How much do we count? Interpretation and error-making in the decennial census

Following a critique of the 1990 decennial census procedures, we conducted a field study among low-income, inner-city residents in 1991 to examine how they conceptualized and managed the civic task of census response. Interpr etations about the purpose and meaning of the census, about commitment to t he task, and about connection to government, singly and together with liter acy skills (e.g., reading and general literacy competence), were associated with errors that are not detectable by evaluative methodologies used regul arly by the Census Bureau. The validity and reliability of census data, and possibly other self-administered survey research, will be increased by gre ater use of knowledge about both interpretation and literacy skills in form ulating data collection procedures.