Geoscience Reference
In-Depth Information
Table 2
Response variables, levels of aggregations and their comparisons
Response
variable
Focal mass
(explanatory variable)
Comparison
First aggregation
Second aggregation
Model
Prey mass
Predator mass
Resolution
A
D
LMM
Prey mass
Predator mass
Grouping
C
F
LMM
PPMR
Predator mass
Resolution
A
D
LMM
PPMR
Predator mass
Grouping
C
F
LMM
Trophic height
Species mass
Grouping
D*
F*
OLS
Prey variance
Predator mass
Resolution
B
D
OLS
Prey variance
Predator mass
Grouping
B
E
OLS
Prey range
Predator mass
Resolution
B
D
OLS
Prey range
Predator mass
Grouping
B
E
Polynomial
In-degree
Species mass
Grouping
D*
F*
OLS
Predator variance
Prey mass
Resolution
B
D
OLS
Predator variance
Predator mass
Grouping
B
E
OLS
Predator range
Predator mass
Resolution
B
D
OLS
Predator range
Predator mass
Grouping
B
E
OLS
Out-degree
Species mass
Grouping
D*
F*
Polynomial
The levels of aggregation (A-F) used in the resolution or grouping comparisons for the various response variables examined and the type of model used to
make the comparison. The raw data of individual predator-prey interactions (A) were aggregated using a taxonomic-based approach to generate different
levels of resolution (A-D) or binned into body size classes using a size-class-based approach to provide groups (E and F) which could be contrasted with
the taxonomic aggregations. For each response variable, comparisons were then done between the first and second aggregation (either between levels of
resolution or between the two approaches to grouping). D* and F* denote food web aggregations needed to calculate and compare some of the response
variables, such as trophic height (see Section
II.B.3
). Focal mass is the explanatory variable for the corresponding response variable. The three sets of
response variables refers to the Size Structure Dimension Set #1-3 described in detail in Section
II.C.1-2
. PPMR stands for predator mass:prey mass ratio,
LMM stands for linear mixed effect model and OLS stands for ordinary least squares. See Section II.E.1 for descriptions of the models used.