Table 1. Deleted terms and order or deletion from hierarchical simplification of three statistical models of male-male aggression.
| Response | Time to first attack | Intensity of first attack bout | Number of attacksa | |||||||||
| Model type | Cox proportional hazardsregression model | Linear model | Linear model | |||||||||
| Variable | order | χ2 | df | P | order | F | df | P | order | F | df | P |
| T:RMS:IMS | 1 | 0.30 | 1 | 0.58 | 1 | 0.32 | 12,13 | 0.58 | 1 | 0.09 | 22,23 | 0.77 |
| T:RMS | 2 | <0.01 | 1 | 0.96 | 4 | <0.01 | 13,14 | 0.99 | 2 | 0.13 | 23,24 | 0.72 |
| T:IMS | 3 | 0.45 | 1 | 0.50 | 3 | <0.01 | 14,15 | 0.96 | 3 | 1.23 | 24,25 | 0.28 |
| RMS:IMS | 4 | 0.60 | 1 | 0.44 | 2 | 0.18 | 15,16 | 0.68 | 4 | 0.93 | 25,26 | 0.34 |
| IMS | 5 | 0.03 | 1 | 0.85 | 5 | 0.40 | 16,17 | 0.53 | 5 | 0.14 | 26,27 | 0.71 |
| T | 6 | 0.56 | 1 | 0.46 | 6 | 2.11 | 17,18 | 0.16 | 6 | 1.20 | 27,28 | 0.28 |
The predictor variables were treatment (T, presence or absence of females prior to trial), resident male size (RMS), intruding male size (IMS) and all their interactions.
a
The response variable was square-root transformed.