Timeline for US Election results 2016: What went wrong with prediction models?
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 10, 2016 at 23:38 | comment | added | colin | my response to comments was necessarily a bit long, so added it as a new answer | |
| Nov 10, 2016 at 4:31 | comment | added | Cliff AB | I don't necessarily disagree with a lot of your points. But I just want to point out that in the case of polling, I think every pollster is extremely aware of the limitations due to data quality, but don't really have any options to improve it (see my answer). Your answer seems to suggests that pollsters want push out any answer, not caring at all about data quality. I think pollster care a lot about data quality, but also realize that the best they can get has serious potential flaws. Do you give up ("50%-50%!") or try to build something that might be reasonable? | |
| Nov 10, 2016 at 1:23 | comment | added | horaceT | Taken far afield I'm sure, can you give examples of the questionable researcher constructs. | |
| Nov 10, 2016 at 0:03 | history | edited | colin | CC BY-SA 3.0 |
added 433 characters in body
|
| Nov 9, 2016 at 23:46 | history | answered | colin | CC BY-SA 3.0 |