Exploratory Graph Analysis in context

Abstract

The current paper presents the network psychometric framework for dimensionality and item analysis termed exploratory graph analysis (EGA). It starts by briefly contextualizing the field of network psychometrics and the early work from the 50’s and 60’s. Then it provides a brief overview of exploratory graph analysis and other recent developments, such as the network loadings (a metric akin to factor loadings), total entropy fit index (verification of dimensionality fit), dynamic EGA, bootstrap EGA for dimensionality and item stability, random intercepts EGA (handling wording effects), and hierarchical EGA to estimate high-order structures (e.g., generalized bifactor models). The goal of the paper is to present the reader with a list of contextualized references.

High-impact communication in nursing leadership

“…Data were analyzed using SPSS Version 27 software and the R/ EGA package. 20,21 Summed scores were used in correlations between MLQ 5X leadership styles and HI communication clusters. Skewness and kurtosis of the summed scores were in the range of normality, and Pearson correlations were used to evaluate the strength and direction of the summed scores.…”

Fundamental Fears in Greece (I)

“…We conducted data analysis using SPSS v. 26 (IBM Corp, 2019) and R (R Core Team, 2021), employing the following R packages: "foreign" (0.8-84; R Core Team, 2022), "lavaan" (v0.6-12; Rosseel, 2012), "psych" (v2.2.9; Revelle, 2022), "EFAtools" (v0.4.4; Steiner, 2023), "EGAnet" (v1.1.1, Golino & Christensen, 2022), "semTools" (v0.5-6; Jorgensen et al, 2020), "ggpubr" (v0.6.0; Kassambara, 2023a), "rstatix" (v0.7.2; Kassambara, 2023b), "car" (v3.1-2, Fox et al, 2023), "responsePatterns (v0.1.0; Rihacek & Gotfried, 2022), and "emmeans" (v1.8.6;Lenth, 2023).…”

Network Psychometrics of the 10-Item Perceived Stress Scale Among Patients With High Cardiovascular and Type 2 Diabetes Risk Using Exploratory Graph Analysis

“…A recent advancement in network psychometrics is exploratory graph analysis (EGA). [15][16][17] EGAworks by first modeling the correlation matrix of the observable variables and then using the graphical least absolute shrinkage and selection operator (LASSO) regularization to gain a sparse inverse covariance matrix (Gaussian graphical model). Next, the Walktrap cluster detection algorithm 18 -a common technique for estimating clusters in weighted networks-identifies a discrete number of dense subgraphs (ie, clusters or communities) of the partial correlation matrix calculated in the previous step.…”

“…In other words, as partial correlations are used in estimating network models, the use of the L1 regularization technique is important in controlling spurious/insignificant connection links between variables. 19 Indeed, 2 recent simulation studies 15,19 demonstrated that EGA achieves the overall highest accuracy compared with a number of traditional factor analytics including scree test and parallel analysis. Golino et al 16 also reported that EGA shows the highest overall accuracy (87.9%) in detecting the number of simulated dimensions, which is followed by the traditional parallel analysis with principal components (83.0%) and parallel analysis using principal axis factoring (81.9%).…”

“…Moreover, the accuracy of EGA is more resilient to variations in sample size, number of scale items per factor, and interdimensional correlations. 15,19,20 EGA is also advantageous over traditional analytics such as exploratory factor analysis (EFA) with respect to having the item-dimension allocation readily interpretable without the researcher's interpretation (or potential a priori assumptions) before producing the factor loading matrix and determining rotation methods to assist in interpretation. 21 Although numerous options are available to help researchers determine the number of factors to extract in EFA (eg, Kaiser-Guttman rule, various maximum likelihood approaches, parallel analysis), it is often difficult to objectively decide which option provides the most appropriate recommendations when discrepancies occur.…”