No matter how successfully Elon Musk and his Department of Government Efficiency (DOGE) identify the waste in federal spending, their achievements will not be secure without a closer look at how Washington became so profligate in the first place. True success requires looking beyond the obvious causes—self-interested lobbying by public unions and the reluctance of politicians to terminate any program with a vocal constituency—but also the false, academically generated worldviews designed to persuade voters that their government is competent to do so many things it cannot.

　　It was the late social philosopher Irving Kristol (1920–2009) who first described this problem in a series of essays written during the mid-1970s. An enthusiastic supporter of President Lyndon Johnson’s Great Society programs a decade earlier, he had become disillusioned with how little the billions spent to alleviate poverty and educate minorities had really achieved.

　　True, the poverty rate for African Americans had fallen a few points over the years, but on average, it had remained essentially unchanged. And the over 100 programs designed to encourage black children to stay in school and to keep poor families intact seemed to have benefitted the government administrators far more than the intended beneficiaries. We were “feeding the horses to feed the sparrows,” as another disillusioned reformer, New York Sen. Daniel Patrick Moynihan (D), sarcastically put it.

　　How, Kristol wondered, was it possible for US taxpayers to be persuaded to subsidize such an expensive campaign to benefit disadvantaged citizens when, as it turned out, government had no useful plan?

　　The answer, he came to see, was the inevitable temptation for economists, healthcare experts, psychologists, and other social scientists to produce findings that validated the desire of their largest funder, the federal bureaucracy, to grow its responsibilities. The more America’s academic researchers—what he called the “New Class”—could persuade the public of their government’s competence to solve various problems, the more federal agencies would continue to subsidize their activities.

　　Importantly, Kristol did not believe that all, or even most, social scientists were intentionally falsifying their data, but rather that the imperfect methodologies of social science (as compared to, say, those of physics) allowed a researcher’s self-interest to have a subtle sway over study outcomes. Just as the priest of ancient times managed to reconcile his high-minded faith with his sovereign’s material ambitions, so the modern social scientist is unconsciously moved to seek truth in ways that end up pleasing his patron.

　　Nevertheless, Irving Kristol’s portrayal of American social science as the marketing arm of an expanding public sector did not go over well with either academic researchers or their DC sponsors. If President Johnson’s vision of a more paternalistic government seemed to be floundering, they countered, it was not because the underlying scholarship was unconsciously biased, but because it was still “evolving.”

　　Indeed, it was not until decades later, in 2005, when Kristol’s explanation for why the government ends up wasting so much money finally got some backing from within the academy itself. That was when Dr. John Ioannidis, co-director of Stanford University’s Meta-Research Innovation Center, published a now famous paper showing that many influential study outcomes, long considered “settled social science,” could not, in fact, be replicated. In other words, by the ultimate test of scientific validity—the ability to get the same result when repeating the same experiment—up to half what was widely held to be true about human behavior had never really been proved.

　　It was not long after Ioannidis’ shocking revelation that Lancet, Nature, Science, and other respected journals began testing many of their own contributors’ studies, only to find that much of what they had previously published was either wrong or, at best, misleading. Even worse, it was becoming clear that the unverifiable studies did not occur randomly but tended to be those that suggested the need for more government regulation, especially in fields like the environment, race relations, and medical care. As National Association of Scholars president Peter Wood put it: while “not all irreproducible research is progressive advocacy [and] not all progressive advocacy is irreproducible, the intersection between the two is very large … [and] a map of much that is wrong with modern science.”

　　If there is any good news about this abuse of social science in the service of wasteful government policy, it is that the remedy lies within science itself.

　　It was also becoming clear to observers like Princeton University Physics Professor Emeritus William Happer that it was the politically biased studies that often had the easiest time getting a federal subsidy. “For many years,” he noted, “researchers willing to demonize carbon dioxide, low-level radiation, meat products, etc., have benefited from generous funding by governments … [but] almost none of it is reproducible.”

　　Around the same time, Hoover Institution economist John F. Cogan similarly documented in his book, The High Cost of Good Intentions, that most of the supposedly scientific research Washington generously sponsored over the years to justify expanding its entitlement programs had proved “consistently wrong.” In fact, he wrote, there is “precious little evidence to support the contention that [any] social welfare services … prevent welfare dependency or help existing recipients achieve self-sufficiency.”

　　Perhaps the most famous example of ready government funding for unreliable research to support a progressive priority occurred in 2009. That was when the National Highway Traffic Safety Administration sponsored a series of evaluations to determine how much automobile pollution could be reduced by paying people to trade their old, inefficient cars for newer, fuel-efficient ones. This led to a $3 billion Car Allowance Rebate System (CARS), or so-called “Cash for Clunkers” program, which did so little to reduce emissions that it soon had to be abandoned.

　　If there is any good news about this abuse of social science in the service of wasteful government policy, it is that the remedy lies within science itself. For while the measurement of human behavior will never be perfect, statisticians have developed some very sophisticated techniques over the years which can force any study to move closer to the truth, no matter how unhappy that might make either its author or its sponsor. As Kristol himself understood, what makes politically slanted research possible is not deliberate fraud but the use of methods sloppy enough to permit an experimenter’s economic self-interest to color a supposedly objective exercise.

　　DOGE can undoubtedly help the country by cleaning up so many of the wasteful programs inspired by the bad science of the past. But agency heads could do even more by insisting that any future social science research their departments sponsor adhere to the strictest experimental standards.

　　These would include having a higher threshold for what is currently considered a “statistically significant” finding, making data and research protocols publicly available, pre-registering study protocols and reporting any mid-stream alteration of them, and using large sample sizes. It would also help if investigators who have met these standards in their previous work were given a funding preference and a significant percentage of the estimated $8.4 and $10.5 billion spent annually on social science research dedicated to replication studies.

　　If the past is any guide, many in the academic community will likely not be pleased with such reforms. In February of 2020, when the Independent Institute sponsored the firstnational conferenceon “Practical Solutions for the Irreproducibility Crisis,” the response on social media was decidedly negative, with the event labeled as everything from misogynistic to white supremacist to climate change denialism. More telling was the fact that two graduate students set to speak at the conference had to withdraw out of their concern for career sabotage.

　　But the rigorous standards needed to tame the problem of social science’s big government bias will not prevent any professor from studying any subject in any way he or she wants, just not at taxpayer expense. Nor will they prevent honest research from yielding outcomes that suggest a new or expanded government program … if that is in fact what they suggest.

　　What the higher standards will do is stop federal bureaucrats and their academic collaborators from inventing a reality that fosters wasteful and even destructive public spending.