cpa: Confirmatory Path Analysis

In cpa: Confirmatory Path Analysis Through 'd-sep' Tests

Confirmatory Path Analysis

Description

cpa performs a confiratory path analysis on causal hypotheses expressed as directed acyclic graphs (DAGs) through the use of d-sep tests.

Usage

cpa()

Details

The function builds a graphical user interface to load the necessary files and perform the analyses. The user is asked to supply a data matrix, a list with the direct interactions implied by a given DAG and the list of variables included in the DAG. The data matrix and the list of the direct interactions should be comma-separated text files. Each direct interaction is coded as a vector of two elements, the first of which is the causal parent (the cause) and the second one the causal child (the effect). Once the files are loaded, it is possible to plot the DAG and start the analyses.

The script, sequentially, builds the basis set, performs the conditional independence tests and the Fisher's C test, calculates the Akaike's Information Criterium according to Shipley (2013), and performs the structural regressions implied by the given DAG. For an in-depth description of the code, the algorithms and the calculations refer to Bellino and co-workers (2015).

Value

The function returns an environment containing (if any) a set of user-defined environments. The user is asked to discard or save the results of each analysis, which will be then stored in an environment containing the following objects:

C	Value of the Fisher's C statistics
P	Null probability of the Fisher's C test
AIC	Akaike's Information Criterium
Bu	Table with the conditional independence claims and the null probability associated
Ti	List containing the conditional independence tests
Ti_summary	List containing the summaries of the conditional independence tests
dctests	List containing the linear models that fit the structural equations implied by the DAG. The names of its elements are the corresponding dependent variables of each model
dctests_summary	List containing the summaries of the linear models contained in 'dctests'

Author(s)

Alessandro Bellino and Daniela Baldantoni

References

Bellino A, Baldantoni D, De Nicola F, Iovieno P, Zaccardelli M, Alfani A. (2015) Compost amendments in agricultural ecosystems: confirmatory path analysis to clarify the effects on soil chemical and biological properties. Journal of Agricultural Science 253,282–295

Shipley B. (2013) The AIC model selection method applied to path analytic models compared using a d-separation test. Ecology 94 (3), 560–564.

See Also

basisSet and shipley.test functions of package ggm for an alternative way to perform the Fisher's C test

Examples

## Start the GUI and save the results of the analyses in an environment

x <- cpa()

## Inspect the content of the environment, it will contain every
## user-defined environment in which the results of each performed
## analysis are stored.

ls(x)

cpa documentation built on June 7, 2022, 1:09 a.m.