4 ms·
The graphs start/end at the 20th/~85th percentiles rather than 0th/100th, which makes the differences seem larger than they are.
by sophistication 8y ago
The graphs start/end at the 20th/~85th percentiles rather than 0th/100th, which makes the differences seem larger than they are.
- marcosdumay 8y agoThe graph starts at 50% where there is a thick line. It grows up and down until it contains the entire variation. The extreme values are not relevant. It is a pretty standard format, and this "warning the graph does not start at 0" kneejerk is tiring. (What would you expect a graph with both positive and negative values should look like?)
- sophistication 8y agoIf there would be no data point below/above 20/85, then it might be warranted to truncate it, but that's clearly not the case, in fact, by definition, there is 20% and 15% of the data respectively. The bars starting from the median (50th percentile) are also a pretty meaningless visual indicator. E.g. the bars for competition visually indicate that men would be twice as competitive, but in the graph without the bar and including the 0th/100th percentile the difference is much less pronounced: https://quanticfoundry.com/wp-content/uploads/2018/08/hardcore_comparison-650x326.png https://quanticfoundry.com/wp-content/uploads/2018/08/hardco... https://i.imgur.com/P9M7AnR.png https://i.imgur.com/P9M7AnR.png I find that especially damning as lots of these gamers are probably not familiar with statistics, so they might indeed draw extreme conclusions from these graphs merely due to the way it is represented.