Fwd: Latest Presidential Election Numbers from www.gallup.com

[Brad DeLong]

>  > ----------
>>  From:	Brad DeLong[SMTP:delong at econ.Berkeley.EDU]
>>  Reply To:	lbo-talk at lists.panix.com
>>  Sent:	Monday, October 23, 2000 7:07 PM
>>  To:	lbo-talk at lists.panix.com
>>  Subject:	Re: Fwd: Latest Presidential Election Numbers from
>>  www.gallup.com
>>
>>  It's been a long time since I've advocated sending *anybody* to the
>>  re-education camp. But a close of the Bush - Gore margin from 11% to
>>  2% in two days is a three-standard-deviation move. This isn't the
>>  first or the second three-standard-deviation move in the Gallup poll
>>  either...
>>
>	---
>
>	And it's a 3-day moving average. Which means the one-day jump was
>even bigger, right?
>
>	Seth

Yep. You drop the old and add the new observation.

Unfortunately, you can't figure out what each day's is unless you 
know at least one of the starting three day results. The inverse 
transformation is singular...

Brad DeLong