Table 3.
P-values and effect sizes for each dataset from starting tree bias analysis
| Dataset | Partitioning method | p-value | effect size |
|---|---|---|---|
| Anderson_2013 | Greedy algorithm | 1 | 0 |
| Anderson_2013 | lnL rates k-means | 0.03 | −0.604 |
| Anderson_2013 | TIGER rates k-means | 1 | 0 |
| Cognato_2001 | Greedy algorithm | 1 | 0 |
| Cognato_2001 | lnL rates k-means | 0.006 | −0.424 |
| Cognato_2001 | TIGER rates k-means | 0.202 | −0.152 |
| Grande_2013 | Greedy algorithm | 1 | 0 |
| Grande_2013 | lnL rates k-means | 0.047 | −0.225 |
| Grande_2013 | TIGER rates k-means | 1 | 0 |
| Kang_2013a | Greedy algorithm | 1 | 0 |
| Kang_2013a | lnL rates k-means | 0.391 | −0.105 |
| Kang_2013a | TIGER rates k-means | 1 | 0.14 |
| Kawahara_2013 | Greedy algorithm | 0.397 | −0.095 |
| Kawahara_2013 | lnL rates k-means | 0.008 | −0.348 |
| Kawahara_2013 | TIGER rates k-means | 0.828 | 0.051 |
| Leavitt_2013 | Greedy algorithm | 1 | 0 |
| Leavitt_2013 | lnL rates k-means | 1 | 0 |
| Leavitt_2013 | TIGER rates k-means | 1 | 0 |
| McCormack_2013a | Greedy algorithm | 0.409 | −0.116 |
| McCormack_2013a | lnL rates k-means | 0.52 | −0.048 |
| McCormack_2013a | TIGER rates k-means | 0.158 | −0.084 |
| Oaks_2011 | Greedy algorithm | 1 | 0.094 |
| Oaks_2011 | lnL rates k-means | 1 | 0.019 |
| Oaks_2011 | TIGER rates k-means | 1 | 0 |
| Sharanowski_2011 | Greedy algorithm | 1 | 0 |
| Sharanowski_2011 | lnL rates k-means | 0.056 | −0.304 |
| Sharanowski_2011 | TIGER rates k-means | 0.069 | −0.222 |