Dear All, What is the minimal group size for level sufficient to use "sts test var, trend"? Specifically, having many levels in var, say, var = 1,2,3,..15 I want to check the trend for survivor functions for 1, 2, ,,,15 but not sure what is the minimal groups size required for each level var = 1, 2, 3, ....15. I know we can use this also for, say, var = 1, 1.5, 2.5 , .... but can't find any suggested minimal group size for each var value to make sure I can use this "sts test var, trend" test. Stata calculates this even having one observation for some levels (so it looks I can do kind of univariate analysis for continous variable, but one-observation survivor function used in testing the trend for all the levels rather doesn't make sense). If using "sts test var, trend strata(dummy)", what would be then minimal group required for all levels in dummy==0 and in dummy==1. Many Thanks.
Related Posts with sts var, trend
Compatability error while creating a spatial lagged explanatory variableI am trying to create a spatially lagged X variable to run a SLX type model Issue: I get a compatib…
number of columns (variables)?Dear All, I found this question here. The question is that how we count the number of columns in the…
asclogit choice model set returns incorrect coefficientsI am using asclogit to analyze a choice model (12 cases with 3 alternatives, 1 independent continuou…
replacing numeric values with datesI have a variable named "year" which randomly displays the values 5 to 22. But I want "year" to disp…
Panel dataset and omitted dummy variables for regressionDear experts, I have a panel dataset of 77 variables and approximately 57.000 observations for the …
Subscribe to:
Post Comments (Atom)
0 Response to sts var, trend
Post a Comment