Comment by philiphodgen | Hacker News Reader

Comment by philiphodgen | Hacker News Reader

Parent Full thread

I'm sorry, but there seems to be an omitted first step in the entire comment thread of critiques -- from all comments coming from all political angles: examining the underlying data.

The NYT graphic tells a terribly compelling story that supports a particular political point of view. That tickles my bullshit sensor, and tells me it is time to question the underlying data and assumptions.

Frankly, this graphic makes me hear the sound of ideological axes grinding. That makes me want to question assumptions, data gathering, and data analysis. This is all thanks to Prof. Eyrich and his Stats and Econometrics classes way back when. (w00t Claremont McKenna College.)

I am not saying the graphic is wrong or it is right in the conclusions it suggests. I just find it telling that the discussion thread jumps straight to an Itchy and Scratchy Show level of discourse, without questioning the underlying data.

Where is HN's passion for analytic rigor? Or does A/B testing and statistical analysis apply to websites but not economics?

> I am not saying the graphic is wrong or it is right in the conclusions it suggests. I just find it telling that the discussion thread jumps straight to an Itchy and Scratchy Show level of discourse, without questioning the underlying data.

So then tell us what's wrong with the data. I assume most people didn't question the data because income inequality since the eighties is rather well documented.

That you would call the data into question without providing any justification suggests you have some ideologies of your own.

While the political story did raise a few flags, that's not what most makes me want to look at the data.

What really surprised me is that both main trends are linear. I presumed both productivity and wages would be exponential. I can believe that the wages are inflation-adjusted, though I saw no mention of this (except in the quintile section), but that productivity would have the same base is unexpected.

It's kind of hard to A/B test an economy.

We're pretty good at it now. A=Republicans, B=Democrats. We switch every couple of election cycles and see how it goes.

It's interesting how the low point on the graph is 1976, but they put the categorical divide at 1980.

Reply on news.ycombinator.com