Eating cheese will make you die by bedsheet...

[OrangeDragon]

And other great spurious correlations:

http://www.tylervigen.com/

[Rain Dog]

And other great spurious correlations:

http://www.tylervigen.com/

I love this stuff as it brings out my "inner stat nerd"

Actually such relationships in Data are often not spurious -- it's just that people are not creative enough to interpret them.

It can be fun (and sometimes even useful) to speculate on "why" there might be a relationship between two variables (rather than just assume it is based on random error of some type).

For example ---

is "Per Capita Consumption of Cheese" correlated with "Obesity" which in turn is correlated with "Sleep Apnea and other sleeping disorders" Which in turn might result in episodes of "fits while sleeping" which in turn might show up in your "death by bedsheets strangulation" stats.

Actually when you see "big" shifts in two apparently unrelated variables, the cause is more often than not an unreported or undocumented changes in data collection, classification, or analytical methodologies. For example -- maybe year to year changes in Budgeting affect how well agencies document certain variables (like bedsheets strangulation) causing seeming shifts in apparently unrelated variables I would bet that a lot of the supposed "spurious" relationships on this site have plausible explanations behind them. Unfortunately you would need full access to understanding the methodologies used and assumptions made.

Anyway, it's "food for thought". (I know, rather cheesy of me -- no point wine-ing about it -- it' is just my dairy reflection on an eclaretic topic) --- Sorry Vlad

[Kuroneko]

And other great spurious correlations:

http://www.tylervigen.com/

I love this stuff as it brings out my "inner stat nerd"

Actually such relationships in Data are often not spurious -- it's just that people are not creative enough to interpret them.

It can be fun (and sometimes even useful) to speculate on "why" there might be a relationship between two variables (rather than just assume it is based on random error of some type).

For example ---

is "Per Capita Consumption of Cheese" correlated with "Obesity" which in turn is correlated with "Sleep Apnea and other sleeping disorders" Which in turn might result in episodes of "fits while sleeping" which in turn might show up in your "death by bedsheets strangulation" stats.

Actually when you see "big" shifts in two apparently unrelated variables, the cause is more often than not an unreported or undocumented changes in data collection, classification, or analytical methodologies. For example -- maybe year to year changes in Budgeting affect how well agencies document certain variables (like bedsheets strangulation) causing seeming shifts in apparently unrelated variables I would bet that a lot of the supposed "spurious" relationships on this site have plausible explanations behind them. Unfortunately you would need full access to understanding the methodologies used and assumptions made.

Anyway, it's "food for thought". (I know, rather cheesy of me -- no point wine-ing about it -- it' is just my dairy reflection on an eclaretic topic) --- Sorry Vlad

I think you would enjoy studying Epidemiology. Heres a paper on using stats to identyfy causal relationships
http://www.stat.duke.edu/~jerry/Papers/causal.pdf
And heres a link to the CDC site where you can download Epi Info 7 a great piece of free software. "The program allows for electronic survey creation, data entry, and analysis. Within the analysis module, analytic routines include t-tests, ANOVA, nonparametric statistics, cross tabulations and stratification with estimates of odds ratios, risk ratios, and risk differences, logistic regression (conditional and unconditional), survival analysis (Kaplan Meier and Cox proportional hazard), and analysis of complex survey data."
http://wwwn.cdc.gov/epiinfo/7/
Who hoo

[Rain Dog]

And other great spurious correlations:

http://www.tylervigen.com/

I love this stuff as it brings out my "inner stat nerd"

Actually such relationships in Data are often not spurious -- it's just that people are not creative enough to interpret them.

It can be fun (and sometimes even useful) to speculate on "why" there might be a relationship between two variables (rather than just assume it is based on random error of some type).

For example ---

is "Per Capita Consumption of Cheese" correlated with "Obesity" which in turn is correlated with "Sleep Apnea and other sleeping disorders" Which in turn might result in episodes of "fits while sleeping" which in turn might show up in your "death by bedsheets strangulation" stats.

Actually when you see "big" shifts in two apparently unrelated variables, the cause is more often than not an unreported or undocumented changes in data collection, classification, or analytical methodologies. For example -- maybe year to year changes in Budgeting affect how well agencies document certain variables (like bedsheets strangulation) causing seeming shifts in apparently unrelated variables I would bet that a lot of the supposed "spurious" relationships on this site have plausible explanations behind them. Unfortunately you would need full access to understanding the methodologies used and assumptions made.

Anyway, it's "food for thought". (I know, rather cheesy of me -- no point wine-ing about it -- it' is just my dairy reflection on an eclaretic topic) --- Sorry Vlad

I think you would enjoy studying Epidemiology. Heres a paper on using stats to identyfy causal relationships
http://www.stat.duke.edu/~jerry/Papers/causal.pdf
And heres a link to the CDC site where you can download Epi Info 7 a great piece of free software. "The program allows for electronic survey creation, data entry, and analysis. Within the analysis module, analytic routines include t-tests, ANOVA, nonparametric statistics, cross tabulations and stratification with estimates of odds ratios, risk ratios, and risk differences, logistic regression (conditional and unconditional), survival analysis (Kaplan Meier and Cox proportional hazard), and analysis of complex survey data."
http://wwwn.cdc.gov/epiinfo/7/
Who hoo

UhOh Porn for stat nerds -- woo hoo indeed

Yea I worked with some of these predictive tools before although I am not familiar with the specific links you just gave me and will definitely take a look. Thanks!

[OrangeDragon]

This thread is getting awesome!

[Kuroneko]

This thread is getting awesome!

CEO goes beyond "the thinking mans forum"

[General Mackevili]

LoL!