Accuracy of Student-Reported Socioeconomic Status: Results From the NAEP/Early Childhood Longitudinal Program—Kindergarten Linking Study

• In Event: 54.032 - Measurement of Socioeconomic Status for American Education Surveys

Abstract

The existing National Assessment of Educational Progress (NAEP) contextual questionnaires contain a set of questions designed to capture information related to household income, such as how many books they have at home, and whether they have a dishwasher and a clothes dryer. As stated in the first paper of this symposium proposal, NAEP is also interested in evaluating the quality of student reported parental educational attainment for 4th-graders and parental occupational status for all grades. The dependency on student self-report poses validity concerns for the measurement of SES, especially at grade 4, due to the age of the students. In particular, there is concern that grade 4 students may be unable to accurately report parental occupation and education.

As part of the Early Childhood Longitudinal Program-Kindergarten Class of 2010-2011 (ECLS-K) grade 4 data collection in 2015, parents of students were interviewed for their occupation, education, and household income. In 2015, approximately 1,500 students took part in both the NAEP and ECLS-K assessments. During the NAEP data collection, the students in the study were given both standard NAEP contextual questionnaires and an extended NAEP student questionnaire that contained questions about parental education and occupation. The aim of the present study was to evaluate the accuracy of grade 4 student reported parental occupation, parental education, and household income related questions in NAEP, using the ECLS-K parent-reported occupation, education, and household income as a reference.

The study found that students could accurately report parental occupation about half (48%) of the time. For 16 of the 21 student-reported parental occupational categories, less than half of the students within the parental occupation categories actually had the reported parental occupation. It was also found that students could accurately report parental education also about half (47%) of the time. Students with at least one college educated parent could accurately report this 75% of the time. However, students with both parents less than college educated were only accurate 14% of the time, instead responding “college” or “I don’t know” 69% of the time. For student-reported income-related survey questions, all questions were seen to relate to household income, however, there were large variation and overlap of household income within each income-related survey questionnaire response. National School Lunch eligibility was fairly accurate (84%) in differentiating students with household income of $65,000 above and below.

The consequences of the findings are discussed in terms of the feasibility of measuring SES with student self-report questionnaires, as well as what sort of self-report SES-related questions are appropriate for grade 4 students and specifically what information about parental occupation and education can be accurately obtained from student self-report.

Authors

• Paul A Jewsbury, Educational Testing Service

• Laura Jerry, Educational Testing Service