An Introduction to Phylogenetic Path Analysis
The questions addressed by macroevolutionary biologists are often impervious to experimental approaches, and alternative methods have to be adopted. The phylogenetic comparative approach is a very powerful one since it combines a large number of species and thus spans long periods of evolutionary change. However, there are limits to the inferences that can be drawn from the results, in part due to the limitations of the most commonly employed analytical methods. In this chapter, we show how confirmatory path analysis can be undertaken explicitly controlling for non-independence due to shared ancestry. The phylogenetic path analysis method we present allows researchers to move beyond the estimation of direct effects and analyze the relative importance of alternative causal models including direct and indirect paths of influence among variables. We begin the chapter with a general introduction to path analysis and then present a step-by-step guide to phylogenetic path analysis using the d-separation method. We also show how the known statistical problems associated with non-independence of data points due to shared ancestry become compounded in path analysis. We finish with a discussion about the potential effects of collinearity and measurement error, and a look toward possible future developments.
We thank László Zsolt Garamszegi for inviting us to write this chapter, as well as him and two anonymous referees for their useful comments and suggestions on a first draft of this chapter.
- Burnham KP, Anderson DR (2002) Model selection and multimodel inference: a practical information-theoretic approach. Springer, New YorkGoogle Scholar
- Fisher RA (1926) The design of experiments, 1st edn. Oliver and Boyd, EdinburghGoogle Scholar
- Harvey PH, Pagel MD (1991) The comparative method in evolutionary biology. Oxford University Press, OxfordGoogle Scholar
- Kline RB (2010) Principles and practice of structural equation modelling methodology in the social sciences, 3rd edn. Guilford Press, New YorkGoogle Scholar
- Pearl J (1988) Probabilistic reasoning in intelligent systems. Morgan and Kaufmann, San MateoGoogle Scholar
- Pugesek BH, Grace JB (1998) On the utility of path modelling for ecological and evolutionary studies. Funct Ecol 12:853–856Google Scholar
- Santos JC (2009) The implementation of phylogenetic structural equation modeling for biological data from variance-covariance matrices, phylogenies, and comparative analyses. The University of Texas at Austin, AustinGoogle Scholar
- Santos JC, Cannatella DC (2011) Phenotypic integration emerges from aposematism and scale in poison frogs. Proc Natl Acad Sci USAGoogle Scholar
- Shipley B (2004) Analysing the allometry of multiple interacting traits. Perspect Plant Ecol Evol Syst 6(235):241Google Scholar
- Stümpke H (1967) The Snouters: form and life of the Rhinogrades (trans: Doubleday & Company I). University of Chicago Press, ChicagoGoogle Scholar
- Team RDC (2013) R: A language and environment for statistical computing. R Foundation for Statistical Computing, ViennaGoogle Scholar
- Verma T, Pearl J (1988) Causal networks: semantics and expressiveness. In: Schachter R, Levitt TS, Kanal LN (eds) Uncertainty in artificial intelligence, vol 4. Elsevier, Amsterdam, pp 69–76Google Scholar