
Greg Hennessy wrote:
> On 20060325, Thomas Smid <thomas.smid@xxxxxxxxx> wrote:
> > I had also a closer look now at the new polarization data as given in
> > http://map.gsfc.nasa.gov/m_mm/pub_papers/polarization/wmap_3yr_pol.pdf
> > ,and this confirms in my opinion that one is merely looking at the
> > effect of statistical errors here: with the relative apparent intensity
> > (temperature) fluctuations of the CMB being of the order of 10^4, one
> > would expect the polarization due to a statistical error be a further
> > factor sqrt(10^4)=10^2 smaller,
>
> I would not expect that. The CMB polarization is not due to
> statistical error, but the percentage polarization of the CMBR is
> about 7 percent.
Where did you get the 7% figure from?
Looking at the 4 data points (green) for the polarization spectrum in
http://lambda.gsfc.nasa.gov/product/map/current/map_images/grandspectrum512.png
and comparing these to the intensity spectrum (black data points at the
top) for the same lvalue, I get the following ratios (going from left
to right):
0.4/40=1% (taking the top of the error bar),
0.3/30=1% (taking the top of the error bar),
0.4/40=1% (taking the middle of the error bar),
1/80=1.25% (taking the bottom of the error bar).
Thomas

