Comparison of single-tree MI vs multi-tree MI on avonet300 (300 species, mixed-type traits).
Downstream model: log(Mass) ~ log(Wing.Length) + log(Beak.Length_Culmen)
Single-tree: MCC tree, m = 50 MC-dropout draws.
Multi-tree: 10 posterior trees × 5 draws = 50 total datasets.
| Term | Method | Estimate | SE | df | FMI | RIV |
|---|---|---|---|---|---|---|
| (Intercept) | Single-tree MI | -7.7734 | 0.4264 | 680.1 | 0.293 | 0.418 |
| (Intercept) | Multi-tree MI (50 trees) | -7.7447 | 0.3106 | 544.9 | 0.311 | 0.451 |
| log(Wing.Length) | Single-tree MI | 2.2892 | 0.1356 | 602.9 | 0.301 | 0.434 |
| log(Wing.Length) | Multi-tree MI (50 trees) | 2.2720 | 0.1115 | 339.2 | 0.389 | 0.633 |
| log(Beak.Length_Culmen) | Single-tree MI | 0.3313 | 0.1301 | 512.9 | 0.322 | 0.474 |
| log(Beak.Length_Culmen) | Multi-tree MI (50 trees) | 0.3504 | 0.1103 | 348.2 | 0.389 | 0.639 |
| Term | Method | Estimate | SE | df | FMI | RIV |
|---|---|---|---|---|---|---|
| (Intercept) | Single-tree MI | -7.3682 | 0.4327 | 379.4 | 0.368 | 0.578 |
| (Intercept) | Multi-tree MI (50 trees) | -7.2205 | 0.4985 | 136.7 | 0.628 | 1.940 |
| log(Wing.Length) | Single-tree MI | 2.1998 | 0.1470 | 247.3 | 0.451 | 0.808 |
| log(Wing.Length) | Multi-tree MI (50 trees) | 2.1008 | 0.1642 | 147.8 | 0.610 | 1.829 |
| log(Beak.Length_Culmen) | Single-tree MI | 0.3292 | 0.1492 | 224.3 | 0.476 | 0.898 |
| log(Beak.Length_Culmen) | Multi-tree MI (50 trees) | 0.4260 | 0.1541 | 162.8 | 0.567 | 1.330 |
| Term | Method | Estimate | SE | df | FMI | RIV |
|---|---|---|---|---|---|---|
| (Intercept) | Single-tree MI | -7.0156 | 0.4859 | 327.5 | 0.410 | 0.713 |
| (Intercept) | Multi-tree MI (50 trees) | -6.8702 | 0.7352 | 88.0 | 0.760 | 3.476 |
| log(Wing.Length) | Single-tree MI | 2.0410 | 0.1474 | 341.5 | 0.392 | 0.642 |
| log(Wing.Length) | Multi-tree MI (50 trees) | 1.9386 | 0.2334 | 91.7 | 0.753 | 3.606 |
| log(Beak.Length_Culmen) | Single-tree MI | 0.4547 | 0.1382 | 308.1 | 0.405 | 0.674 |
| log(Beak.Length_Culmen) | Multi-tree MI (50 trees) | 0.5650 | 0.2022 | 112.0 | 0.687 | 2.463 |
Values > 1 indicate that phylogenetic uncertainty widens standard errors beyond what single-tree MI provides.
| Missing % | Term | SE (single) | SE (multi) | Ratio |
|---|---|---|---|---|
| 20% | (Intercept) | 0.4264 | 0.3106 | 0.73 |
| 20% | log(Wing.Length) | 0.1356 | 0.1115 | 0.82 |
| 20% | log(Beak.Length_Culmen) | 0.1301 | 0.1103 | 0.85 |
| 40% | (Intercept) | 0.4327 | 0.4985 | 1.15 |
| 40% | log(Wing.Length) | 0.1470 | 0.1642 | 1.12 |
| 40% | log(Beak.Length_Culmen) | 0.1492 | 0.1541 | 1.03 |
| 60% | (Intercept) | 0.4859 | 0.7352 | 1.51 |
| 60% | log(Wing.Length) | 0.1474 | 0.2334 | 1.58 |
| 60% | log(Beak.Length_Culmen) | 0.1382 | 0.2022 | 1.46 |
Wall time: 279.4 min. Generated: 2026-04-17 16:01
Reference: Nakagawa S, de Villemereuil P (2019). Systematic Biology 68(4): 632-641.