Table 2
Testing Null Hypothesis I.
| CMB Removal | Implementation 1 | Implementation 2 | ||||||
|---|---|---|---|---|---|---|---|---|
| PR3 | SRoll2 | PR3 | SRoll2 | |||||
| diff. | p-value | diff. | p-value | diff. | p-value | diff. | p-value | |
| None | 0.22 ± 0.02 | 7 × 10−34 | 0.28 ± 0.02 | 2 × 10−48 | 0.19 ± 0.017 | 4 × 10−29 | 0.25 ± 0.018 | 6 × 10−44 |
| commander | 0.20 ± 0.02 | 7 × 10−34 | 0.24 ± 0.02 | 4 × 10−48 | 0.17 ± 0.015 | 1 × 10−28 | 0.21 ± 0.016 | 1 × 10−40 |
| nilc | 0.20 ± 0.02 | 2 × 10−35 | 0.26 ± 0.02 | 2 × 10−54 | 0.18 ± 0.015 | 4 × 10−32 | 0.23 ± 0.016 | 6 × 10−49 |
| sevem | 0.19 ± 0.02 | 3 × 10−33 | 0.24 ± 0.02 | 7 × 10−46 | 0.17 ± 0.015 | 2 × 10−29 | 0.21 ± 0.016 | 1 × 10−40 |
| smica | 0.20 ± 0.02 | 4 × 10−36 | 0.25 ± 0.02 | 3 × 10−49 | 0.18 ± 0.015 | 9 × 10−32 | 0.22 ± 0.016 | 5 × 10−45 |
Notes. Probability distribution of the difference of 
values computed for control and target, with their uncertainties, computed from the sampling uncertainties in
and
, in turn obtained through unweighted bootstrapping. The one-sided p-value gives the probability that 
. Results are given for both the PR3 and SRoll2 polarization maps; for removal of the CMB polarization as estimated from the different Legacy component-separation methods, as well as for no CMB removal; and for our two implementations of the target
pixel selection, presented in Sect. 4.1.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.