PURPOSE: The validity of parental occupation recalled by adult children is not established, yet it is commonly used to measure childhood socioeconomic status (SES). We investigated the feasibility of using data from historical records to validate recalled parental SES. METHODS: Data from death certificates and applications for Social Security numbers (parents' names, date and place of birth) were used to locate birth certificates and 1930 census records of 416 decedents in Forsyth County, NC, to verify parental occupation and childhood residence. RESULTS: Birth certificates and/or census records were located for 85% of decedents. Of 257 for whom both records were searched, both were found for 60%, only a census record for 10%, and only a birth certificate for 24%. Among those with father's occupation recorded on both records (n = 138), occupational category matched on 89% of records (kappa = 0.86). Place of residence/birth, which can be linked with census-based county socioeconomic indicators, was also highly concordant across records. CONCLUSIONS: These results demonstrate that birth and census records can be located for most decedents and that the childhood SES data contained therein is highly concordant. Thus they are an alternative to recalled childhood SES and a source of validation data in life course studies.
PURPOSE: The validity of parental occupation recalled by adult children is not established, yet it is commonly used to measure childhood socioeconomic status (SES). We investigated the feasibility of using data from historical records to validate recalled parental SES. METHODS: Data from death certificates and applications for Social Security numbers (parents' names, date and place of birth) were used to locate birth certificates and 1930 census records of 416 decedents in Forsyth County, NC, to verify parental occupation and childhood residence. RESULTS: Birth certificates and/or census records were located for 85% of decedents. Of 257 for whom both records were searched, both were found for 60%, only a census record for 10%, and only a birth certificate for 24%. Among those with father's occupation recorded on both records (n = 138), occupational category matched on 89% of records (kappa = 0.86). Place of residence/birth, which can be linked with census-based county socioeconomic indicators, was also highly concordant across records. CONCLUSIONS: These results demonstrate that birth and census records can be located for most decedents and that the childhood SES data contained therein is highly concordant. Thus they are an alternative to recalled childhood SES and a source of validation data in life course studies.
Authors: Calpurnyia B Roberts; David J Couper; Patricia P Chang; Sherman A James; Wayne D Rosamond; Gerardo Heiss Journal: Am J Epidemiol Date: 2010-08-08 Impact factor: 4.897