In the 1984, 1990, 1991, 1992, and1993 SIPP panels, a value for ethnicity is included on each wave file. Thus, there are actually as many ethnicity variables as there are waves of the survey and some changes occur across waves as a result of data collection error. Ethnicity is chosen by creating an array of variables ethncty1-ethncty{max number of waves} and choosing the first non-missing value. Thus, ethnicity comes from the first wave in which the individual was interviewed instead of from a fixed point in the survey. Respondents are coded as Hispanic if they have an ethnicity code between 14 and 20. In the 1996, 2001, 2004, and 2008 panels, the longitudinally-edited version contains only one value for ethnicity across all waves (EORIGIN) and this value is used. Respondents are coded as Hispanic if they have an ethnicity code between 20 and 28 in 1996 and 2001, or if they have an ethnicity code of 1 in 2004 2008.