Table B.1.
Statistical test results (p-values) comparing FSRQ and RQ NTD distributions for the full samples and different NTD regimes.
| Sample | ECDF | Wasserstein | Std. mean diff. | Stddev ratio | High moment diff. |
|---|---|---|---|---|---|
| Full Sample | |||||
| K-S | 1.426 × 10−9 | 1.725 × 10−9 | 1.426 × 10−9 | 1.614 × 10−9 | 1.582 × 10−9 |
| A-D | < 1 × 10−4 | < 1 × 10−4 | < 1 × 10−4 | < 1 × 10−4 | < 1 × 10−4 |
| MWU | 2.960 × 10−6 | 2.470 × 10−6 | 2.923 × 10−6 | 1.201 × 10−6 | 4.716 × 10−5 |
| Disk-Dominated | |||||
| K-S | 0.198 | 0.198 | 0.240 | 0.240 | 0.084 |
| A-D | 0.259 | 0.358 | 0.269 | 0.343 | 0.139 |
| MWU | 0.350 | 0.389 | 0.364 | 0.495 | 0.143 |
| Jet-Dominated | |||||
| K-S | 3.715 × 10−11 | 1.223 × 10−10 | 8.255 × 10−11 | 3.076 × 10−12 | 3.710 × 10−11 |
| A-D | < 1 × 10−4 | < 1 × 10−4 | < 1 × 10−4 | < 1 × 10−4 | < 1 × 10−4 |
| MWU | 1.577 × 10−13 | 1.197 × 10−13 | 1.197 × 10−13 | 3.061 × 10−15 | 1.706 × 10−13 |
| NTD)<1 | |||||
| K-S | 1.080 × 10−8 | 2.985 × 10−8 | 4.665 × 10−9 | 3.737 × 10−9 | 3.187 × 10−9 |
| A-D | < 1 × 10−4 | < 1 × 10−4 | < 1 × 10−4 | < 1 × 10−4 | < 1 × 10−4 |
| MWU | 5.888 × 10−12 | 1.096 × 10−11 | 7.598 × 10−12 | 2.358 × 10−12 | 3.566 × 10−13 |
Note 1: The comparisons are based on different metrics used to select the optimal bootstrap subsample: Empirical Cumulative Distribution Function (ECDF, column 2), Wasserstein distance (column 3), standardized mean difference (column 4), standard deviation ratio (column 5), and higher moment differences (column 6). p-values lower than 0.05 indicate statistically significant differences. Note 2: P-values reported as < 1 × 10−4 for the Anderson-Darling (A-D) test indicate that the calculated value is below the lower precision limit of the scipy.stats.anderson_ksamp function in Python.
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.