# Detecting the gorilla: Results summary Here we report on some supplementary analyses that were not included in the manuscript for this study. The data and analytic script can be found in an rMarkdown document in this project component. When we preregistered this study, we proposed three different criteria for what counted as detecting the gorilla. We ended up reporting primarily on the most lenient criteria as it captured those who may not have been willing to mention the gorilla during the trial itself, but did later when prompted by follow-up questions. Here are the data for each criteria. ## Lenient criteria With this criteria, any participant who mentioned the gorilla during the trial or said they noticed the gorilla in response to any of the three follow-up questions was coded as having noticed the gorilla. #### Separate chi square analyses In our preregistration we stated that we would analyse each condition separately using a chi-square test. An equal number of experts (6/20) and novices (4/20) noticed the local gorilla, χ2 = 0.533, *p* = .465 (Cohen's w = 0.12). Fewer experts (2/20) detected the gorilla than novices (9/20) in the global condition χ2 = 6.144, *p* = .013 (cohen's w = 0.39). #### Binomial logistic regression In running a binomial logistic regression we found no significant main effects, but we did find a significant interaction between Expertise and Condition, *b* = -.634, *p* = .027. ## Medium criteria With this criteria, any participant who mentioned the gorilla during the trial, or said they noticed the gorilla in response to either of the first two follow-up questions was coded as having noticed the gorilla. #### Separate chi square analyses In running a chi-square analyses separately for each condition, we found that experts (4/20) and novices (2/20) detected the local gorilla at equal rates, χ2 = 0.784, *p* = .376 (Cohen's w = 0.14), but fewer experts (2/20) than novices (8/20) noticed the gorilla in the global condition, χ2 = 4.800, *p* = .028 (Cohen's w = 0.35). #### Binomial logistic regression In testing for an interaction, we found no main effects but there was a significant interaction, *b* = -.651, *p* = .042. ## Strict criteria With this criteria, only participants who mentioned the gorilla during the ciritical trial were coded as having noticed the gorilla. #### Separate chi square analyses Chi-square analyses for each condition separately revealed that experts (3/20) and novices (2/20) detect the gorilla in the local condition, χ2 = 0.229, *p* = .633 (Cohen's w = 0.08), and an equal number of experts (2/20) and novices (5/20) detected the gorilla in the global condition, χ2 = 1.558, *p* = .212 (Cohen's w = 0.20). #### Binomial logistic regression We found no significant main effect nor an interaction (*p*s < .05) when we conducted a binomial logistic regression using the strict criteria.