Talk:Proxy (statistics)

Typo or example of singular usage?
The line as a prox for gender seems to be a typo but I am not familiar with the term so I don't know. Perhaps someone can verify whether or not the word should read as prox rather than proxy. BigNate37T·C 06:42, 17 July 2006 (UTC)

You were right BigNate, it was a typo. There is no "prox" in statistics.

A better example?
What was the person who wrote the "sex as a proxy for gender" phrase thinking? This makes very little sense, statistical or otherwise.


 * I second this. For most research we're not interested in "gender" vs "sex" and people with nonstandard gender identification don't necessarily cluster with either gender/sex. Taw (talk) 23:44, 4 December 2008 (UTC)

Proxy (climate) and Proxy (statistics)
Barring any objections I'll move the example from proxy {climate) here and set a redirect. ~ trialsanderrors 19:28, 20 March 2007 (UTC)


 * Please don't. They are different enough to not merit merging. And there are plenty of other climate proxies that ought to be added William M. Connolley 20:19, 20 March 2007 (UTC)
 * I don't see where they differ. Proxies are observable independent variables that are used as stand-ins for unobservable causal variables under the assumption that causal variable and proxy are highly correlated, and both articles state exactly that. We could start articles on proxy (political economy) and proxy (biology) offering field-specific examples, but the would simply be confusing and counterproductive. The concept of a proxy is the same in empirical research, no matter the field. ~ trialsanderrors 20:51, 20 March 2007 (UTC)


 * "different" was the wrong word. Its well enough used in climate to need its own article. There is far more that ought to be added - tree rings; coral; etc. If there are useful proxies in bio, then yes they probably do want their own articles William M. Connolley 21:17, 20 March 2007 (UTC)


 * I agree with WMC, the climate proxies are significant and notable enough for the independant article. If other fields use proxies in a sufficiently notable manner then they too, should have independant articles. Therefore - no merge. Vsmith 00:32, 21 March 2007 (UTC)
 * I'm unconvinced, but I'm not particularly inclined to argue about it. I've removed the merge headers. ~ trialsanderrors 18:09, 21 March 2007 (UTC)

Rename to "proxy variable"?
Would this article be better named proxy variable? Nurg (talk) 22:50, 9 October 2008 (UTC)

Is this a good definition?
I have to challenge the current definition: "In statistics, a proxy variable is something that is probably not in itself of any great interest, but from which a variable of interest can be obtained." as I find it misleading, uninformative and somewhat misinformative. In applied econometrics (I do not know on the usage of proxy variables in other areas) it is a variable that substitutes an unobservable or immesurable variable. It is true that the proxy used in place of the unobservable variable is not of great interest but its usage facilitates the construction and estimation of the model. It is not true that the variable of interest can be obtained through the proxy variable. Take for example risk-free interest rate in the area of empirical finance. It is frequently proxied by short-term interest rates on securities issued by the government (such as the 3-month U.S. Treasury bill interest rate), and this is done because the risk-free rate is directly unobservable (and to a large extent a fiction) and it is not true that this interest rate (e. g. the 3-month U.S. Treasury bill interest rate) may help recover or obtain the true unobservable risk-free rate. Similarly, under the currently displayed example: "Per-capita GDP is often used as a proxy for measures of standard of living or quality of life." One cannot obtain a good measure of standard of living or quality of life from the GDP per capita. If it were so, one would not employ the GDP per capita but directly the measure of standard of living or quality of live instead. — Preceding unsigned comment added by 195.168.229.59 (talk) 22:07, 5 August 2011 (UTC)

Merge to top-level article Proxy (science)
This article and Proxy (climate) should be merged into a new article, Proxy (science).

Proxy methods are used all over in science. While proxies used for paleoclimate reconstructions are probably the most controversial recently, they are used in many more sciences--http://www.nature.com/nature/journal/v402/n6759/abs/402288a0.html, for example-- and there should just be one article, not many. Geoffrey.landis (talk) 22:42, 12 July 2015 (UTC)