|
Here is a composite plot that looks at the data splitting for netvar
alternate seconds, minutes and hours. This seems to show the following:
There is a difference between pre and post for all 3 blockings.
The post is at the .002 - .001 level for all 3.
This accords with our main result.
Two things that don't fit a neat picture are:
1. The post 9/11 correlations seems to weaken at the blocking is longer.
But careful. This may be insignificant or just stats.
2. The pre-9/11 correlation at minute-blocking seems to have a
significant correlation on timescales 2-4 weeks (= "n" index of 35 thru' 80).
Clearly there's more thinking to do. But basically this is very good
news. The post 9/11 correlation is robust at the .001 level for alternate
blockings all the way out to 3600 seconds.
Important to note that the post 9/11 level obtains for all blockings.
The significant pre-9/11 result at minute blocking is
destroyed if the blocking is 1-sec or 1-hour.
|