Statistical Modeling, Causal Inference, and Social Science: Peeking behind the curtain, or, What's (not) the matter with Portugal?

Peeking behind the curtain, or, What's (not) the matter with Portugal?

This is pretty embarrassing, but I think it's better to tell all, if for no other reason than to make others aware of the challenges of working with data . . .

Peter

What about Israel?

It seems to be the only one where the rich vote more liberally than the poor. But, with so many parties there, I'm curious as to how you rank them.

LemmusLemmus

I think this illustrates a common problem with data analysis: Suspicious results get double-checked more often. Yes, you can call it Bayesian, but it still biases the analysis in favour of unsuspicious results (which are less likely to get double-checked).

(Don't get me wrong, I'm not blaming you here. I still remember quite vividly the afternoon I discovered that all our previous analysis were plain wrong because we had forgotten to recode some of the missings as zeros - and that was two days before the report was due. I came into the office at six in the morning the next day and then the computers didn's work - but that's another story.)

GDP Guy

What's the matter with the GDP/capita order?

Peter, Israel, Ireland, Portugal and Brazil have the same pattern for the rich.

Kaiser

Lemmus^2: can call this "narrative bias" or "optimism bias". that's why Andrew said "we better check all the party names carefully now".

James Annan

You might also consider that a small compensation for the publication bias that favours "exciting new results".