Estimating and examining the replicability of belief system networks

M.J. Brandt*

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

9 Downloads (Pure)


Belief system structure can be investigated by estimating belief systems as networks of interacting political attitudes, but we do not know if these estimates are replicable. In a sample of 31 countries from the World Values Survey (N = 52,826), I find that countries’ belief system networks are relatively replicable in terms of connectivity, proportion of positive edges, some centrality measures (e.g., expected influence), and the estimates of individual edges. Betweenness, closeness, and strength centrality estimates are more unstable. Belief system networks estimated with smaller samples or in countries with more unstable political systems tend to be less replicable than networks estimated with larger samples in stable political systems. Although these analyses are restricted to the items available in the World Values Survey, they show that belief system networks can be replicable, but that this replicability is related to features of the study design and the political system.
Keywords: Belief systems, networks, replication, political stability
Original languageEnglish
Article number24
JournalCollabra: Psychology
Issue number1
Publication statusPublished - 2020

Fingerprint Dive into the research topics of 'Estimating and examining the replicability of belief system networks'. Together they form a unique fingerprint.

  • Cite this