We do it every day: correctly classifying people, objects, words. What seems easy for humans is computationally very complex. Together with Jonathan D. Nelson and Björn Meder (both Center for Adaptive Behavior and Cognition, MPIB Berlin) I hypothesized and found that a relatively simple statistical principle called class-conditional independence (aka naive Bayes) describes human classification learning.
Human behavior changes across contexts like family, health, or job situations. I measure and model whether the aspects of situations determine our attitudes towards risk; specifically how our attitudes change across ten evolutionary domains. Together with Andreas Wilke (Clarkson University, NY, USA), I ask which processes underly domain differences.
People often state that they prefer one thing over the other. Whether these preferences are learned, constructed, discovered, or something completely different, is unclear. We ask: What is the cognitive process underlying preference formation? And we have a pretty neat idea (with Jörg Rieskamp at the University of Basel).
Situational Attributes for Risk Taking How the retrieved attributes relate to domain differences in risk taking was analyzed in an exploratory fashion because … The tables can be found in Appendix A of the paper. Setup library(data.table) # fantastic and fast data manipulation, see library(scales) # for percent foramt library(lsr) # for cohens D library(coin) # for wilcox_test library(ltm) # for point biseral correlations # Helper functions: download the utils folder and maybe change the path path argument sapply(list.files(path="utils/", pattern="*.R", […]
Replication of Wilke (2014) Eight of 10 main effects of domain and gender have the same direction as in Wilke et al. (2014). Eight of nine effects had the same direction in both studies, and the ninth was small (and insignificant) in both studies The tables can be found in Appendix A of the paper. Setup # Libraries & Directories library(reshape2) #To change from long to short format library(data.table) #To handle data.frames more efficiently library(lsr) #For effect sizes library(pwr) library(xtable) […]