How to Do Ecology by unknow

Author:unknow
Language: eng
Format: epub
ISBN: 9781400851263
Publisher: Princeton UP
Published: 2014-09-15T05:00:00+00:00

PATH ANALYSIS AND CAUSAL RELATIONSHIPS

Path analysis is a technique that allows a researcher to consider different causal schemes (paths) and to evaluate which of those paths does the best job of explaining the observed patterns, that is, which factors are likely to cause which other factors (Shipley 2000, Mitchell 2001, Grace 2006). The diagrams with arrows that connected weather, plant stress, and herbivore numbers in chapter 2 are path diagrams. The easiest way to evaluate paths is by using partial regression coefficients to estimate the strength of individual paths (arrows in the diagrams) along with goodness-of-fit statistics to estimate the fit of the overall model (see Mitchell 2001 for a very clear description with worked examples, and Grace 2006 for a more complete, but still accessible, account). Partial regression coefficients provide an estimate of the effect of one variable on another when the effects of other variables are held constant.

Path analysis can shed light on the likely validity of several different cause-and-effect hypotheses (Shipley 2000, Mitchell 2001, Grace 2006). It is particularly useful for evaluating the roles of indirect effects (fig. 7).

Path analysis that includes structural equation modeling and maximum likelihood methods can be used to generate plausible hypotheses when several are possible, to estimate effect sizes, and also to confirm that one causal hypothesis is more likely than others. It can be particularly valuable when there are many factors that could potentially interact in a variety of different ways; in such cases it may be difficult or impossible to experimentally manipulate all of them.

Path analysis can be extremely useful early in a study before you commit to time-consuming manipulations, and it is particularly valuable when used along with experimental manipulations. The simple act of drawing out hypothesized relationships as paths is a good opportunity to formally think about the way you view your system. Path analysis using structural equation modeling allows you to increase your confidence when inferring causality from the results of observational studies. In cases where both experimental manipulations and observations have been conducted, they can be combined in a single path analysis.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.