Date of Award
The norm samples that are developed in Westernized countries are typically created using a routine process in test development that involves stratification of a range of independent variables. However, this method of creating norm-referenced samples may be discriminatory against individuals from low SES backgrounds since sampling across the full range of SES invariably results in an “average” that characterizes middle-SES individuals best. Because SES is generally stratified by category, according to U.S. Census demographics, any such sample developed in this manner is likely to be appropriate only for individuals within the “average” SES range. The literature contains a wealth of knowledge indicating that SES and development are highly related and that individuals from higher SES backgrounds have advantages that individuals from lower SES backgrounds do not. As such, use of a middle-class SES standard for sampling purposes may not provide an accurate comparison regarding development for lower SES individuals as it would for those having been raised in more average SES homes and would likely underestimates their abilities. The present study was interested in investigating the appropriateness of norm samples for lower SES individuals by examining whether individuals from lower SES backgrounds would perform significantly differently on a measure of language development than those from higher SES backgrounds. It was hypothesized that the mean receptive language score on a test of language development for the lower SES group based on maternal education or family income would be significantly and meaningfully lower than the score for the higher SES group based on maternal education. Independent-samples t-tests were conducted to compare language development scores in the higher SES group and the lower SES group. Overall, contrary to what was predicted, findings of this study did not support the hypothesis that the lower SES group based on annual family income and or maternal education would be significantly and meaningfully lower than the mean language development score for the higher SES group based on annual family income and or maternal education. However, when examining English Learners alone and when dividing the higher and lower SES groups at $65,000 for annual family income and 14 years for maternal education a significant difference between the groups was found. As such, it seems plausible that the stratification of SES by sampling across the range as a way of controlling for presumed differences to create middle-class representation is likely to be discriminatory for lower SES individuals and may require a different procedure to assure fairness and equity in testing. It is important for school psychologists to recognize that SES likely impacts language development performance at and below a certain income level and maternal education level and that use of tests with norm samples that are constructed via aggregation of SES levels may not be valid for use with individuals below a particular threshold of maternal education and family income.
Franco, Gavriel, "NORM SAMPLE STRATIFICATION AND SOCIOECONOMIC STATUS: EXAMINATION OF LANGUAGE DEVELOPMENT ACROSS EDUCATION AND INCOME" (2022). Theses and Dissertations. 419.