Create account

replied 2373d
Sk8eM dUb
i understand if you've got personal experience (hard to beat) but if you refuse to measure the world around you won't make sense of it and will end up guided by your feelings.
Sk8eM dUb
replied 2373d
My point is that it's impossible to be objective with a wickedly complicated metric such as "crime rate". Are you familiar with p-hacking? https://en.wikipedia.org/wiki/Data_dredging
replied 2372d
It is dogmatic to throw out all studies because they may have data dredging. It is prudent to carefully consider studies individually and look for any flaws in their methodology.
Sk8eM dUb
replied 2372d
Furthermore, it's impossible to give proper weight to the impact of the actions of a few white collar criminals(think Ben Bernanke). e.g. How many died because of the 2008 crash?
replied 2370d
Sk8eM dUb
replied 2369d
Dope, yeah I got a good list going for books I need to read. Is there a book recommendations topic yet?
replied 2369d
haha same. so much reading, so little time. was some interest in DPR Book Club
replied 2371d
there are efforts to estimate government control of the economy including regulation https://www.macdonaldlaurier.ca/files/pdf/MLIPCrossSizeOfGovernmentPaper0514.pdf
Sk8eM dUb
replied 2372d
it's different in every area and sometimes those laws are stupid. So you're not really measuring crime rate, you're measuring *obedience rate*.
Sk8eM dUb
replied 2372d
I completely agree, but the data sets for what we're talking about here depend on corrupt judicial systems reporting data accurately. Then you have to consider what's legal and illegal
replied 2372d
Example on pg 3-4. establishing individualist culture’s causal effect on economic growth https://eml.berkeley.edu/~ygorodni/gorrol_culture.pdf
replied 2372d
There are ways to statistically establish causality https://en.wikipedia.org/wiki/Instrumental_variables_estimation
replied 2372d
"patterns in data that can be presented as statistically significant, without first devising a specific hypothesis as to the underlying causality.”
replied 2372d
yes, p-hacking is an issue (esp with big data). The issue comes from fishing through all relations looking for statistically significant ones regardless of any potential causal basis.