Generalized variance functions (GVFs) are used to produce official standard errors for the Current Population Survey (CPS) estimates included in The Employment Situation , a Principal Federal Economic Indicator (PFEI) published monthly by the Bureau of Labor Statistics. The parameters of the GVF models are also published and available to data users. However, variance function coverage beyond The Employment Situation is limited, and no GVF parameters are published for non-PFEI news releases. In this paper, recently developed models are extended in multiple dimensions, including: from time-dependent to time-robust parameters; from medians to other percentiles; from full-sample to partial-sample data; and from national to state-level estimates. Strengths and weaknesses of various research models are considered, specifically in relation to the quality of estimated standard errors and the convenience of the model form for parameter publication.