One common refrain in debates around education is that standardized exams negatively impact applicants from disadvantaged backgrounds. In recent years, many colleges and selective K–12 programs around the U.S. have been limiting or abandoning these measures—a trend only accelerated by the Covid pandemic.
The use of standardized exams, however, may function to replace other, more-discretionary criteria like subjective personality assessments or personal and family connections when judging applicants. Indeed, these assessments were devised in the twentieth century partly as a way to minimize the influence of prestigious boarding schools, social class, and geography in college admissions. This meritocratic impulse built on an earlier push in the late nineteenth century to use standardized tests for awarding government jobs.
A new NBER study draft from University of California, Davis economists Diana Moreira and Santiago Perez explores the social and equity impacts of the 1883 Pendleton Act, which mandated that federal civil servants be hired on the basis of competitive exams, rather than the “spoils” system of political patronage used before. What can we learn from studying its effects?
The analysts constructed a new dataset over the period 1871–93 by matching federal employee names to census records, including their sex, race, immigration history, and geographic origin. Moreira and Perez also managed to match thousands of employee files to information related to socioeconomic class, including the individual civil servants’ former careers, families’ careers, and approximate household wealth. They then divided careers—both individual and family—into five categories based on education levels: professional, white collar, farmer, blue collar, and unskilled laborer.
They then created a “social background index” aggregating information on parental wealth, parental occupations, and demographics like race and immigration history, such that a low score corresponds to a disadvantaged background. Then they divided employees into a “treated” group, those subjected to testing beginning in 1883, and a “control” group of federal employees still hired by discretion, who were usually at the top and bottom of the pay scale.
Moreira and Perez find that after the introduction of testing, the median social background index score for treated federal employees rose by 0.17 points. In real terms, this captures the researchers’ observation that, out of a pool of thousands of employees, from 1881 to 1885, the adoption of standardized testing for mid-level federal civil servants is associated with a shift in their background away from immigrant and upper- and lower-class Americans and toward native-born and middle- and upper-middle (professional-) class Americans.
The researchers interpret this to mean that the Pendleton Act reforms increased the share of middle-class “educated outsiders” in the federal bureaucracy. While the legislation reduced the share of individuals from disadvantaged backgrounds entering the federal workforce, Moreira and Perez suggest that this may have been a consequence of the reduced influence of the “spoils” system for awarding federal jobs, which favored political connections like Irish-immigrant urban machines such as New York’s famous Tammany Hall. The researchers also mention that, although the Civil Service Commission argued that the Pendleton Act reforms would increase the number of African Americans employed, statistical estimates find no impact or change in racial representation in the federal workforce.
As for how this may relate to standardized testing in American education, Moreira and Perez argue that their findings point to an equity-efficiency tradeoff, in that the use of standardized exams may improve the quality of the bureaucracy at the cost of “equitable” or diverse representation across social backgrounds. Yet this is dubious. To validly constitute such a tradeoff, you’d have to somehow lose out on “fairness” to gain efficiency. If the researchers believe that adopting standardized exams in 1883 cost society fairness, they are embedding several assumptions into this welfare analysis: that the federal bureaucracy in 1885 was better than in 1881; that innate talent was uniformly distributed along class, education, and immigration lines; that the “spoils” system better reflected this uniform distribution than standardized tests; and that the federal government should have targeted “innate” talent rather than demonstrated credentials.
Nevertheless, Moreira and Perez’s exploration of the social impacts of early standardized testing in the United States is thought-provoking, in that it points to the way hiring (or admissions) decisions are never made ideally, but are instead made under limited information, using “noisy signals” of candidates’ abilities like credentials and recommendations. Deciding which sets of information to include or exclude from such decisions may have powerful social implications. If nothing else, Moreira and Perez’s study should offer some context to the public policy debates around the use of standardized testing for K–12 and higher education admissions, affirmative action, equality, equity, race, and class in America today.
SOURCE: Diana Moreira and Santiago Pérez, Who Benefits from Meritocracy?, NBER (2021).