Is SERIAL code (household serial number) the same across different years?

I am trying to analyze the pattern of household across different years in Puerto Rico. I am wondering can I track the changes of a single household across years by using its serial number? Again taken Puerto Rico as an example, if I only keep the one sereial number for each year, should the PUMA code vary in different years? So does this mean that the household with the same serial code does not necessarily to be the same one in different years?

Please refer to my attached excel output generated by the following sas code, where the data ‘incomep’ means the original data.


data series; set incomep; keep serial year costwatr puma;run;

proc sort data=series nodup; by serial;run;

proc print data=series (obs=220);run;


From the output in the excel file, you can see that the PUMA code varied quite often, and also the COSTWATR variable also varied significantly. I’m not quite sure what does the Comparibility and Universe for this variable (SERIAL) mean as showed in the website.

Thank you very much for your help in advance.

output for Serial-1.xlsx (18 KB)

Hi, in the Frequently Asked Questions, it says:

“Three variables constitute a unique identifier for each household record in the IPUMS: YEAR, DATANUM, and SERIAL (year, data set number, and household serial number).”

So, based on this, is there a way that I can track the changes of variables for a household level across years, for example household income, electricty cost, etc?

The variable SERIAL is created by IPUMS-USA for each data set and is only meant to uniquely identify households within a single dataset. SERIAL cannot be used to identify households across samples.

I hope this helps.