013

Olympics 2008: Boxing 25 54 0.4630 [ 0.3369; 0.5940] 4.6% 4.6% World Boxing Championship 2007

and 6 World Boxing result Influence on overall

Olympics 2012: Boxing 25 55 0.4545 [ 0.3302; 0.5848] 4.7% 4.7% Olympics 2012: Boxing Olympics 2016: Boxing Championships. Olympics 2016: Boxing 27 44 0.6136 [ 0.4660; 0.7431] 3.6% 3.6% Olympics 2000: Taekwondo 13 23 0.5652 [ 0.3679; 0.7439] 2.0% 2.0% Olympics 2008: Boxing Olympics 2004: Taekwondo 12 17 0.7059 [ 0.4657; 0.8701] 1.2% 1.2% Non• -CognitiveFor Olympic Predictors of events, Student Success: we coded Olympics 2008: Taekwondo 0 3 0.0000 [−0.0560; 0.6175] 0.2% 0.2% Non-Cognitive Predictors of StudentOlympics 2004:Olympics Greco− 2004:Roman Success: Taekwondo Wrestling A Predictive Validity Comparison Between Domestic and A Predictiveg Championship 2015 Validity Comparison Between Domestic and Olympics 2012: Taekwondo 14 25 0.5600 [ 0.3705; 0.7335] 2.1% 2.1% Olympics 1996: Free−Style WrestlingOlympics 2016: Taekwondo InternationalTaekwondo, Students Greco-Roman Internationalics 2000: Greco−Roman Wrestling Students Olympics 2016: Taekwondo 10 15 0.6667 [ 0.4150; 0.8504] 1.2% 1.2% 08:Olympics Free−Style 2004: Wrestling Free−Style Wrestling R0Taekwondooman: FreeTaekwondo −WrestlingStyle Wrestling Olympics 2008: Taekwondo pW 2005restling os2restlingtlingxing011 Olympics 1996: Greco−Roman Wrestling 9 18 0.5000 [ 0.2903; 0.7097] 1.6% 1.6% 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14

Wrestling, Freestyle Wrestling, Olympics 2000: Greco−Roman Wrestling 11 18 0.6111 [ 0.3854; 0.7977] 1.5% 1.5% 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Olympics 2004: Greco−Roman Wrestling 12 17 0.7059 [ 0.4657; 0.8701] 1.2% 1.2% Contribution to overall heterogeneity and Boxing. Olympics 2008: Greco−Roman Wrestling 7 15 0.4667 [ 0.2480; 0.6989] 1.3% 1.3% Baujat plot for close contests (quartiles)) Olympics 2012: Greco−Roman Wrestling 9 17 0.5294 [ 0.3096; 0.7384] 1.5% 1.5% • Following Hill & Barton, we Olympics 1996: Free−Style Wrestling 8 19 0.4211 [ 0.2311; 0.6376] 1.6% 1.6% Olympics 2000: Free−Style Wrestling 5 8 0.6250 [ 0.3038; 0.8651] 0.6% 0.6% agreed upon a pre-specified Olympics 2004: Free−Style Wrestling 4 10 0.4000 [ 0.1671; 0.6884] 0.8% 0.8% Olympics 2008: Free−Style Wrestling 9 15 0.6000 [ 0.3567; 0.8025] 1.2% 1.2% protocol for close contests, Olympics 2012: Free−Style Wrestling 10 18 0.5556 [ 0.3370; 0.7546] 1.5% 1.5% World Boxing Championship 2005 35 68 0.5147 [ 0.3983; 0.6295] 5.9% 5.9% and selected the first quartile World Boxing Championship 2007 50 103 0.4854 [ 0.3912; 0.5807] 8.9% 8.9% World Boxing Championship 2009 57 118 0.4831 [ 0.3949; 0.5723] 10.2% 10.2% World Boxing Championship 2011 71 135 0.5259 [ 0.4421; 0.6083] 11.6% 11.6% of closest matches. World Boxing Championship 2013 55 113 0.4867 [ 0.3965; 0.5778] 9.7% 9.7% • We used the 'meta' package(4) World Boxing Championship 2015 39 70 0.5571 [ 0.4408; 0.6675] 6.0% 6.0% Fixed effect model 1181 0.5245 [ 0.4957; 0.5531] 100.0% −− to synthesise our findings Random effects model 0.5245 [ 0.4957; 0.5531] −− 100.0% Prediction interval [ 0.4942; 0.5545] (note that other analyses Heterogeneity: I 2 = 0%, τ2 = 0, p = 0.83 including Bayes Factors have 0 0.2 0.4 0.6 0.8 also been used but are not Forest plot for close contests (quartiles) Take a picture to reported here). download the poster

(Key) Results

• See Figures. References [1] Elliot, A.J. & Maier, M.A. (2012). “Color-in- Context Theory”. In: Advances in Experimental Discussion Social Psychology. Ed. by A Devine and P Plant. Vol. 45. Academic Press, pp. 61–125. doi: • No statistically significant 10.1016/B978-0-12-394286-9.00002-0. effects of colour (but [2] Elliot, A.J. & Niesta, D. (2008). Romantic Red: Red Enhances Men’s Attraction to Women. Journal suggestion of small effect of Personality and Social Psychology 95(5), 1150– 1164. doi: 10.1037/0022-3514.95.5.1150. when competition is close). [3] Hill, R.A. & Barton, R.A. (2005). Red Enhances Human Performance in Contests. Nature 435, 293. • Alternative explanations? doi:10.1038/435293a. (Time?, Small but noisy effect?, Picture by Boxing AIBA, https://commons.wikimedia.org/wiki/File:Rio_2016_Olympic_Games_-_Day_3_(28777000891).png [4] Schwarzer, G. Meta: Meta-Analysis with R. R Package Version 4.4–0. .((2008 J. Leißing, & B., Strauss, N., Hagemann, statistical outlier(s)?) [5] Contact: When the Referee Sees Red …. Psychological 10.1111/j.1467- doi: 19(8):769–771. [email protected] Science 9280.2008.02155.x @tvpollet