A reviewer asked me to look at possible heteroscedasticity (HS) in some models I ran. It took me on a detour that ended with making some new convenience functions for resting for HS, whether linear or nonlinear. HS is when the variance of the residual depends on predictor values; i.e., the amount of error is a function of the predictors. Errors are supposed to be totally random normal values. HS biases the standard errors of the model, though does not normally result in bias in the effect sizes/slopes (I think).
R notebook for https://mankindquarterly.org/upcoming/issue/60-4/5
plot data from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4584427/
R notebook for http://emilkirkegaard.dk/en/?p=5263
Are there any effects of active video games on cognitive functioning? a reanalysis of Stanmore et al (2017)
Orig. study: http://www.sciencedirect.com/science/article/pii/S014976341730129X
Baggrund for http://teknologipartiet.dk/2017/02/23/indvandrere-opsummering/
Visual explanation of the effects of excluding lower scoring students from PISA samples.
Visualizations based on Charles Murray's great book, Human Accomplishment.
Commentary on Wicherts 2016. http://www.sciencedirect.com/science/article/pii/S0160289616303233
Examples of ggplot2 helpers functions in kirkegaard package.
An example of a system for translating names of political units to ISO names, and using them to join datasets.
Joining tables based on fuzzy string matches in an optimal way.
It has been claimed that gender differences in academic fields results from stereotypes about how difficult the fields are (Leslie, 2015). This is based on the finding that perceptions of brilliance requirement of different fields correlated strongly with the gender distribution. It apparently did not dawn on the people making this claim that the stereotypes were probably very accurate: people who study physics really are smarter than those who study psychology. This knitr visualizes data from a paper by Templer and Tomeo from 2002.
A brief tutorial for the choroplethr package.
A showcasing of some string functions in the Kirkegaard package.
A comment on Gwern's analysis.