If blessed with excess applicants, how should colleges and universities screen and select those most likely to thrive academically, to graduate, and ultimately to vocationally succeed?   As a general rule, the best predictor of future behavior is past behavior. Thus, high-achieving high school students, as reflected in their grades, will tend to become high-achieving college students. And of course, high school GPA reflects students’ combined ability and diligence—the very traits that will predict their success in college and beyond. But then came grade inflation, with American applicants’ GPA now mostly ranging from 3.5 to 4.0, rather than the previous 2.5 to 4.0. And when the range of any predictor shrinks, so does its predictive power. Among women basketball players ranging from 5’0” to 6’4,” height will predict rebounds snagged, but much less so among those ranging from 6’1” to 6’4.” With high school GPA having become less usefully predictive, admissions officers have looked to other indicators of potential student excellence. Using a more holistic process, they assess the lucidity of students’ essays. They take note of applicants’ extracurricular music lessons, international travel, and volunteerism. They may prioritize students from elite high schools. And they may also prioritize legacy students—those with family ties to, or financial gifts to, their school. Such considerations privilege students from higher income families—those that can fund the best schooling, essay-writing coaches, and extracurricular enrichments for applicants who are less likely than lower-income students to be working after-school jobs. Thus, admissions officers have wished for a better way to identify overlooked talent in unexpected places. From that wish was born test-based selection and the SAT (formerly Scholastic Aptitude Test). Harvard’s Steven Pinker notes that in its initial phase, from the 1930s through most of the twentieth century, standardized testing was “the enlightened policy . . .  since it [could] level a hereditary caste system by favoring the Jenny Cavilleris (poor and smart) over the Oliver Barretts (rich and stupid).” In a second phase, from the early 2000s through the Covid era, critics, mindful of test score disparities among income and racial groups, many argued that the SAT and its cousin, the ACT (originally American College Testing), were biased—with content that favored privileged social groups. To increase student diversity and preclude discrimination, many colleges therefore went test-optional or even “test blind.” But standardized testing advocates cautioned against blaming tests for exposing unequal experiences. If, with malnutrition, young people suffer stunted growth, don’t blame the measuring stick that reveals it. If unequal past experiences affect future achievements, a valid aptitude test will detect such. And an unbiased test will have the same predictive validity—it will work equally well—for people of any social group. Such is the SAT. The SAT correlates about +.5 with first-year undergraduate GPA—and does so for Black students as for White students, for women as for men, for lower-income as for higher-income applicants. This is roughly equivalent to the predictive power of today’s high school GPA across a diverse sample of U.S. colleges and universities. (Before the increase in grade inflation, high school GPA was the better predictor.) So, for most American schools, high school GPA and the SAT (or ACT) both provide useful information. But among highly selective schools, for which applicant GPAs tend to be uniformly high, “high school GPA and class rank now offer little additional predictive power,” notes a new analysis by Dartmouth economists. Likewise, at the selective University of California, Berkeley, “test scores are currently better predictors of first-year GPA than high school grade point average.” Ditto the Ivy League and sister schools. The Dartmouth researchers also document how test-optional policies at selective schools “disproportionately harm” high-achieving applicants “from disadvantaged backgrounds.” In the absence of the standardized tests, selection becomes more arbitrary and subject to family privilege bias. How does one pick among all the eager 4.0 students? Thus, although most colleges in 2025 remain test-optional, MIT, Yale, Brown, and Dartmouth, among others, have returned to requiring standardized tests. In defense of this current third phase—judicious aptitude test use by selective schools or for competitive scholarships—educational researchers have reemphasized these key points: Aptitude tests work. They predict not just college grades, but college persistence through graduation, later vocational success, and even life longevity. Twelve-year olds with SAT scores approximating a top 5 percent college applicant have later earned doctorates at twenty times the normal rate, and have disproportionately produced patents and publications. Test preparation courses are minimally effective, mostly aiding the math component—for which pre-calculus math instruction is also beneficial. Household income better predicts applicants’ essay quality than does SAT scores. Thus shifting from SATs to other criteria such as essays can increase inequality of opportunity. Schools can still prioritize giving opportunity to all social groups and to enriching their campus with diversity. “Once we brought the test requirement back,” explained MIT admissions dean Stuart Schmill, “we admitted the most diverse class that we ever had in our history,” including 31 percent Black and Hispanic students. There is, to be sure, much more to academic and vocational success than academic aptitude and general intelligence. Conscientiousness matters. Grit matters. Social skill matters. Curiosity matters. Creativity matters. Courage matters. And aptitude matters. That is why the SAT serves a prosocial purpose when it enables identification of otherwise unnoticed talent. Such is the case of one West African student to whom I alerted my college—someone who, after earning near perfect SAT scores as a 16-year-old, is now excelling here in physics courses and research, and destined for a high-level STEM career. One New Yorker letter writer—the daughter of a single, uneducated immigrant—explained that the SAT was her springboard to her state’s flagship university, “and, from there, on to medical school. Flawed thought it is, the SAT afforded me, as it has thousands of others, a way to prove that a poor, public-school kid who never had any test prep can do just as well as, if not better than, her better-off peers.” Napkin.AI visual synopsis of this essay: David Myers, a Hope College social psychologist, authors psychology textbooks and trade books, including his recent essay collection, How Do We Know Ourselves? Curiosities and Marvels of the Human Mind.   ... View more

At long last, artificial intelligence (AI)—and its main subset, machine learning—is beginning to fulfill its promise. When fed massive amounts of data, computers can discern patterns (as in speech recognition) and make predictions or decisions. AlphaZero, a Google-related computer system, started playing chess, shogi (Japanese chess), and GO against itself. Before long, thanks to machine learning, AlphaZero progressed from no knowledge of each game to “the best player, human or computer, the world has ever seen.” DrAfter123/DigitalVision Vectors/Getty Images I’ve had recent opportunities to witness the growing excitement about machine learning in the human future, through conversations with Adrian Weller (a Cambridge University scholar who is program director for the UK’s national institute for data science and AI). Andrew Briggs (Oxford’s Professor of Nanomaterials, who is using machine learning to direct his quantum computing experiments and, like Weller, is pondering what machine learning portends for human flourishing). Brian Odegaard (a UCLA post-doc psychologist who uses machine learning to identify brain networks that underlie human consciousness and perception). Two new medical ventures (to which—full disclosure—my family foundation has given investment support) illustrate machine learning’s potential: Fifth Eye, a University of Michigan spinoff, has had computers mine data on millions of heartbeats from critically ill hospital patients—to identify invisible, nuanced signs of deterioration. By detecting patterns that predict patient crashes, the system aims to provide a potentially life-saving early warning system (well ahead of doctors or nurses detecting anything amiss). Delphinus, which offers a new ultrasound alternative to mammography, will similarly use machine learning from thousands of breast scans to help radiologists spot potent cancer cells. Other machine-learning diagnostic systems are helping physicians to identify strokes, retinal pathology, and (using sensors and language predictors) the risk of depression or suicide. Machine learning of locked-in ALS patients’ brain wave patterns associated with “Yes” and “No” answers has enabled them to communicate their thoughts and feelings. And it is enabling researchers to translate brain activity into speech.   Consider, too, a new Pew Research Center study of gender representation in Google images. Pew researchers first harvested an archive of 26,981 gender-labeled human faces from different countries and ethnic groups. They fed 80 percent of these images into a computer, which used machine learning to discriminate male and female faces. When tested on the other 20 percent, the system achieved 95 percent accuracy.   Pew researchers next had the system use its new human-like gender-discrimination ability to  identify the gender of persons shown in 10,000 Google images associated with 105 common occupations. Would the gender representation in the image search results overrepresent, underrepresent, or accurately represent their proportions, as reported by U.S. Bureau of Labor Statistics (BLS) data summaries?   The result? Women, relative to their presence in the working world, were significantly underrepresented in some categories and overrepresented in others. For example, the BLS reports that 57 percent of bartenders are female—as are only 29 percent of the first 100 people shown in Google image searches of “bartender” (as you can see for yourself). Searches for “medical records technician,” “probation officer,” “general manager,” “chief executive,” and “security guard” showed a similar underrepresentation. But women were overrepresented, relative to their working proportion, in Google images for “police,” “computer programmer,” “mechanic,” and “singer.” Across all 105 jobs, men are 54 percent of those employed and 60 percent of those pictured. The bottom line: Machine learning reveals (in Google users’ engagement) a subtle new form of gender bias.   As these examples illustrate, machine learning holds promise for helpful application and research. But it will also entail some difficult ethical questions.   Imagine, for example, that age, race, gender, or sexual orientation are incorporated into algorithms that predict recidivism among released prisoners. Would it be discriminatory, or ethical, to use such demographic predictors in making parole decisions?   Such questions already exist in human judgments, but may become more acute if and when we ask machines to make these decisions. Or is there reason to hope that it will be easier to examine and tweak the inner workings of an algorithmic system than to do so with a human mind?   (For David Myers’ other essays on psychological science and everyday life visit www.TalkPsych.com.) ... View more

Originally posted on April 22, 2014. Critics have used the SAT test redesign to denounce the SAT and aptitude testing.  The multiple choice SAT has “never been a good predictor of academic achievement,” Bard College president Leon Botstein argued in Time. Better, to “look at the complex portrait” of college applicants’ lives, including “what their schools are like.” said Colby College English professor Jennifer Finney Boylan in a New York Times essay. The SAT only measures “those skills … necessary for the SATs,” surmised New Yorker staff writer Elizabeth Kolbert.  In a new Slate essay, David Hambrick and Christopher Chabris, distinguished experimental psychologists at Michigan State University and Union College, rebut such assertions.  Massive data, they argue, show that •    SAT scores do predict first-year GPA, whole-college GPA, and graduation likelihood, with the best prediction coming from a combination of both high school grades and aptitude scores. •    SAT scores of 13-year-old predict future advanced degrees and income, much as kindred and strongly-related IQ scores predict job training and vocational success. •    In one famous nationwide sample, the IQ scores of Scottish 11-year-olds predicted their later-life longevity, even after adjusting for socioeconomic status. •    Although SAT scores are slightly higher among students from high income families, the SAT also provides an opportunity for students from nonelite public school to display their potential—rather than to be judged by “what their schools are like.”  Thus SAT scores, when compared with assessments influenced by income-related school quality, have a social levelling effect.  •    Test preparation courses often taken by higher income prep school students “don’t change SAT scores much.” Ergo, say Hambrick and Chabris, while other traits such as grit, social skill, conscientiousness, and creativity matter, too, “the idea that standardized tests and ‘general intelligence’ are meaningless is wishful thinking.” ... View more

Originally posted on October 7, 2014. The October APS Observer is out with an essay by Nathan, “Once a Psychopath, Always a Psychopath?” on people who “commit horrific crimes, experience little guilt or remorse, and then commit similar crimes again.” What is their potential for change, and how can we teach students about them? In the same issue, I offer “The Story of My Life and Yours: Stability and Change.” It’s a celebration of what I regard as one of the great studies in the history of psychological science...Ian Deary and colleagues’ discovery of the intelligence scores of virtually all Scottish 11-year-olds in 1932, and then their retesting of samples of that population up to age 90.  The bottom line:  our lives are defined by a remarkable stability that feeds our identity, and also by a potential for change that enables us to grow and to hope for a brighter future. ... View more