David French at National Review criticized Nicholas Kristof’s New York Times op-ed where he wrote that many other hazards like bathtubs, stairs, and lightning strikes are deadlier than foreign-born terrorism on U.S. soil. French is correct that there is a big difference between dying as the result of an accident and dying as a result of murder (intentional killing committed by another person). Murder is scarier than an accidental death so people are willing to tolerate more precautionary measures to prevent it. The costs of death appear to be the same to the victim but many risk analysts disagree. A 2010 report endorsed by the Department of Homeland Security (DHS) estimated the value of each life saved at $6.5 million but argued that $13 million was justifiable. Another report estimated the value of a statistical life at $15 million. People seem to intrinsically be repelled when the chance of being murdered in a terrorist attack is compared to your chance of being killed in a bathtub.

My preferred comparison is your annual chance of being murdered by a terrorist versus being murdered in a homicide. One in 3.6 million people were murdered in a terrorist attack by a foreign-born terrorist on U.S. soil per year from 1975 through the end of 2015 (one in 3.2 million per year for all terrorist attacks committed on U.S. soil by any perpetrator). One in 14,219 people per year were murdered on U.S. soil by a non-terrorist homicide during the same period. Comparing murder by terrorist and deaths by accident is useful for explaining the frequency of rare events. The understandable desire to compare the likelihood of infrequent yet scary events to other infrequent but less scary events is a useful academic, mathematical, or cost-benefit exercise but it is not effective at convincing people to more rationally view the real risk from terrorism. People understand there is an important distinction between those who die as a result of murder and those who die from accidents.

However, French’s criticism of Kristof for focusing on terrorism deaths caused by foreigners on U.S. soil is off base. French’s sarcastic comparison to World War II summarizes his position:

By that logic, never mind about those Nazis. Much ado about nothing. After all, ladders, bathtubs, toddlers, and husbands were all more deadly “in America” from 1939–1945 than the SS or the Wermacht. Millions of died overseas, including hundreds of thousands of Americans, but the Germans couldn’t strike us here at home. So all that hysteria over Hitler? Fearmongering, really. He couldn’t hurt us.

The reason Kristof distinguishes between murders committed by terrorists on the homeland and those killed overseas is because Trump’s executive order is entirely concerned with stopping terrorists from entering the United States and committing attacks here.

The title of Trump’s executive order is: “EXECUTIVE ORDER: PROTECTING THE NATION FROM FOREIGN TERRORIST ENTRY INTO THE UNITED STATES.”

The first header of the executive order is “Protecting the Nation from Foreign Terrorist Entry into the United States.”

The first full sentence is:

By the authority vested in me as President by the Constitution and laws of the United States of America, including the Immigration and Nationality Act (INA), 8 U.S.C. 1101 et seq., and section 301 of title 3, United States Code, and to protect the American people from terrorist attacks by foreign nationals admitted to the United States, it is hereby ordered as follows [emphasis added].

President Trump’s executive order limited the scope of the debate to foreign-born terrorists on U.S. soil, so it is perfectly reasonable for Kristof to stick to this issue when criticizing Trump’s executive order. There may be better justifications for the executive order than those offered by the administration itself but that doesn’t change the fact that the official justification written into the text of the order itself is to prevent terrorist attacks committed by foreign nationals on U.S. soil.

French also unintentionally makes a compelling argument that the government’s overreaction to terrorism killed far more people than it saved. He writes that we should count the 7,000 U.S. soldiers who have died abroad fighting jihadists as victims of terrorism. Assuming his numbers are correct, that means 2.35 U.S. soldiers died abroad after 9/11 in the wars started as a result of 9/11 (even though Iraq wasn’t initially about terrorism) for each innocent victim killed during that terrorist attack. That seems like a bad exchange.

Those 7,000 U.S. military deaths overseas would have to have prevented at least an equivalent number of American deaths from terrorist attacks to justify the loss of life even though the Iraq War was not initially justified with stopping terrorism and had nothing to do with 9/11. To prevent 7,000 terrorism deaths since 9/11 would mean stopping about 30 San Bernardino-scale attacks each year. To put that in perspective, only 80 people have been killed on U.S. soil in any terrorist attack from 2002 to the end of 2015. It is extremely unlikely that our government’s decision to fight several wars overseas in response to 9/11, even though the Iraq War was not a direct response to that attack, prevented 7,000 deaths from terrorism. This is just a back of the envelope calculation that excludes non-American civilian deaths in war zones, the costs of property damage not incurred in terrorist attacks that didn’t happen, the cost to U.S. taxpayers of fighting wars overseas, and the potential for blowback. Including all of those costs and savings would yield a far more depressing number even if the deaths of foreign civilians was not considered.

French is correct to criticize it when Kristof or others compare dying from a terrorist attack to dying from accidents like slipping in a bathtub. However, it is unfair for him to criticize Kristof for arguing over Trump’s executive order on rhetorical ground chosen by the Trump administration.