A common assumption of latent variables is that their means are zero and their variances are one. In practice, this assumption is not always true, particularly in the case of endogenous latent variables. At the same time, visual representations of latent variable interactions often rely on plotting effects at values corresponding to standard deviations of the latent moderating variable. This discussion draws attention to this often unmet assumption, highlights some potential consequences of failing to meet it, and offers a generalized solution to estimating the mean and variance of endogenous latent variables.