Hi all,
Does anyone know about a statistical test which will indicate which points
on a time series chart are significantly different than the rest, where
points are rolling averages. Does anyone know of any sig test on moving
average data?
Example...
If we have a tracking study where we conduct 50 interviews every week and
are viewing this continous data on a four weekly roll i would have data
points like:
1. Week 1 to Week 4 : data point (Y=) 5%
2 Week 2 to Week 5 : (Y=) 7%
(Base for each data point = 200)
Is there a statistical test to prove that Y of 5% is not significantly
lower than Y 7% with a CI of 95%.
regards
Vince
|