A Good Guy With a Gun
“The only thing that stops a bad guy with a gun is a good guy with a gun.”
Armed bystanders do sometimes stop attacks, but rarely. Of 160 active shooter incidents the FBI studied from 2000 to 2013, 5 ended when armed people who were not police fired back; 21 ended when unarmed people restrained the shooter. And CDC data show 62 percent of American gun deaths in 2024 were suicides.
Wayne LaPierre, then the head of the National Rifle Association, gave the line its modern form at a press conference one week after the shooting at Sandy Hook Elementary School in December 2012. In its strongest version the argument says that police take minutes to arrive, that an armed citizen is already on the scene, and that Americans use guns to protect themselves far more often than the news reports.
That last point is a genuine dispute among researchers. Estimates of defensive gun use run from roughly 100,000 a year in the federal government’s crime victim survey to 2.5 million in a 1990s survey by the criminologist Gary Kleck, a gap of twentyfold or more that the RAND Corporation traces to how each survey defines a defensive use and how it asks about one. Anyone who quotes a single number is choosing a side of that dispute.
On mass attacks, the record is narrower than the slogan. The FBI’s study of 160 active shooter incidents from 2000 to 2013 found that 5 ended when armed people who were not police exchanged fire with the shooter, while 21 ended when unarmed people restrained the shooter. Armed bystanders do sometimes stop attacks, and that part of the claim holds. Laws that make it easier to carry a gun in public have a weaker record: the most detailed recent study found violent crime rose in the states that adopted them.
The larger problem is that a gun bought for protection is also in the house on the worst night of someone’s life. In 2024, 62 percent of American gun deaths were suicides, and a study that followed 26 million Californians found that men who acquired a handgun died by gun suicide at nearly eight times the rate of men who did not.
Gun Laws Don’t Work
“Chicago has the toughest gun laws in the country and it’s a war zone. Gun laws only disarm law-abiding people.”
Chicago’s laws stop at the city line. The city’s 2017 trace report found 60 percent of guns recovered in crimes came from dealers in other states, more than 20 percent from Indiana alone. Statewide, CDC data for 2024 put the lowest gun death rates in Hawaii, Massachusetts, New Jersey, and New York, all strict-law states.
The Chicago version of the argument has one strong fact behind it: the city has had strict local gun rules and a serious gun violence problem at the same time. What it leaves out is where the guns come from. The city’s 2017 gun trace report found that most guns recovered in Chicago crimes were first sold outside Illinois, many of them a short drive away in Indiana. A city cannot wall itself off from its neighbors’ laws.
Comparing whole states is the fairer test, and it cuts the other way. In 2024, according to CDC figures compiled by the Pew Research Center, the states with the highest gun death rates were Mississippi, New Mexico, Alaska, Alabama, and Wyoming, and the lowest were Hawaii, Massachusetts, New Jersey, New York, and Rhode Island. Gun laws are not the only difference between those states, but the pattern runs opposite to the talking point.
Abroad, the case most often cited is Australia’s 1996 National Firearms Agreement, which banned rapid-fire long guns and bought back hundreds of thousands of them after the Port Arthur massacre. Australia went two decades without a mass shooting of the kind it had seen 13 times in the 18 years before, though the researchers who documented that were careful to say the law alone could not be proven to be the cause. At home, the federal assault weapons ban of 1994 to 2004 produced split findings: the Justice Department’s own evaluation expected small effects on gun crime overall, while a later study found mass shooting deaths were lower while it was in force.
The Second Amendment framing changed in 2008. In District of Columbia v. Heller, the Supreme Court held for the first time that the amendment protects an individual right to keep a handgun at home, while writing that the right, like most rights, is not unlimited. Later cases set a history-based test for gun laws and then upheld a ban on guns for people under domestic violence restraining orders. Meanwhile, since 2020, guns have been the leading cause of death for American children and teenagers.
Voter Fraud and Voter ID
“Elections are full of fraud, mail ballots make it easy, and voter ID is the only thing stopping it.”
The Heritage Foundation’s own fraud database, built to prove the problem, listed 1,298 cases going back to 1979 when FRONTLINE reviewed it in 2020: across four decades of elections. In 2021 a Republican-commissioned hand review in Arizona’s largest county found Biden had won by slightly more than the official count.
The strongest form of the argument goes like this: most states never ask voters to prove who they are, mail ballots leave the supervision of the polling place, and the Heritage Foundation keeps a database of proven fraud cases that runs well into four figures. Every one of those points is true as far as it goes.
The trouble is what the database shows. It is a real list of real cases, built by people who believe fraud is a serious threat, and it is also the best evidence of how rare fraud is. When FRONTLINE and USA Today reporters went through it in 2020, they found 1,298 entries reaching back to 1979, most involving a handful of votes. Mail voting has produced real cases, most famously a 2018 congressional race in North Carolina that had to be rerun, but the MIT Election Data and Science Lab finds no reason to think fraud is more common in states that vote by mail.
After 2020, the claim was tested directly. A review commissioned by Arizona Senate Republicans and run by a firm with no election experience recounted Maricopa County’s ballots by hand and found Joe Biden had won by slightly more than the official tally. The broader story of how the stolen-election claim was built is told in the it-worked-last-time series.
Voter ID is where the argument is strongest. ID requirements poll well across parties, and the fear that they would crush turnout has mostly not come true: a study of 1.6 billion voter records from 2008 to 2018 found strict ID laws had no measurable effect on registration or turnout for any group. An earlier Government Accountability Office study did find turnout drops of two to three points in Kansas and Tennessee in 2012. What neither side’s research shows is that ID laws stop much fraud, because the kind they target, someone pretending to be another voter at the polls, was already rare.
Makers and Takers
“Welfare pays people not to work, and it’s riddled with fraud.”
Most people on these programs are children, older adults, or working. USDA figures for 2024 show about 39 percent of SNAP recipients were children and 20 percent were 60 or older. KFF found that in 2023 most Medicaid adults were working, and nearly all the rest were caregiving, ill or disabled, or in school.
The modern version of the argument traces back to Ronald Reagan’s 1976 campaign speeches about a Chicago woman who cheated welfare out of a fortune. The strongest form says that aid without conditions discourages work, that taxpayers carry people who could support themselves, and that the programs leak money to fraud. The history of how welfare became a coded racial message is told on the site’s The Code Word page.
Start with who receives the help. In 2024, according to the Department of Agriculture, about 39 percent of SNAP participants were children and 20 percent were 60 or older; among SNAP households with children, more than half had earnings from work. In Medicaid, KFF found that nearly two-thirds of adult enrollees were working in 2023, and most of the rest were caregiving, sick or disabled, or in school.
Fraud exists, and the woman in Reagan’s story was real. But the widely quoted error rates for these programs measure mistakes, in both directions, not theft: SNAP’s payment error rate counts underpayments along with overpayments, most of them paperwork and calculation errors by states or households.
Work requirements are the test case. When Arkansas became the first state to enforce one in Medicaid in 2018, about 18,000 people lost coverage within months, and researchers found no gain in employment. The 2025 budget law makes a national version mandatory starting in 2027, and the Congressional Budget Office expects millions more uninsured.
Source: USDA Economic Research Service, SNAP Key Statistics, fiscal year 2024.
The Climate Has Always Changed
“The climate has always changed. Scientists disagree, they predicted an ice age in the 1970s, the models have been wrong, and China is the real problem anyway.”
The climate has always changed, but not this fast: NASA says the past century’s warming ran about ten times faster than warming after the last ice age. A review of 88,125 studies found over 99 percent agreement that humans are the cause. And Americans emit about 14 tonnes of carbon dioxide per person; China, about 9.
The climate talking point is really five claims stacked together, and each has a kernel of truth. The climate has changed before, many times. That is how scientists know what drives it. NASA’s Earth Observatory notes that as the world came out of the last ice age, it warmed by 4 to 7 degrees Celsius over about 5,000 years; the warming of the past century has run roughly ten times faster. The United Nations’ climate panel calls human influence on the warming “unequivocal”.
“Scientists disagree” is no longer true in the literature. A 2021 review of 88,125 climate papers estimated agreement above 99 percent. The doubt was organized: a 2002 Republican strategy memo by the pollster Frank Luntz advised keeping the science looking unsettled, a story the site’s Doubt Is Our Product page traces back to the tobacco industry.
The 1970s ice age is mostly a myth: a survey of the research of that decade found far more papers predicting warming than cooling. The models have held up too: a 2020 evaluation of 17 projections made from 1970 to 2007 found most matched what happened.
China is where the claim is partly right. China is the world’s largest emitter, with about a third of global carbon dioxide from fossil fuels in 2024. But per person, Americans emit far more, and the United States has put more carbon into the air since 1750 than any other country.
Source: Our World in Data and Global Carbon Project, fossil fuel and industry emissions, 2024.
A Christian Nation
“America was founded as a Christian nation, and the separation of church and state is a modern invention.”
The founders were mostly Christians, but they did not found a Christian government. The Constitution never mentions God and bans religious tests for federal office. In 1797 the Senate approved, and President John Adams ratified, a treaty whose English text says the United States government is not in any sense founded on the Christian religion.
This one is popular: a 2022 Pew survey found 45 percent of Americans say the country should be a Christian nation, and 60 percent believe the founders intended one. Its strongest advocates, such as David Barton’s WallBuilders, point to the founders’ own words about faith, and there is plenty to quote.
The record supports part of that. Most founders were Protestant Christians, Congress opened with prayer, and several states kept tax-supported churches after independence. Jefferson and Madison, the founders most associated with separation, attended Sunday services held in the House of Representatives.
But the federal government they built was deliberately secular. The Constitution never mentions God or Christianity and forbids any religious test for federal office. The Treaty of Tripoli, approved by the Senate and ratified by President John Adams in 1797, says outright that the government is not founded on the Christian religion. And the phrase “wall of separation between church and state” is not modern: it comes from Jefferson’s 1802 letter to the Danbury Baptists, a minority church that wanted protection from the established churches of its own state.
How the Christian nation idea became a political program, through Christian Reconstructionism and the Quiverfull movement, is the subject of the site’s Joshua Generation page.
Sports, Bathrooms, and Medical Care
“Men are taking over women’s sports, predators are using women’s bathrooms, and doctors are transitioning children.”
The fairness question in sports is real, but small in scale: in December 2024 the NCAA’s president told a Senate committee there were fewer than 10 transgender athletes among its 510,000. And a 2024 study of insured Americans found no gender surgeries on children 12 or younger, and very few on older teenagers.
Transgender policy is where the talking points mix the strongest arguments with the weakest. In its best form, the argument says that male puberty gives lasting physical advantages, that women’s sports exist to remove those advantages, and that the medical evidence for treating minors is thinner than its supporters admit. Each of those claims has serious support.
On sports, a 2020 review by two sports scientists found that a year of testosterone suppression reduces transgender women’s muscle mass and strength by only about 5 percent. Other researchers dispute how much those findings apply to actual athletes, but the fairness concern is a real one. The scale is not: the NCAA’s president said in December 2024 that fewer than 10 of its 510,000 athletes were transgender. In February 2025, an executive order directed federal agencies to bar transgender girls and women from female sports.
On medical care, England’s Cass Review called the evidence for treating young people with puberty blockers and hormones weak, and its conclusions are still debated. But the picture of children being operated on is wrong: a study of 22.8 million insured minors found no gender surgeries on children 12 or younger. Regret appears rare in the best long-term data, a Dutch cohort in which 98 percent of adolescents who began treatment were still using hormones years later. In June 2025, the Supreme Court upheld Tennessee’s ban on such treatment for minors.
On bathrooms, the danger mostly runs the other way: a 2019 study of 3,673 transgender teenagers found those barred from restrooms matching their identity were more likely to be sexually assaulted. Short, personal conversations change minds on these questions more than statistics do, as a field experiment in South Florida showed.
The Liberal Media and Free Speech
“The mainstream media is liberal, and conservatives are being silenced.”
Most journalists do lean left: a 2022 survey found about 36 percent call themselves Democrats and 3 percent Republicans. But a 2020 study in Science Advances found no bias in which political stories journalists choose to cover, and Twitter’s own experiment found its algorithm amplified the political right more than the left in six of seven countries.
Half of this talking point is correct. Surveys of American journalists show they identify as Democrats far more often than as Republicans, and the gap has grown. Conservatives are right to notice it.
Whether that changes the news is a separate question, and the evidence is mixed. A large 2020 study of political journalists found no bias against conservatives, or liberals, in which stories they chose to cover. Meanwhile the media people actually see is not all mainstream newsrooms: Sinclair Broadcast Group owned more than 190 local television stations in 2018, and Twitter’s own randomized experiment found its algorithm amplified right-leaning politicians and news more than left-leaning ones.
On free speech, the First Amendment limits the government, not employers, platforms, or audiences. Campus speech is a real problem across the spectrum: FIRE’s 2026 rankings found record shares of students saying it is acceptable to shout down a speaker. The clearest recent threats, though, have come from government power. In 2024 the Supreme Court ruled unanimously for the National Rifle Association in a case against a New York regulator, holding that officials cannot coerce private companies to punish speech they dislike. In September 2025, ABC pulled Jimmy Kimmel’s show after the chairman of the Federal Communications Commission warned broadcasters they could do it “the easy way or the hard way.” The principle in the NRA case applies to both.
How a Talking Point Is Built
Talking points are not folk wisdom. Many are designed, tested, and distributed. A useful starting point is the memo Lewis Powell sent to the U.S. Chamber of Commerce in 1971, which urged business to fund a long campaign of ideas. Think tanks and advocacy groups became the factories, turning research into short lines that politicians and broadcasters could repeat.
Then comes the wording. The pollster Frank Luntz has described how he tests phrases with focus groups: “estate tax” became “death tax”, and his memo on the environment advised Republicans to keep the climate science looking unsettled.
Repetition does the rest. The illusory truth effect means a claim heard many times feels true, even to people who know better. The site’s It Feels True page shows how this is used on social media.
How to Answer in One Sentence
A slogan takes five seconds to say and five minutes to correct. Research on why corrections struggle explains the imbalance: a false claim keeps shaping how people reason even after they accept that it was wrong, unless the correction gives them something better to put in its place.
That suggests the shape of a good answer. Lead with the fact rather than the myth, so the true statement is the one repeated. Explain why the myth seems plausible. Better still, reach people before the myth does: short videos that teach the tricks of manipulation made viewers better at spotting them in a test with 22,632 YouTube users.
In practice, the strongest one-sentence answers to these talking points share three features. They concede what is true in the claim, which earns a hearing. They give one number with a named source. And they come from someone the listener already trusts, which is why a personal conversation beats a statistic.
Where the Evidence Cuts the Other Way
Honest fact-checking runs in both directions, and a few popular progressive talking points do not hold up as usually stated.
“Women earn 82 cents on the dollar for the same work” mixes two things. The raw gap is real, but it compares all working women with all working men, not people in the same job. The research of the economist Claudia Goldin, who won the 2023 Nobel Prize for it, finds that most of the gap now arises within occupations and opens with the birth of a first child, which points to how jobs reward long, inflexible hours rather than to employers paying women less for identical work.
“Assault weapons are driving gun deaths” is not supported by the data. FBI figures for 2024 show rifles of all kinds were involved in about 3 percent of gun murders; handguns were involved in more than half. The case for restricting them rests on mass shootings, not everyday violence.
“Nuclear power is too dangerous” also runs against the record: per unit of electricity, nuclear has caused far fewer deaths than coal, oil, or gas. The voter ID research above is another example, since strict ID laws did not cut turnout the way opponents predicted.
Socialism, Big Government, and States’ Rights
“Democrats want socialism. Look at Venezuela. Conservatives stand for small government and states’ rights.”
The Nordic countries Democrats point to are not socialist: in 2015 Denmark’s prime minister told a Harvard audience, “Denmark is far from a socialist planned economy. Denmark is a market economy.” They have private businesses, free trade, and high taxes that pay for universal programs. Venezuela nationalized hundreds of businesses and ended term limits; no Democratic platform proposes either.
“Socialism” in American politics usually means two different things. In the textbook sense, it means government ownership of the means of production. In current use, it often means a larger welfare state, such as universal health coverage, paid family leave, or free public college. The talking point gets its force by sliding from the second meaning to the first, and from Denmark to Venezuela.
The Nordic countries are market economies with large welfare states and high taxes, as Denmark’s prime minister explained at Harvard in 2015. Conservatives have a fair point here, and it is the same one he made: the services are paid for by taxes far higher than Americans pay, including on the middle class. The Nordic model is a trade-off, not free money. How those countries came to their systems is part of the site’s Norse and Vikings page.
There is also real enthusiasm on the left for the word itself. Gallup found in 2019 that about half of young adults viewed socialism positively, as many as viewed capitalism positively.
Small government and states’ rights are principles both parties invoke when they are out of power and drop when they hold it. The phrase “states’ rights” also has a history in the fight over segregation, told on the site’s The Code Word page, and the claim that Democrats are still the party of slavery and the Klan is answered in full on The Party Switch.