Table 2.
The residual maximum likelihood (REML) table for soil organic carbon (SOC) (g 100 g−1; %) under long‐term grass, arable (1949–) or fallow (1959–) treatments in 2008 with the REML structure for splines outlined in models (4), (5), (6)
| Fixed term | Wald statistic | n.d.f. | F statistic | d.d.f. | P |
|---|---|---|---|---|---|
| Experiment | 153.1 | 1 | 153.1 | 29.2 | < 0.001 |
| Depthv | 545.2 | 1 | 545.2 | 29.1 | < 0.001 |
| Experiment × Treatment | 137.8 | 1 | 137.8 | 29.0 | < 0.001 |
| Experiment × Depthv | 100.5 | 1 | 100.5 | 29.1 | < 0.001 |
| Experiment × Treatment × Depthv | 129.0 | 1 | 129.0 | 29.0 | < 0.001 |
The table gives the degrees of freedom (d.f.) associated with the numerator (n) and denominator (d), the Wald statistic, the variance ratio (F) statistic and the probability level associated with F (P) to assess experiment and treatment factors, the depth variate (v) and their interactions. Note that the Wald statistic is identical to the F statistic because the numerator d.f. was 1. SOC is shown in Figure 2(a).