Lack Of Diversity In Genetic Databases Hampers Research

Similar issues in Canada:

When Lalita Manrai went to see her doctor for treatment of kidney disease, she noticed that some of the blood test results had different “normal” ranges for African Americans compared with everybody else.

When she asked her doctor which range applied to her — a woman born in India — he said the “everybody else” category was actually based on a study of Europeans, so neither category was right.

Instead, he said, he calculated “normal” for her by averaging the two values.

“It’s ridiculous,” says Arjun Manrai, a medical researcher at Harvard Medical School, who recounted this story of his mother, who died in 2018. But there simply isn’t good information about a lot of medical issues that may vary based on a person’s ancestry. “In this vacuum of information, this was what [the doctor] was doing as his approach to staging her kidney disease,” Manrai says.

It’s important to get those laboratory results right, because they influence a patient’s treatment, Manrai says.

The same problem comes up in other common situations, such as the A1C test that is used to diagnose and manage diabetes, and in genetic variants that can identify people at risk of sudden death from heart disease.

These factual gaps exist because much of the research used to understand these genetic tests and lab values comes from predominantly European populations. Manrai is part of a growing effort to correct the skewed picture that results.

One of the most widely used resources for studying the genetics of disease is the U.K. Biobank, which contains samples from half a million middle-aged British people, 95% of whom are of European ancestry.

“At the time they were recruited and the age group that were recruited, that largely reflected the average across the U.K.,” says Dr. Cathie Sudlow, the biobank’s chief scientist. “So because the study was in the U.K., that’s what we got.”

The biobank has been a boon to scientists who want to identify the genes that are involved in disease. Genes are universal. But the ethnically skewed resource doesn’t work as well to identify the genetic variants that differ based on ancestry.

“There is no one cohort anywhere in the world that can answer all questions for all people,” Sudlow says. So the biobank is working to help develop much more diverse resources.

The U.K. Biobank has helped establish large repositories in Mexico and China. In the United States, Sudlow and her colleagues have been offering advice to the National Institutes of Health, which is gradually putting together a biobank that aims to have a diverse population of a million volunteers.

There are dozens and dozens of collections like this scattered around the world, some in private hands and others accessible to scientists. Nobody knows exactly how many of these collections exist, but “broadly we’re talking about at least millions of people,” says Ewan Birney, co-director of the European Bioinformatics Institute.

He is part of an effort to find ways to link some of these resources together so scientists can quickly see how a discovery in one group applies to people with different ancestries. Birney says even though most of the initial work has been in European populations, a lot of it is relevant to everybody.

“How genetics works in different countries — sort of a surprise — is that very often the genetics is pretty much the same as you move between different countries,” Birney says.

Where biobank study conclusions can be misleading is in the details. The same genes and proteins are involved in diseases such as diabetes, but the variants that can affect a person’s risk of disease differ based on a person’s genetic heritage.

Birney expects that the new and linked databases not only will help identify issues of concern to a particular ethnic group but will identify genes that are important for everybody’s health. He’s particularly eager to learn what comes out of a biobank project taking shape in sub-Saharan Africa.

“Because Africa is the birthplace of humans, there’s the highest amount of genetic diversity inside of sub-Saharan Africa,” he says. “And it’s really clear if you are a geneticist, we should be spending an awful lot more time studying humans there.”

Birney is mindful of simply allowing scientists from rich companies to swoop in on this resource, so right now the African scientists developing biobanks will have an opportunity to study the data first. Birney says it’s “really important that we do that in a way that is empowering and enabling for the scientists who come from these different countries.”

Manrai at Harvard is tapping into data that’s already available, including medical databases curated by the National Institutes of Health and the Centers for Disease Control and Prevention.

“I think understanding ancestry, race, ethnicity is an area that we’re going to see a tremendous amount of work in over the next 10 years,” he says.

Source: Lack Of Diversity In Genetic Databases Hampers Research

Cato’s 2018 Immigration Research in Review | Cato @ Liberty

While I am far from being a libertarian, I do find that Cato’s analysis of immigration issues, and particularly of immigration-related data, impressive and worth following. Their round-up provides a good sense of the scope of their research and analysis, particularly with respect to some of the myths circulating or being propagated by the Trump administration:

Cato’s immigration policy team was very busy in 2018.  My colleagues David Bier and Andrew Forrester, in addition to some contributions by myself and numerous outside authors like the stupendous Michelangelo Landgrave, worked non-stop to produce almost 180 pieces this year in the form of blog posts, op-eds, Cato research papers, and peer-reviewed academic articles.  David Bier summarized many of these pieces in a twitter thread for those on Twitter.

Of those, I’m most proud of the pieces that discovered original facts and figures to illuminate the immigration issue.  With rare exceptions, the most valuable immigration policy research is that which produces original facts and figures, as too much of the debate over this topic is emotional and ungrounded.  We are trying to make the debate about the facts and contributing those that we have discovered on our own in the process. Below is a rundown of the original facts and figures that Cato scholars have calculated in 2018 by subtopic with links to our research.


The recent surge in immigrants along the border are low-skilled, poorly educated, and from Central America – but that doesn’t stop them and their descendants from learning English, converging to American wages, and joining the military at rates comparable to or higher than native-born Americans.

Border Security, the Wall, and Interior Immigration Enforcement

Much of the national immigration debate proceeds under the implicit and incorrect assumption that immigration enforcement only harms illegal immigrants. My colleague Matthew Feeney waded into the immigration debate with an excellent primer on how increased immigration enforcement, both at the border and in the interior of the United States, will infringe upon the civil liberties of American citizens and lawful permanent residents as well as an examination of legal protections that can help mitigate the lost rights.  Complementing Feeney’s paper is our finding, based on data from Travis County in Texas, that Immigration and Customs Enforcement (ICE) targeted at least 228 American citizens as illegal immigrants in that county over 12 years – or about 0.9 percent of all those detained.

Related to interior immigration reform is the E-Verify program, which is an electronic eligibility for employment verification system run by the federal government.  Congress created it in an attempt to turn off the magnet that attracts illegal immigrants to the United States in the first place: higher wages and low unemployment.  In theory, E-Verify would allow employers to check the identity information of new hires against government databases to see if they are legally eligible to work and to deny illegal immigrants.  For years, members of Congress have introduced bills to make E-Verify a national mandate to be used whenever a business hires somebody – including American citizens.

Four states have mandated E-Verify for all new hires, but only 56 percent of new hires in those states were run through E-Verify in the second quarter of 2017.  To be effective, a much higher percentage of new hires must be checked through E-Verify.  The four states that mandated E-Verify are Arizona, Alabama, Mississippi, and South Carolina.  Over time, the rate of new hires has barely budged in those states – even in South Carolina where the state conducts random audits of employers to supposedly guarantee compliance.  If those conservative states can’t effectively enforce an E-Verify mandate, there is no hope for doing so nationally.

Our next piece of original research confirmed that California’s TRUST Act, which limited state law enforcement cooperation with ICE, dramatically reduced deportations from that state.  Although deportations from California were falling prior to the TRUST Act going into effect in 2014, deportations from California that year dropped 39 percent relative to 2013.  In the rest of the country, the number of deportations only dropped 9 percent over the same period.

Much of the rest of our original research focused on border enforcement.  Republicans introduced a bill in 2018 to spend more on Border Patrol in the next five years than has been spent over the last 5 decades – in real terms.  A portion of that extra money would be spent on drones to patrol the border, an enforcement tool that has already been used on the border and is responsible for 0.5 percent of all border apprehensions at an astonishing cost of $32,000 per arrest.  Apprehended border crossers, whether discovered by drones or more traditional methods, spent an average of 39 hours in detention in late 2014 and 2015 or 12.8 million hours total.  Of course, all of this extra enforcement is unnecessary as the lesson of marijuana legalization on the state level shows that smuggling can more effectively be cut with better laws that allow cross-border flows rather than crackdowns.

Part of the justification for more spending and technology on the border is that Border Patrol agents face severe threats on the job.  While they certainly do, it’s not nearly as dangerous as many assume.  Thirty-three Border Patrol agents died on the job from 2003 through 2017 or about one death for every 7,968 agents per year.  Six of those agents were murdered on the job while the other 27 died in accidents or in unknown circumstances.  Their on-the-job murder rate is about 1 in 43,824 per year from 2003 onwards, much lower than the 1 in 19,431 annual murder rate for Americans during the same time period.  Every one of those murders or deaths is a tragedy, but those rates do not indicate an exceedingly dangerous job.


In 2016, illegal immigrants were 47 percent less likely to be incarcerated than native-born Americans and legal immigrants were 78 percent less likely to be incarcerated than natives.  By race and ethnicity, legal immigrants and illegal immigrants were less likely to be incarcerated than their native-born co-ethnics.  In the state of Texas, which actually counts criminal convictions by immigration status, the illegal immigrant criminal conviction rate is about half that of native-born Texans and the legal immigrant conviction rate was 66 percent below.  In Texas, that pattern also holds for crimes like homicide, larceny, and sex crimes.  Nationwide, only about 11 percent of “criminal aliens” actually committed a violent or property crime and 60 percent of those “criminal aliens” deported committed only a victimless crime. Related to these findings, DACA recipients were far less likely to be arrested than those who were not in DACA.

Illegal immigrants could commit more crimes and escape punishment for them by fleeing back to their home countries, but police clearance rates (the rate as which police solve crimes) are not correlated with the size of the illegal immigrant population even with numerous controls.  There is even some evidence that motor vehicle theft and burglary are solved as slightly higher rates in states with more illegal immigrants as a proportion of their population.  This is consistent with our finding that the interior immigration enforcement program had no effect on crime rates in North Carolina although it did increase assaults against police officers.  Interestingly, Arizona’s passage of an E-Verify mandate in 2007 drastically increased the flow of non-citizen offenders into Arizona state prisons – a serious potential side-effect of increased immigration enforcement that E-Verify supporters have yet to address.

Crime in Mexico along the U.S. border is a serious problem, but we found a negative correlation between homicide rates in Northern Mexico border states and homicide rates in American border states.  Expanding on the theme of crime flowing over the border, only about 0.2 percent of all border apprehensions in the first half of 2018 belonged to a gang.

DACA and Legalizing Unlawful Immigrants

President Trump’s slow-motion cancellation of the DACA program made for DACA-recipients and other Dreamers a big political issue in 2018 and several bills to do so in exchange for a border wall were proposed.  Many of those bills would have legalized only a small proportion of the Dreamer population, about half the number that President Trump claimed.  Another proposal would have denied a path to citizenship for 82 percent of Dreamers.

Economic Growth, Fiscal Effects, and Wages

Former visiting fellow Ike Brannon estimated that reversing DACA would cost the U.S. economy $351 billion from 2019 to 2028 in lost income and that the U.S. Treasury would lose $92.9 billion in tax revenue.  Under Trump’s proposal to halve legal immigration, we used a simple model to show that it would reduce the size of the U.S. economy by about $19 trillion in 2060 relative to what it would have been under the status quo, mainly by reducing the growth of the American population by 26 million.

Wage and economic assimilation of new immigrants is vitally important. Newly arrived immigrants have wages lower than otherwise identical natives, but those wage differences diminish greatly or disappear entirely after about two decades of working in the United States.  The immigrant wage gap has diminished in recent years.  Furthermore, illegal immigrants initially faced a hefty wage penalty of about 11.3 percent relative to legal immigrants due to their lack of legal work status.


Many commentators expressed fear that immigrants, especially those in the migrant caravans, would spread disease once they arrive.  However, vaccination rates in Mexico  and Central America are generally higher than or about the same as those in the United States.

Immigration Affects the Fundamentals of Economic Growth

The best criticism of expanded legal immigration is that the new Americans and their descendants could vote for bad policies that diminish the prosperity of the United States.  On its face, this is plausible as immigrants generally come from countries with worse economic and political institutions than the United States.  Immigrants today are coming from more democratic countries than immigrants who came in the past.  Additionally, we published a working paper that examined a quasi-natural experiment in Jordan where a large and sudden exogenous shock of migrants permanently moved there.  We found that the migration significantly increased economic freedom.  That paper was accepted for publication in the World Bank Economic Review, the 28th best peer-reviewed academic economic journal in the world.  More impressively, that publication marks the first peer-reviewed publication for my talented colleague Andrew Forrester.

Unrelated to immigrant effects on public policy, we investigated whether immigrants could worsen U.S. economic growth by reducing the quality of firm management and found precisely nothing.

Immigration Policies in Foreign Countries

No analysis of American immigration policies is complete without a comparison to policies in other countries.  The United States ranks in the bottom third of wealthy countries in terms of net new immigrants as a share of total population from 2015 to 2017 as well as total foreign-born residents as a share of total population.  Singapore’s relatively open immigration policy provides a possible model for the United States.  About 47 percent of Singapore’s population is foreign-born, more than three-times greater than the United States as a whole and larger than any American urban area, but with fantastic economic effects compared to its neighbors.

Legal Immigration

One of President Trump’s immigration reform frameworks would have cut 22 million legal immigrants over the next 50 years and, if it was in place since 1965, it would have reduced legal immigration by about 23 million.  That latter figure doesn’t include the tens of millions of our fellow citizens born here since 1965 who would not be Americans if that framework was applied retroactively.  Consistent with the President’s plans to cut legal immigration, his administration has increased the denial rate for visas by 37 percent.

President Trump and those who want to cut legal immigrants have frequently said that they want to reduce low-skilled immigration and boost the number of highly-skilled immigrants so that our immigration system looks more like the Canadian system.  This is unnecessary as our immigration system, on its own, is already admitting far more skilled immigrants than it used to.  On paper, the proportion of skilled new immigrants admitted to the United States from 2012-2016 is about the same as in Canada during that time:  49 percent with a bachelors or above education admitted to the United States compared to 52 percent in Canada.  Even immigrants who arrive via family-reunification and on the diversity visa are more educated than native-born Americans.

Although our legal immigration system is admitting more skilled immigrants on its own, serious problems remain.  For instance, Indian immigrants with advanced degrees face a 150-year wait for employment-based green cards.  That is shockingly unfair and economically destructive, even for a government bureaucracy.  Small tweaks to our immigration system could reduce that problem significantly.  More importantly, a small administrative change that is consistent with current law could increase legal immigration by 27 percent across the board and allow in far more skilled immigrants.

Refugees and Asylum Seekers

President Trump’s so-called Muslim ban has cut Muslim refugees, immigrants, and travelers by 91 percent, 26 percent, and 60 percent, respectively.  Related to that, Trump’s refugee policy has also cut the number of Christian refugees by 64 percent.  Additionally, signing a Free Trade Agreement with the United States does not boost the number of refugees or asylum-seekers who come from those countries.  The Syrian Civil War is winding down, but a persistent criticism over recent years is that rich Gulf States have not sponsored any Syrian refugees.  While legally true, that analysis ignores the fact that the Gulf States have allowed over 1.2 million Syrians to enter and remain on their territory on non-refugee visas over that time in response to the humanitarian crisis.


President Trump favored “extreme vetting” for new immigrants and travelers to prevent future terrorist attacks. But since the 9/11 attacks, the U.S government has done an admirable job screening out terrorists.  From 2002-2016, the government issued one visa to a radicalized terrorist for every 29 million non-terrorists and issued 379 million visas for each deadly terrorist.  The government undertakes many more counterterrorism activities than just visa vetting.  Since 9/11, they have spent $2.8 trillion on counterterrorism.  Assuming the statistical value of life is $15 million, that spending would have to have prevented about 188,740 murders in terrorist attacks during that time to break even – or over 1,000 times as many people as were actually murdered in terror attacks on U.S. soil since 9/11.  That is extremely unlikely.

About 3,518 Americans have been murdered in terrorist attacks on U.S. soil from 1975 through the end of 2017.  That’s about a one in 3.3 million chance per year of being murdered in a terrorist attack here committed by any terrorist.  By comparison, 7,548 people have been murdered by animals during that time – a death rate about double that caused by terrorists.  The annual chance of dying in a terrorist attack in the United Kingdom during that time is higher at about 1 in 1.1 million per year.  Since 9/11, the chance of being murdered in a terror attack in France has been about 7-times higher than in the United States.  Terrorism is obviously a threat to Americans that the government should seek to keep low, but its deadliness should not be exaggerated.

The migrant caravan dominated headlines in 2018, but the terrorist threat from asylum-seekers and illegal immigrants has been very low since 1975 and not a single terrorist from Mexico or Central America has entered during that time.  The last year that illegal border crossers who were eventually convicted of planning a terrorist attack on U.S. soil entered the United States was in 1984.  They came as children and were arrested in 2007 before they killed or injured anybody.  Furthermore, those apprehended along the border from Muslim countries haven’t committed any attacks on U.S. soil and none of the examples given meet that criteria.


On the basis of monetary value, immigrants individually consume about 39 percent fewer welfare benefits than native-born AmericansImmigrants and their native-born children consume about 33 percent less welfare individually than native-born Americans whose ancestors have been here for at least two generations.


Immigration has been one of the top policy issues since 2015.  Cato scholars have been at the forefront of publishing new facts and figures to illuminate this debate.  This post does not include our other activities such as our work with Rep. Grothman (R-WI) to reduce immigrant welfare consumption, our numerous public debates, summations of outside research, and weekly analysis of immigration-related events.  We hope to continue this pace of original research in 2019 and beyond.

Source: Cato’s 2018 Immigration Research in Review | Cato @ Liberty

ICYMI: Stephen Harper pledges $10M to research terrorism, radicalization

Good investment and one that any government should maintain and possibly strengthen:

On the anniversary of the Sept. 11 attacks on the U.S., Conservative Leader Stephen Harper is committing new money to research terrorism and radicalization.

Harper said Friday that a Tory government would provide $10 million over five years to the Kanishka Project, an initiative — established in 2011 and named in recognition of the 1985 Air India bombing that killed 331 people — to better understand radicalization and effective ways to prevent attacks.

The Kanishka Project is administered through Public Safety Canada and has funded research by academics both in Canada and abroad. For example, in October 2014 the government put out a call soliciting research on how jihadists use the internet, while in July it was announced that the project would provide $170,000 over two years to an Australian sociologist studying why some Canadians convert to Islam.

Source: Stephen Harper pledges $10M to research terrorism, radicalization – Politics – CBC News

Research based on social media data can contain hidden biases that ‘misrepresent real world,’ critics say

Good article on some of the limits in using social media for research, as compared to IRL (In Real Life):

One is ensuring a representative sample, a problem that is sometimes, but not always, solved by ever greater numbers. Another is that few studies try to “disentangle the human from the platform,” to distinguish the user’s motives from what the media are enabling and encouraging him to do.

Another is that data can be distorted by processes not designed primarily for research. Google, for example, stores only the search terms used after auto-completion, not the text the user actually typed. Another is simply that many social media are largely populated by non-human robots, which mimic the behaviour of real people.

Even the cultural preference in academia for “positive results” can conceal the prevalence of null findings, the authors write.

“The biases and issues highlighted above will not affect all research in the same way,” the authors write. “[But] they share in common the need for increased awareness of what is actually being analyzed when working with social media data.”

Research based on social media data can contain hidden biases that ‘misrepresent real world,’ critics say

Justice Canada chops research budget by $1.2-million – The Globe and Mail

Worrisome. Not the decision itself to cut research funding as much as the reason: not liking the results, and wanting to align research to government policy, rather than understanding of society. While public service research sometimes was less neutral and impartial than it should have been (had my experience in multiculturalism research in this regard), this change abandons any claim to independent and objective research.

Abdication of “fearless advice” role of public service. Hopefully this will provoke more serious reflection among senior ranks of the public service rather than the rather shallow Destination 2020 initiative:

The result is a diminished research capacity, which now must be better controlled from the top to ensure it supports the government policies, says the report.

“The review confirmed that there have been examples of work that was not aligned with government or departmental priorities,” says the October 2013 document, obtained by The Canadian Press under the Access to Information Act.

Some past projects have “at times left the impression that research is undermining government decisions.”

The report did not cite specific studies, but a department report last year on public confidence in the justice system appeared to be at odds with the Conservative government’s agenda.

Researcher Charlotte Fraser found many Canadians lacked confidence in the courts and prison system, but suggested it was the result of misunderstanding rather than any failures in the system, and that education could rectify the problem.

Justice Canada chops research budget by $1.2-million – The Globe and Mail.

Le budget de la recherche en droit fond de 1,2 million (La Presse)

Book Review: The Myth of Research-Based Policy and Practice | LSE Review of Books

An interesting mini-review of an interesting book, The Myth of Research-Based Policy and Practice, by Martyn Hammersly, questioning the limits of evidence-based policy. I like the conclusion of the review, as in the end, still better than the alternatives:

In the end, I find myself describing evidence-based policy as Churchill described democracy – the worst option excepting all others. Although this book dispelled some of the mythology, when it comes to evidence-based policy, to borrow a phrase from The X-Files, “I want to believe.”

Book Review: The Myth of Research-Based Policy and Practice | LSE Review of Books.

What We Don’t Know Can’t Hurt Us (Right?) | The Census Project Blog

A bit tongue-in-cheek on the US Census debates. Orwellian “ignorance is strength”.

What We Don’t Know Can’t Hurt Us (Right?) | The Census Project Blog.

Column: What’s the evidence for evidence-based policy?

William Watson raises some valid and important points about evidence-based policy and the limits. While some data and evidence is largely neutral and firm (e.g., Census data) other evidence can be subject to confirmation and other biases, in addition to the limits of our understanding of the complexity of society and behaviour. Evidence is still better than anecdote, but it limits also need to be understood. #W2P #GOC

Column: What’s the evidence for evidence-based policy?.