Response Bias in Children's Reports of Parental Status Characteristics

Abstract

There is increasing evidence of the differential accuracy of children's reports of parental status characteristics (Colfax & Allen, 1967; St. John, 1970; Borus & Nestel, 1971; Kayser & Summers, 1973). These studies point to differential abilities of children to accurately replicate parental reports of their income, education, and occupation. Borus and Nestel (1971) using son's and father's reports of father's education found identical reports more frequent for whites, for sons in school, for younger sons, for better educated sons, for sons from smaller families, and for sons from less populated areas. St. John (1970) in addition to finding low accuracy of reports of parental education, found that the sixth grader non-response rate was 45% for father's education and 33% for mother's education. Non-response was associated with being female, being black, and with low mother's education. Overreporting education was associated with being male, with being a poor reader, with low actual parental education, with the presence of a report of the father's education, with the presence of a report of education for the same-sex parent, with having a father absent, and with being black. Colfax and Allen (1967) reported non-response rates of only 6% on father's occupation for sixth graders, and no missing data for similar reports by ninth and twelfth graders. All students were found to upgrade the father's occupational status if permitted the opportunity to give other than open-ended answers. Kerckhoff, Mason, and Poss (1973) attempted to replicate and expand the findings of St. John (1970) by using a sample of 6th, 9th, and 12th grade boys and interviewing both parents. Non-response was low and inversely related to the grade level of the student. In addition, increased response rate was not associated with increased error. Percentage of identical reports of parental education based on matching the son's and parent's reports ranged from 45% to 82% for mother's education and from 47% to 73% for father's education. Their study found minimal evidence of upgrading of status characteristics, with less dispersion of responses for whites. With increases in age levels of the groups, there was evidence of increased predictability in the son's estimates. However,