There are some questions in immigration that remain almost entirely unaddressed. Academic experts on the left tend to shy away from them, and policymakers on the right are not interested in the nuance. What the public thinks about how we enforce immigration law and how to make it better is one of them. So I’m excited to welcome a guest post from Mollie Gerver, an associate professor in international ethics at King’s College London. She is one of the most nuanced thinkers I know, a philosopher who combines ethical reasoning with empirical data on the issue. I hope you enjoy it.
Nearly everyone agrees that immigration enforcement is justified at least some of the time. However, immigration enforcement can also be done badly, entailing a great deal of violence and disproportionate harm. Harm can be deadly, as when Indian border officials shoot at Bangladeshi migrants attempting to enter India. Or it can be traumatising, as when UK officials wake up children at dawn, force them into vans, and bring them to airports for deportation. Sometimes harm is against those with a legal right to remain in a country, as when American immigration agents grabbed a seven-year-old and her mother, locked both in detention, and asked them to self-deport despite reportedly having visas to remain in the US.
If immigration enforcement is harmful, there is a question of when and whether it should continue. Many political philosophers claim nearly all current enforcement should discontinue, and while others disagree, all accept there is a debate to be had. This debate is less common in the field of public opinion. Public opinion research focuses less on citizens’ views on enforcement than on which immigrants they support admitting. This is a problem: understanding views on enforcement matters for democratic politics, because governments currently enforce immigration law in citizens’ name without knowing which methods those citizens actually support. Importantly, understanding views on enforcement could show governments that they have more room to enforce immigration law humanely than they assume.
Opinion research on enforcement is still limited
Public opinion researchers have paid much less attention to views on immigration enforcement as compared to views on how popular immigration is in general and which immigrants are popular in particular. Researchers often examine this using conjoint experiments. Conjoint experiments involve presenting respondents with two migrants, randomly varying the migrants’ gender, religion, occupation, criminal record, and so forth, and then asking respondents which migrants they prefer admitting or letting remain. If respondents are more likely on average to prefer migrants with certain characteristics, such as those not holding a criminal record, this is evidence that respondents’ views are influenced by this characteristic. Researchers also utilise factorial vignette experiments, where respondents are asked how much they support admitting a given migrant, but not asked to compare pairs of migrants. These experiments, too, randomly vary characteristics to understand which characteristics influence support or opposition to admitting a given migrant.
These and other surveys tell us that most citizens in most countries oppose increasing immigration in general, oppose migrants arriving with a criminal record in particular, and prefer migrants who are high-skilled over those arriving for humanitarian reasons or to be reunited with family (see here, here and here). European citizens also likely prefer Christian and non-religious migrants over Muslims, and citizens throughout the world prefer migrants who share the majority religion of their country.
All these findings, and many more, have one thing in common: they focus on which migrants citizens want, and not how migrants should be treated. They do not ask if citizens support deporting, detaining, shooting at, and blocking migrants from remaining or entering.
The importance of understanding views on enforcement is made salient if we compare international migration with internal migration. With internal migration, nobody claims citizens’ views on the migrants they support always track their views on how migrants should be treated. For example, when San Francisco natives complain that too many Ohioans are arriving, calling them “transplants”, nobody suggests these San Franciscans want the Ohioans forced onto the next flight out of town. For all we know, citizens think similarly about international migration: they might prefer no increase in migration or only some migrants, but still oppose some enforcement necessary to prevent these migrants from arriving. Current public opinion research often overlooks this possibility.
Because it does, researchers sometimes assume most citizens are not very humanitarian. For example, Alexander Kustov notes that citizens generally do not support admitting more migrants for humanitarian reasons, preferring migrants who will help the economy, such as by filling labour shortages. He concludes that citizens are driven less by humanitarian concern in general, and more by concern for the welfare of their co-nationals in particular. While this might be true, citizens could be more humanitarian than his evidence suggests: citizens might say they prefer fewer migrants, or only migrants who will help the economy, just as many San Franciscans prefer fewer Ohioans, or only Ohioans who will help the economy, but they might still oppose treating them in ways deemed sufficiently harmful.
This is partly because citizens might care about fairness. We know citizens think some immigration policies are unfair in general, and they might similarly view immigration enforcement as unfair when harms against migrants far outweigh the benefits for citizens. Alternatively, perhaps citizens don’t think in terms of fairness, but just feel - in the pit of their stomach - that some treatment is always wrong, including imprisoning children for lengthy amounts of time, shooting in the directions of migrants attempting to enter, and intercepting patients trying to reach hospitals. Finding out citizens’ views on these and other enforcement mechanisms is essential.
Because it is, some scholars have indeed begun examining this topic, but tend to only ask about a limited range of enforcement, leaving details out. For example, research by Vrânceanu et al finds that many Europeans supporting the return of migrants to Turkey still oppose pushing the boats of migrants back to Turkey, but these and other studies in the US and Africa do not ask individuals if they would oppose enforcement procedures even if necessary to achieve the goal of returning migrants. After all, people opposing pushbacks at sea and other procedures might think migrants can be returned on flights, but would support pushbacks and other procedures if there were no other way of preventing migrants from arriving.
Nor do surveys typically examine whether respondents shift their views if told about harms. While one well-designed survey examines levels of support for deporting gay migrants as compared to others, with the authors noting that gay migrants might face greater harm, this survey does not examine if telling respondents about these harms would shift views. Similarly, another study examines whether support for deportation is impacted by the legal, cultural and economic characteristics of migrants, but not the harms from deportation. An exception is a ground-breaking study demonstrating that respondents are less supportive of deportation if told about the harms migrants will face in their home countries, but even this excellent study does not examine views on harms from enforcement itself, including malnutrition in detention, injury trying to cross a border, and post-traumatic stress disorder from being handcuffed to a seat on a deportation flight.
Outside of academia, politicians and journalists sometimes claim that violent enforcement can contribute to election losses, as when one news host insisted that ICE violence led to many US citizens voting Democrat in the 2026 midterm elections. However, it is difficult to know if this is true, given that many previous election years did not include polls asking about a range of enforcement. Even in 2026, a year where quite a few polls asked about enforcement, details are limited: an NBC exit poll found that 57% of US voters thought immigration enforcement had gone too far, and a Pew poll found that most citizens oppose ICE agents arresting individuals at schools and hospitals, but neither of these polls asked about a broader range of enforcement procedures. In fact, the European Social Survey even discontinued a question asking respondents’ views on deportation because this was not “sufficiently fundamental theoretically or important in policy terms.”
Why this is a problem
Not understanding citizens’ views on enforcement is a problem partly because of ethics. The ethics of immigration is not only about who should be admitted or allowed to remain, but how migrants should be treated in practice. That is why moral philosophers generally hold that detaining migrants for a very long time is a disproportionate response for limiting migration, even if limiting migration is justified. We do not know if citizens agree with this assessment, and it matters if they do.
This matters partly because, even if citizens don’t agree with the moral philosophers, they might at least perceive moral trade-offs. For example, if many citizens oppose increasing the rate of children and elderly migrants arriving, they might also oppose detaining or forcibly deporting children and elderly migrants, facing a moral trade-off: they either progress the value of reaching their desired immigration goals, or they progress the value of not harming vulnerable migrants in their country. Citizens might be uncertain which value they should prioritise, feeling torn. This feeling is important if we want to understand not only what citizens think, but what citizens are uncertain about thinking.
Understanding citizens’ views might also tell us what is feasible for policymakers to do. For example, even if prolonged detention is often wrong, if very popular then policymakers trying to end this practice might be voted out of office before they can. In contrast, if the public opposes this practice, policymakers cannot claim they are unable to end this practice due to public attitudes. The question, then, is what the public’s attitudes actually are; this knowledge is largely missing within the field of public opinion.
Lack of this knowledge is even a problem for democracy. If immigration restrictions are only democratically legitimate if popular, it’s tempting to claim restrictions are democratically legitimate because most people do not want more migrants, but this ignores enforcement itself. If most citizens do not support detention, deportation, pushbacks, and other enforcement policies, these policies cannot be justified by appealing to their popularity. There is therefore value in understanding whether these policies are indeed popular.
Finding out what is popular
We can evaluate when enforcement is popular with better conjoint and vignette experiments: when asking respondents to decide which of two types of migrants they prefer, or whether they support a given migrant entering or remaining in a country, an experiment can include not only the migrant’s gender, occupation, criminal history, and so forth, but the harm from enforcement necessary to prevent them from entering or remaining. Information should include whether the migrants are detained, forced onto a flight, arrested at work, blocked from crossing a border via a wall, and so forth. Information should also include what will likely happen to migrants as a result. This approach has already been integrated into three experiments I conducted with colleagues in the UK and US, and we found that the average opinion is against a range of harmful enforcement practices even when necessary to prevent migrants from entering or remaining.
Importantly, we found that even those opposing more immigration don’t exactly support life-threatening or injurious enforcement: anti-immigrant respondents are pretty ambivalent about a lot of deportations, detention procedures, and so forth, especially against refugees. This was true even though we asked respondents about enforcement procedures that were necessary to prevent individuals or groups of migrants from entering, still finding that support for a range of enforcement was lukewarm or non-existent. In general, views on immigration do not perfectly track views on enforcement.
Future surveys can go a step further in examining not only people’s views, but how these views are impacted by real events. We know that real events impact citizens’ attitudes about immigration in general, and they might impact attitudes about enforcement in particular. For example, in 2015 two-year-old Syrian refugee Alan Kurdi was found dead after drowning in the Aegean Sea, a death that could have been prevented had he and his family been permitted to board a flight. Kurdi’s death was followed by widespread protests in Europe against barriers refugees faced trying to reach safety. It may be that opposition to certain enforcement mechanisms shifted, yet public opinion polls at the time did not evaluate this possibility. Future public opinion polls should do so, asking individuals when and whether they support a range of enforcement mechanisms.
Such polls may also help us evaluate if enforcement impacts people’s actual votes. As noted, some claim ICE violence – like against the seven-year-old arrested with her mother – led to US voters voting Democrat in 2026, but we don’t know if this is true without understanding people’s views on ICE violence. Understanding people’s views requires not only understanding if people support ICE in general, but their views on what ICE does in particular.
Such surveys should also ask respondents about enforcement not reaching the news; surveys should not, for example, only ask about ICE when ICE is becoming more violent than it has ever been. Surveys should instead focus, or also focus, on seemingly run-of-the-mill detention and deportation that has been occurring for over a hundred years.
Surveys should also examine views on enforcement procedures that don’t seem harmful, but are, such as carrier sanctions. Carrier sanctions involve a government fining an airline if it lets an individual board a flight without a visa, which is why Alan Kurdi and his family couldn’t board a flight in Turkey; no airline would let them board due to threats of fines by European governments. Carrier sanctions may be unpopular. In the studies I conducted with colleagues on enforcement we found that respondents are far less likely to support carrier sanctions if utilised against migrants whose lives are at risk in their home countries. Future studies should include more detailed questions about these sanctions, explaining what airlines do to physically block people from boarding a flight with their purchased ticket, and what happens to migrants as a result.
***
Understanding these and other views on enforcement is essential not only as an academic exercise: it can potentially create enforcement that is more humane. Humane enforcement is rare partly because governments decide how harsh to be by guessing voters’ views, or relying on general polls about how many immigrants citizens want. This is a mistake, because if citizens wanting fewer immigrants oppose arresting children at schools, separating families in detention, and threatening migrants with bullets, then governments might have more room than they assume.
They might be able to enforce immigration law while dropping practices that cause the most harm and win the least support. In contrast, if citizens really do support harsh methods even when understanding what they entail, those of us objecting to these methods will need to make the moral case for their end, rather than assuming the public is on our side. Either way, advocating for less harmful enforcement requires understanding citizens’ opinions on this topic, and how these opinions differ from their general skepticism about immigration.





There are a number of "soft" approaches here like e-verify, removal of welfare benefits, or deporting immigrants who get arrested for other reasons
My 5 cent to this topic:
Implementation and enforcement of any given policy is inseperatable from the policy itself. If we don't want to hurt humans in the process of trying to control their lifes, then maybe the policy itself is faulty as well. This is also why you cannot just ask what people prefer. It's the migrant itself that chooses his own behavior, not the policy makers. That is, unless you implement control mechanisms with their respective cost, be it monetary or human.