What Google Search Trends Can Tell Us the Rise of White Nationalism

lawsuit.org

5 points·kristintynski··6 comments

If you need to rely on stats for Google search terms as opposed to observing real life, then you can safely say there isn't a rise.

> If you need to rely on stats for Google search terms as opposed to observing real life, then you can safely say there isn't a rise.

We don't "need to rely on stats for [sic] Google" to, though.

[sic] Seriously?

That scatterplot of random data with a clean line through it is practically art. Statisticians will draw a line through a plot of noise and infer that the slope shows correlation. The race war thing was a conservative meme for a time. There may be some correlation there. Racial bias is a spectrum independent of political affiliation. Liberals do like to ride a high horse though.

I mean, if you don't like the slope of the line on county level (included in the article), the state level lines (click on the graph to get to the tableau) have a stronger slope. But overall, you're accusing the author of bias in cherry picking the data or you're doubting statistics because the data is significant but not eye poppingly so. Really, the weakest link is the title - searching for race war isn't an act of white nationalism all the time, and may be searched by both pro and anti factions to understand why it is trending.

Overall, I thought it was really cool to not just publish the data, but also present the graphics in a way you can confirm yourself.

Only a fool would trust statistics. Confidence is proportional to the p-value. False positives are very common. The author acknowledged that the results are not significant.

Reply on news.ycombinator.com