Table B.1.
Effective number of constrained parameters.
| Split type | Setup | NΘ, χ2 | NΘ, BMD | dχ2 | dBMD | ||
|---|---|---|---|---|---|---|---|
| Fiducial - 1cosmo (COSEBIs) | 5.60 | 6.00 | - | - | |||
| Fiducial - 1cosmo (Band powers) | 5.23 | 4.16 | - | - | |||
| Fiducial - 1cosmo (2PCFs) | 7.79 | 7.79 | - | - | |||
| Data vector | Bin 1 | 8.05 | 7.60 | 2.45 | 1.60 | ||
| Bin 2 | 8.20 | 5.52 | 2.60 | -0.48 | |||
| Bin 3 | 8.48 | 7.83 | 2.88 | 1.83 | |||
| Bin 4 | 8.69 | 9.05 | 3.09 | 3.05 | |||
| Bin 5 | 9.01 | 8.51 | 3.41 | 2.51 | |||
| Bin 6 | 8.85 | 8.06 | 3.25 | 2.06 | |||
| Redshift bin AC vs CC | 8.53 | 8.81 | 2.93 | 2.81 | |||
| En = 1 vs En > 1 | 7.35 | 7.22 | 1.75 | 1.22 | |||
| Small multipoles vs large multipoles (Band powers) | 7.70 | 5.98 | 2.47 | 1.82 | |||
| Small scales vs large scales (2PCFs) | 10.67 | 10.35 | 2.88 | 2.56 | |||
| Catalogue | North vs South | 9.17 | 7.45 | 3.03 | 1.81 | ||
| Red vs blue, TB = 3.0 | 9.90 | 7.55 | 3.63 | 0.80 | |||
| Red vs blue, TB = 1.9 | 7.03 | 8.70 | 3.40 | 3.80 | |||
| North vs South - 1cosmo | 6.14 | 5.64 | - | - | |||
| Red vs blue - 1cosmo, TB = 3.0 | 6.27 | 6.75 | - | - | |||
| Red vs blue - 1cosmo, TB = 1.9 | 3.63 | 5.54 | - | - | |||
| Statistic | COSEBIs vs Band powers | 10.22 | 10.69 | 3.26 | 4.59 | ||
| COSEBIs vs 2PCFs | 9.45 | 11.16 | 3.80 | 3.40 | |||
| Band powers vs 2PCFs | 8.57 | 10.01 | 3.59 | 3.03 | |||
| COSEBIs vs Band powers - 1cosmo | 6.96 | 6.10 | - | - | |||
| COSEBIs vs 2PCFs - 1cosmo | 5.65 | 7.76 | - | - | |||
| Band powers vs 2PCFs - 1cosmo | 4.98 | 6.98 | - | - | |||
Notes. For each analysis setup we provide the effective number of constrained parameters inferred via χ2 minimisation of mock data vectors (Joachimi et al. 2021) in the third column, and the Bayesian model dimensionality (Handley & Lemos 2019a), inferred directly from the chain, in the fourth column. The first three rows show the results in an analysis with the fiducial setup, dubbed ‘1cosmo‘, in which we do not apply any splits to the data. The remaining rows present results when splitting the data into subsets. Here, we report the effective number of constrained parameters with the split cosmological model. For the splits at the catalogue level and the splits by summary statistic we additionally report the results with a shared set of parameters. The final two columns report the difference in the number of constrained parameters between the fiducial cosmological model and the split cosmological model. We employ COSEBIs as the summary statistic except when explicitly stated.
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.