Figure 1From: Improving statistical inference on pathogen densities estimated by quantitative molecular methods: malaria gametocytaemia as a case study Heteroscedasticity of calibration or standard curves from quantitative nucleic acid sequence-based amplification (QT-NASBA) assays. Panel A depicts the conditional raw residuals of a homoscedastic linear mixed model with a correlated random intercept and slope (the homoscedastic homologue of the selected Model 7 in Table 2) plotted against the fitted time to positivity (TTP). The data points in panel B depict the natural logarithm (ln) of the variance of the residuals presented in panel A grouped by deciles (10-quantiles) of fitted TTP. The solid line and shaded area are for presentation purposes only, representing a linear regression ± 2 standard deviations respectively.Back to article page