omitted variable bias
Definition
Omitted variable bias refers to systematic distortion in network psychometric models that arises when important variables are excluded from the analysis, resulting in spurious edges (false connections) appearing in the network or the suppression of true connections that ultimately distort the model's structure. In cross-sectional network psychometric applications, omitted variables can lead to decreased statistical power and increased false positive rates in hypothesis testing. When combined with measurement error, omitted variable bias can result in researchers incorrectly identifying the most central nodes, potentially leading to ineffective or misguided intervention strategies.
Sources: Henry & Ye (2024)
Related Terms
Applications
Omitted Variable Bias and Measurement Error
Omitted variables and measurement error represent parallel sources of model misspecification in network psychometric models, both capable of distorting network structure and edge estimation. Omitted variables can create spurious edges or suppress true connections, while measurement error affects edge estimation, with both problems compromising the accurate identification of central nodes and symptom relationships in clinical applications.
Sources: Henry & Ye (2024)
