income. According to the authors of the ELS 2002/04 study, the SES variable is an,
“NLS-72/HS&B/NELS:88-comparable composite variable constructed from parent
questionnaire data when available and student substitutions when not” (Ingels et al. 2005:
18). It is worth mentioning that the ELS 2002/04 data set offers a choice between two,
slightly different, composite SES variables. One type of SES variable is based on the
Duncan Socioeconomic Index (SEI) scale for ranking various occupational titles.
Socioeconomic index scores were originally calculated by Otis Dudley Duncan based on
the National Opinion Research Center (NORC) 1947 North-Hatt prestige study and the
1950 Census (Davis, Smith, and Marsden 2007: 37). The second type of SES composite
is based on the 1989 NORC / General Social Survey (GSS) Occupational Prestige Scale.
In the ELS 2002/04 data set, the first SES variable is based on occupational codes
supplied by the respondents according to the older, 1961, Duncan SEI occupational
prestige scores. The second SES variable in the ELS 2002/04 data set is based on
respondent-supplied occupational codes derived from the 1989 NORC / General Social
Survey (GSS) occupational prestige scores. Given the similarities in the two available
SES variables, the decision was made to use the SES quartile variable which is based on
newer NORC/GSS based SES. Within the composite SES variable, occupational status
for each parent ranges from unemployed to professional, and parental education levels
range from lowest, those parents who did not finish high school, to highest, those parents
who completed degrees at the level of Ph.D. or M.D.