Factor or Network Model? Predictions From Neural Networks

Christensen, Alexander P.; Golino, Hudson

doi:10.35566/jbds/v1n1/p5

Cited by 12 publications

(4 citation statements)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Network loadings of .15, .25, and .35 are equivalent to low (.40), moderate (.55), and high (.70) network loadings, respectively (Christensen & Golino, 2021c ). The development of network loadings opened new lines of research, such as the development of metric invariance using EGA and permutation tests in a network perspective (Jamison et al, 2022 ), and determining whether data are generated from a factor or network model (Christensen & Golino, 2021b ).…”

Section: Methodsmentioning

confidence: 99%

The Misinformation Susceptibility Test (MIST): A psychometrically validated measure of news veracity discernment

Maertens,

Götz,

Golino

et al. 2023

Behav Res

View full text Add to dashboard Cite

Interest in the psychology of misinformation has exploded in recent years. Despite ample research, to date there is no validated framework to measure misinformation susceptibility. Therefore, we introduce Verificationdone, a nuanced interpretation schema and assessment tool that simultaneously considers Veracity discernment, and its distinct, measurable abilities (real/fake news detection), and biases (distrust/naïvité—negative/positive judgment bias). We then conduct three studies with seven independent samples (Ntotal = 8504) to show how to develop, validate, and apply the Misinformation Susceptibility Test (MIST). In Study 1 (N = 409) we use a neural network language model to generate items, and use three psychometric methods—factor analysis, item response theory, and exploratory graph analysis—to create the MIST-20 (20 items; completion time < 2 minutes), the MIST-16 (16 items; < 2 minutes), and the MIST-8 (8 items; < 1 minute). In Study 2 (N = 7674) we confirm the internal and predictive validity of the MIST in five national quota samples (US, UK), across 2 years, from three different sampling platforms—Respondi, CloudResearch, and Prolific. We also explore the MIST’s nomological net and generate age-, region-, and country-specific norm tables. In Study 3 (N = 421) we demonstrate how the MIST—in conjunction with Verificationdone—can provide novel insights on existing psychological interventions, thereby advancing theory development. Finally, we outline the versatile implementations of the MIST as a screening tool, covariate, and intervention evaluation framework. As all methods are transparently reported and detailed, this work will allow other researchers to create similar scales or adapt them for any population of interest.

show abstract

Section: Methodsmentioning

confidence: 99%

The Misinformation Susceptibility Test (MIST): A psychometrically validated measure of news veracity discernment

Maertens,

Götz,

Golino

et al. 2023

Behav Res

View full text Add to dashboard Cite

show abstract

“…Woodbury matrix identity [Woodbury, 1950] that oblique factors are statistically consistent with clusters of nodes (i.e., sets of connected nodes) and orthogonal factors are statistically consistent with unconnected clusters in GGM, when the data generation mechanism is a factor model. Christensen and Golino [2021] showed that factor loadings are statistically consistent with a modified version of node strengths (i.e., sum of all connections to a node) that takes into consideration of the dimensionality structure, represented as network loadings.…”

Section: Introductionmentioning

confidence: 67%

A Bayesian Approach for Dimensionality Assessment in Psychological Networks

Shi¹,

Christensen²,

Day³

et al. 2023

Preprint

View full text Add to dashboard Cite

To understand psychological data, it is crucial to examine the structure and dimensions of variables. In this study, two Bayesian approaches were developed in network psychometric models to explore the dimensionality structure of the data, termed Bayesian exploratory graph analysis or BEGA. Unlike traditional approaches that provide fixed parameter estimates, BEGA estimates posterior probabilities of graphical structures to assess the conditional dependence relations and then employs the Louvain community detection algorithm to partition and identify groups of nodes, which enables to detect the multidimensional factor structures. Monte Carlo simulations suggested that the two BEGA methods had comparable or better performance when compared with the other network-based method (EGA) and conventional parallel analysis (PA). When estimating the multidimensional factor structure, the analytically based method (i.e., BEGA.A) showed the best balance between accuracy and mean biased/absolute errors, with the second highest accuracy slightly lower than EGA but with the smallest errors. The sampling-based approach (BEGA.S) yielded higher accuracy and smaller errors than PA; lower accuracy but also lower errors than EGA. Both Bayesian techniques had more stable performance than EGA and PA across different data conditions. When estimating the unidimensional structure, the PA technique performed the best, followed closely by BEGA.A, EGA, and then BEGA.S. The study recommends using BEGA.A as an alternative tool for assessing dimensionalities and advocates for the usefulness of BEGA.S as a valuable alternate technique. The study suggested the potential to extend the regularization-based network modeling EGA method to the Bayesian framework.

show abstract

“…They showed a simulation study that network loadings are akin to factor loadings, opening a new line of research within the EGA approach. These loadings have opened to door to assessing measurement (metric) invariance (Jamison et al, 2022) from the network perspective as well as determining whether data are generated from a factor or network model (Christensen & Golino, 2021b).…”

Section: Other Developments and Future Directionsmentioning

confidence: 99%

Exploratory Graph Analysis in context

Golino¹,

Christensen²,

Garrido

2022

PTP

Self Cite

View full text Add to dashboard Cite

This 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 1950s and 1960s. Then, it provides a brief overview of EGA 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.

show abstract

Factor or Network Model? Predictions From Neural Networks

Cited by 12 publications

References 59 publications

The Misinformation Susceptibility Test (MIST): A psychometrically validated measure of news veracity discernment

The Misinformation Susceptibility Test (MIST): A psychometrically validated measure of news veracity discernment

A Bayesian Approach for Dimensionality Assessment in Psychological Networks

Exploratory Graph Analysis in context

Contact Info

Product

Resources

About