I had my hard drive formatted and had to download an old CPS dataset again. After this, some regression results have changed a little. I notice that in my old STATA dofile, the WTSUPP variable is said to occupy 8 bytes, while in the new file it is 10 bytes. Therefore I wonder if the WTSUPP variable has been modified between my two downloads? The first download was in December 2011.
Yes, WTSUPP was lengthened to accommodate additional implied decimals in our new supplement data files. However, the additional digits in the March WTSUPP values should always be zero, and so I don’t believe this would affect your regression results.