Network Analysis

Networks

Networks can be estimated from data derived from different designs, including cross-sectional, longitudinal, and experimental studies. Furthermore, the analysis can be conducted on an individual (within-person) level or at a group level. Within-person network analysis provides insights into a specific individual over time. Group-level analysis can estimate the network structure of nodes and edges that represent a specific population, whereas between-group network analysis compares networks produced by different populations.

Psychological research examines patterns of relationship between variables: The network of relationships between the variables constitutes the psychological phenomenon to be understood using network analysis. Networks are structures comprising variables (represented by nodes) and the relationships (edges) between the nodes. Within the network literature, nodes are sometimes referred to as vertices, and edges are sometimes referred to as links. A node represents the variable of interest, which can be measured in a variety of ways. For example, the node may represent the response to a questionnaire, a clinician’s rating of psychological state, observed behavior, activation in neural pathways, or levels of a specific hormone in the blood. The edge represents the nature of the relationship between the nodes. For example, it may reflect the comorbidity of psychological symptoms, the relationship between beliefs and behaviors, and the social connections between individuals.

Within a network, two types of edges are distinguished. A directed edge refers to the connection between the nodes indicating a one-way effect, represented by the head of the edge having a pointed arrowhead. In contrast, an undirected edge connects the nodes to reflect a shared relationship (e.g., correlation), but there is no indication of direction of effect. Networks may be fully directed (i.e., all edges are directed), fully undirected (i.e., no edges are directed), or mixed (i.e., both direct and undirected edges are present in the network). A directed network can be cyclic (i.e., one can follow the directed edges from a given node to end up back at that node) or acyclic (i.e., one cannot start at a node and end up back at that node again by following the directed edges).

...