20 Jan 2017 | Sacha Epskamp, Denny Borsboom and Eiko I. Fried
This tutorial paper by Epskamp, Borsboom, and Fried introduces the concept of psychological networks, which conceptualize behavior as a complex interplay of psychological and other components. The authors highlight the importance of assessing the accuracy and stability of network estimates, particularly in the context of limited sample sizes common in psychological research. They outline methods for evaluating the accuracy of edge weights and the stability of centrality indices, including the use of bootstrap routines. Two novel statistical methods are introduced: the correlation stability coefficient (CS-coefficient) for assessing the stability of centrality indices and the bootstrapped difference test for comparing edge weights and centrality indices. The paper also presents simulation studies to evaluate the performance of these methods and introduces the R package *bootnet*, which facilitates the estimation of psychological networks and the application of the proposed bootstrap methods. The tutorial is accompanied by a step-by-step guide using a dataset of 359 women with post-traumatic stress disorder, demonstrating how to estimate a network, assess its accuracy, and interpret the results.This tutorial paper by Epskamp, Borsboom, and Fried introduces the concept of psychological networks, which conceptualize behavior as a complex interplay of psychological and other components. The authors highlight the importance of assessing the accuracy and stability of network estimates, particularly in the context of limited sample sizes common in psychological research. They outline methods for evaluating the accuracy of edge weights and the stability of centrality indices, including the use of bootstrap routines. Two novel statistical methods are introduced: the correlation stability coefficient (CS-coefficient) for assessing the stability of centrality indices and the bootstrapped difference test for comparing edge weights and centrality indices. The paper also presents simulation studies to evaluate the performance of these methods and introduces the R package *bootnet*, which facilitates the estimation of psychological networks and the application of the proposed bootstrap methods. The tutorial is accompanied by a step-by-step guide using a dataset of 359 women with post-traumatic stress disorder, demonstrating how to estimate a network, assess its accuracy, and interpret the results.