Model Diagnostic Plots for Repeated Measures Data

Abstract

SummaryIn the analysis of repeated measurements, multivariate regression models that account for the correlations among the observations from the same subject are widely used. Like the usual univariate regression models, these multivariate regression models also need some model diagnostic procedures. Though these models have been widely used, not many studies have been performed in model diagnostic areas. In this paper, we propose simple residual plots to investigate the goodness of model fit for repeated measures data. Here, we mainly focus on the mean model diagnostics. The proposed residual plots are based on the quantile-quantile(Q-Q) plots of a c 2 distribution and a normal distribution. In particular, the proposed model is useful in comparing several models simultaneously. The proposed method is illustrated using two examples.

Using Search Engine Data as a Tool to Predict Syphilis

“…Normalized residuals plots for the initial trained models displayed heteroscedasticity and bias, which was improved by transforming the dependent variable (P&S syphilis cases) using the natural log. 24,25 Applying the natural log transformation and retraining the models with LMM-LASSO produced more robust models due to relatively improved residual plots. Natural log transformation in regression introduced statistical bias, thus bias correction was applied as part of the back-transformation.…”

Goodness-of-fit tests in mixed models

“…The QQ-plots of Park and Lee (2004) for longitudinal data are based on the fact that, under normality, a quadratic form in the residuals Y − Xβ is approximately chisquared distributed when estimated variances are inserted in the covariance matrix.…”

Goodness-of-Fit Tests in Mixed Models

“…The QQ-plots of Park and Lee (2004) for longitudinal data are based on the fact that, under normality, a quadratic form in the residuals Y −X β is approximately chi-squared distributed when estimated variances are inserted in the covariance matrix.…”