Timeline for OLS vs WLS in a Heteroskedastic Trending Regression
Current License: CC BY-SA 4.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 14 at 14:34 | vote | accept | Joseph Basford | ||
| Jan 5 at 9:43 | comment | added | Joseph Basford | @BigBendRegion Unfortunately, I was constrained in my approach as this is from a past paper with a few subparts which guide the answer. I did manage to adapt what you said to find the problem and get a solution. Thanks! | |
| Jan 5 at 9:25 | answer | added | Joseph Basford | timeline score: 1 | |
| Jan 4 at 19:19 | comment | added | BigBendRegion | Don't know where your mistake is, but there is no need for normalization or asymptotics here. The finite sample variance of WLS is smaller than that of OLS, simply by the G-M theorem. That should be apparent as well by inspection of the function forms of the finite sample variances, and an application of some inequality like C-S. | |
| Jan 4 at 14:36 | history | asked | Joseph Basford | CC BY-SA 4.0 |