Matrix dimensions bias demographic inferences: implications for comparative plant demography

Area of Science:

• Ecology
• Population Biology
• Quantitative Biology

Background:

• Plant demography uses projection matrices for comparative studies.
• Matrix dimension variation complicates interspecific comparisons.
• Collapsing matrices to a common dimension may introduce bias.

Purpose of the Study:

• Examine how matrix dimension affects demographic elasticities.
• Evaluate different matrix collapsing criteria.
• Identify optimal methods for standardizing matrix dimensions in plant demography.

Main Methods:

• Analyzed 13x13 matrices for nine plant species.
• Collapsed matrices to smaller dimensions (7x7, 5x5, 4x4, 3x3).
• Applied various collapsing criteria, including one minimizing elasticity dissimilarities.

Main Results:

• Reduced matrix dimensions increased stasis and fecundity elasticities.
• Reduced matrix dimensions decreased progression and retrogression elasticities.
• Herbaceous perennial matrices showed particular sensitivity to dimensionality changes.

Conclusions:

• Projection matrix dimension significantly affects demographic parameters.
• A novel collapsing criterion improves matrix standardization for comparative studies.
• Recommends collapsing higher classes while retaining initial classes for normalization.