Skip to main content
13 events
when toggle format what by license comment
Feb 28, 2017 at 0:00 vote accept FarrukhJ
Feb 10, 2017 at 15:13 comment added Richard Hardy Perhaps more importantly, the OLS estimator of the coefficient on a linear trend converges a whole order of magnitude faster (at a rate $n^{-3/2}$) to its true value than for stationary regressors ($n^{-1/2}$), which means you can consistently estimate the trend even if you neglect the stationary variables. This is in contrast to estimating the effects of stationary variables one by one, where you lose consistency if you omit variables.
Feb 10, 2017 at 14:21 comment added Ben Bolker I don't have the time to write a proper answer/document this, but in general serial correlation does not bias the results of a linear regression (it alters the appropriate computation of the standard errors, confidence intervals, etc.). This makes the classic two-stage approach (detrend, then analyze for correlation) sensible. (e.g. some googling of "serial correlation linear regression unbiased leads to fmwww.bc.edu/ec-c/f2010/228/EC228.f2010.nn12.pdf )
S Feb 10, 2017 at 8:10 history suggested smci CC BY-SA 3.0
clarify question
Feb 10, 2017 at 8:06 history tweeted twitter.com/StackStats/status/829964730609577984
Feb 10, 2017 at 8:01 comment added Matthew Gunn What do you mean precisely by detrend?
Feb 10, 2017 at 7:51 review Suggested edits
S Feb 10, 2017 at 8:10
Feb 10, 2017 at 7:34 answer added DeltaIV timeline score: 3
Feb 10, 2017 at 7:27 answer added Matthew Gunn timeline score: 15
Feb 10, 2017 at 6:00 answer added G_B timeline score: 6
Feb 10, 2017 at 5:36 comment added Christoph Hanck It is not generally true that we make the assumption that the "data" is i.i.d.
Feb 10, 2017 at 4:54 review First posts
Feb 10, 2017 at 6:00
Feb 10, 2017 at 4:51 history asked FarrukhJ CC BY-SA 3.0