Estimating True Changes when Categorical Panel Data are Affected by Uncorrelated and Correlated Classification Errors

• DOI: 10.1177/0049124100029002003
• Publication Date: 2007-03-11T06:02:55Z
• Authors: FRANCESCA BASSI; JACQUES A. HAGENAARS; MARCEL A. CROON; JEROEN K. VERMUNT
• Keywords: Categorical variable; Uncorrelated; Random error

Conclusions about changes in categorical characteristics based on observed panel data can be incorrect when (even a small amount of) measurement error is present. Random errors, referred to as independent classification usually lead over-estimation of the total gross change, whereas systematic, correlated errors cause underestimation transitions. Furthermore, patterns true change may seriously distorted by or systematic errors. Latent class models and directed log-linear analysis are excellent tools correct for both An extensive example labor market states taken from Survey Income Program Participation presented.