Due to in conssitent of PUMA boundry in the 2010-2014 5-yr data, i plan to use “multyear” variable to extact the 2012-2014 data to use it as a three-year average. Can I still use the weight for tabulation and estimation? e.g the perwt? Thanks!
It is possible to limit the years of a multiyear file. To do this, you’ll need to use the MULTYEAR variable, which identifies the actual year of the survey. Weighting this new ‘restricted’ multiyear date file, however, may be tricky and potentially problematic. The reason for this is the PERWT variable included in the 2010-2014 5-year sample is designed to represent the entire 5-year period. What might be best is to construct your new restricted data file by combining the 2012, 2013, and 2014 single year ACS files. Through this method the PERWT variable is designed to represent the true population in a given year. When these three single-year files are combined, you’ll need to divide the PERWT varaiable by 3 in order to calculate accurate population count for that three year period.