Skip to main content

Table 1 Model selection statistics (multistate approach) for semi-annual survival (Φ), recapture (p) and state-transition (ψ) probabilities of M. murinus depending on hair cortisol concentration and general body condition (measured as scaled mass index)

From: Hair cortisol concentrations correlate negatively with survival in a wild primate population

Rank Model K QDEV QAICc Δ i w i
Hair cortisol concentration (categorization cut-off: median)
 1 \(\varPhi \left( c \right)\;p\left( t \right) \, \psi \left( c \right)\) 8 200.78 515.10 0 0.147
 2 \(\varPhi \left( s \right)\;p\left( t \right) \, \psi \left( c \right)\) 8 201.90 516.23 1.13 0.142
 3 \(\varPhi \left( {c + s} \right) \, p\left( t \right) \, \psi \left( c \right)\) 9 199.96 516.44 1.34 0.126
 4 \(\varPhi \left( c \right) \, p\left( \cdot \right) \, \psi \left( c \right)\) 5 208.91 516.88 1.77 0.120
 53 \(\varPhi \left( {c*s + t} \right) \, p\left( {s + t} \right) \, \psi \left( {c*t} \right)\) 20 192.77 534.29 19.19 0.0001
Hair cortisol concentration (categorization cut-off: third quartile)
 1 \(\varPhi \left( c \right) \, p\left( t \right) \, \psi \left( c \right)\) 8 146.94 465.86 0 0.230
 2 \(\varPhi \left( {c + s} \right) \, p\left( t \right) \, \psi \left( c \right)\) 9 146.08 467.16 1.30 0.120
 3 \(\varPhi \left( c \right) \, p\left( \cdot \right) \, \psi \left( c \right)\) 5 154.75 467.31 1.45 0.111
 4 \(\varPhi \left( s \right) \, p\left( t \right) \, \psi \left( c \right)\) 8 148.66 467.59 1.73 0.097
 54 \(\varPhi \left( {c*s + t} \right) \, p\left( {s + t} \right) \, \psi \left( {c*t} \right)\) 20 138.21 484.33 18.47 0.00002
Scaled mass index (categorization cut-off: median)
 1 \(\varPhi \left( s \right) \, p\left( t \right) \, \psi \left( c \right)\) 7 114.35 330.82 0 0.272
 2 \(\varPhi \left( c \right) \, p\left( t \right) \, \psi \left( c \right)\) 7 115.43 331.90 1.08 0.158
 54 \(\varPhi \left( {c*s + t} \right) \, p\left( {s + t} \right) \, \psi \left( {c*t} \right)\) 16 109.73 346.67 15.85 0.0004
Scaled mass index (categorization cut-off: third quartile)
 1 \(\varPhi \left( c \right) \, p\left( t \right) \, \psi \left( c \right)\) 7 102.53 323.87 0 0.229
 2 \(\varPhi \left( {c + s} \right) \, p\left( t \right) \, \psi \left( c \right)\) 8 101.34 324.85 0.99 0.140
 3 \(\varPhi \left( s \right) \, p\left( t \right) \, \psi \left( c \right)\) 7 103.72 325.06 1.19 0.127
 4 \(\varPhi \left( c \right) \, p\left( {s + t} \right) \, \psi \left( c \right)\) 8 102.03 325.55 1.68 0.099
 54 \(\varPhi \left( {c*s + t} \right) \, p\left( {s + t} \right) \, \psi \left( {c*t} \right)\) 16 96.91 338.72 14.85 0.0001
  1. Only models with Δ i  ≤ 2 and the global model (in italic) are shown here with the number of parameters (K), the quasi-likelihood adjusted deviance (QDEV), the quasi-likelihood adjusted AIC for small sample size (QAICc), the difference between the QAICc of the top model and a given model i (Δ i ) and the Akaike weights (w i ). Variables considered are the condition index (c, which can indicate HCC or SMI values), sex (s) and time (t). Constant parameters are noted (.). Interactions are indicated by (*) and additive effects by (+)