I will present the paper 'Error Bars Considered Harmful' from InfoVis 2014. It argues why error bars don't accurately convey statistical information, and presents three alternative representations. Interestingly, the authors ran a rather sizable experiment to demonstrate that using different representations for the error gave the participants a better understanding of the statistics and how confident they should be in the data. I'll also show a few other visual representations of statistical uncertainty found in the wild.