Skip to main content
. 2022 Apr 20;6(3):igac027. doi: 10.1093/geroni/igac027

Figure 1.

Figure 1.

Joint latent growth model of cognitive functioning (initial cognitive status and rate of cognitive change) and discrete-time survival model (mortality risk), including the predictor variable (low-quality responding) and covariate age. Additional covariates included in the models are not shown for visual clarity. c1–c6 are observed continuous test scores for each assessment wave, and they are used as indicators of intercept (initial status) and linear change (rate of change) factors in a latent growth model. d1–d10 are coded as 0 if the respondent is still alive, 1 if the respondent died during that year, and missing in subsequent years following death; they serve as indicators of a mortality risk factor with loadings constrained at 1 following the assumption of proportional hazard odds.