Thursday, November 20, 2014

Guns, Crime, and Statistics: It's Complicated

Despite being deeply contested in academic circles, a paper published in 1997 by John Lott and David Mustard, "Crime, Deterrence, and Right-to-Carry Concealed Handguns" (and the book, More Guns, Less Crime, that followed) has had a profound impact on public policy.  Lott and Mustard's argument that the adoption of "shall issue" or "right to carry" (RTC) handgun laws has been instrumental in reducing violent crime has been used by legislators and special interest groups to push for more states to pass laws to make it easier to get permits to carry concealed loaded guns, and to lessen the barriers for those permit holders to take guns in more places (schools, bars, places of worship, ...).

A new analysis of the impact of right to carry (RTC) laws by Stanford University's Abhay Aneja and John Donohue and Johns Hopkins University's Alexandria Zhang debunks the "more guns, less crime" thesis.  More importantly perhaps, the authors also sound a cautionary note about interpreting the findings of any one study--including their own.

Aneja, Donohue, and Zhang uncover four types of issues with Lott and Mustard's original study:

  1. "The comparison of crime between RTC and non-RTC states is inherently misleading because of factors such as deprivation, drugs, and gang activity, which vary significantly across gun-friendly and non-gun-friendly states (and are often difficult to quantify). To the extent that the relatively better crime performance seen in shall-issue states during the late 1980s and early 1990s was the product of these other factors, researchers may be obtaining biased estimates."
  2. Lott and Mustard's study covered the period 1977-1992.  "Crime rates declined sharply across the board beginning in 1992. [...] Moreover, the average crime rates in non-RTC states seemed to have dropped even more drastically than those in RTC states, which suggests that crime-reducing factors other than RTC laws were at work."
  3. The data used by Lott and Mustard are problematic: Aneja, Donohue, and Zhang found multiple coding errors.  In addition, county data are notoriously inaccurate, inconsistent, and incomplete.  Finally, Lott and Mustard used a flawed contemporaneous arrest rate variable.
  4. The statistical model specification used by Lott and Mustard is problematic in a number of dimensions--omitted variable bias, endogeneity, collinearity, and treatment of standard errors.

The authors correct for some of these problems (fixing data inaccuracies, poorly constructed arrest ratios, incorrect standard errors, and using an extended 1977-2000 dataset).  They find that "RTC laws increase crime--for rape, aggravated assault, robbery, auto theft, burglary, and larceny.  There is not even a hint of any crime decline."

In an effort to probe the robustness of the results and improve the model, Aneja, Donohue, and Zhang then proceed to revise the model (using state rather than county data; extending the study to 2006, then 2010; removing collinear variables; including better controls; ...) and run a rigorous and extensive set of analyses with varying model specifications.

Their results make three important points:

  • No matter the model specification, they can find no evidence to support the "more guns, less crime" claim.  On the contrary, the evidence suggests that the effects of RTC laws on crime are positive, meaning that adopting RTC laws appears to result in crime increases.  This effect is strongest and most consistent for aggravated assault.
  • "Different models yield different estimated effects": while no model shows that RTC laws decrease crime, the impact of RTC laws varies from model to model.  In some models, RTC laws are associated with substantial and statistically significant crime increases across multiple crime categories.  In others, the extent of the crime increase is more modest, or observable only in one crime category.  In other words, the results are not robust to model specification.
  • "Our ability to ascertain the best model is imperfect."  The authors conclude that "the 'best practices' in econometrics are evolving.  Researchers and policymakers should keep an open mind about controversial policy topics in light of new and better empirical evidence or methodologies."

For an in-depth discussion of the very complex statistical issues involved in analyzing the relationship between RTC and crime, please refer to Aneja, Donohue, and Zhang's article.

Read more: resources:
Gun Violence in America (
Fear of Crime (
Crime Victimization in the US: A Data-Driven Learning Guide (
Generational Trends in Attitudes about Gun Ownership: A Data-Driven Learning Guide (
Crime and Victims Statistics (

Monday, November 17, 2014

Record One In 30 U.S. Children Is Homeless, According to Latest Report

A report just issued by the National Center on Family Homelessness, "America's Youngest Outcasts," presents the latest figures on child homelessness in the United States.  The report is based on over a dozen established data sets and the most recent federal data that comprehensively counts homeless children.

According to the report, in the wake of the Great Recession, the number of children lacking homes rose from 1.2 in 2007 to 1.6 million in 2010. The number increased 8-10 percent in 31 states and the District of Columbia between 2012 and 2013 to reach 2.5 million (or one in every 30 U.S. children) in 2013 -- a historic high.  California accounts for more than one-fifth of the homeless children population.

The National Center on Family Homelessness also ranked states according to a composite score that reflects each state's overall performance across four domains: prevalence of child homelessness; child well-being; risk for child homelessness; and state policy planning and efforts.  Minnesota, Nebraska, and Massachussetts ranked highest, while California, Mississippi, and Alabama ranked lowest.

The authors explain that:
"Major causes of homelessness for children in the U.S. include: (1) the nation’s high poverty rate; (2) lack of affordable housing across the nation; (3) continuing impacts of the Great Recession; (4) racial disparities; (5) the challenges of single parenting; and (6) the ways in which traumatic experiences, especially domestic violence, precede and prolong homelessness for families. 
The impact of homelessness on the children, especially young children, is devastating and may lead to changes in brain architecture that can interfere with learning, emotional self-regulation, cognitive skills, and social relationships. The unrelenting stress experienced by the parents, most of whom are women parenting alone, may contribute to residential instability, unemployment, ineffective parenting, and poor health."

Read more: resources:
Homelessness: A Data-Driven Learning Guide (

Wednesday, November 12, 2014

Online Dating Aggression Linked to Neighborhood Violence Exposure

A University of Michigan research team led by Quyen Epstein-Ngo, a research assistant professor at the U-M Institute for Research on Women and Gender and a fellow with the U-M Injury Center, has begun investigating  risk and promotive factors associated with technology-delivered dating aggression (TDA) and relations between types of violence (physical dating/nondating, community violence, and TDA).  Preliminary baseline data from their ongoing study examining violent experiences among urban youth were recently published in the journal Violence and Gender.

The researchers hypothesized that: (1) risk factors would be positively associated with TDA, (2) promotive factors would be negatively associated with TDA, and (3) TDA would be associated with higher levels of physical dating violence, physical nondating violence, and community violence exposure.

They found that:

  • 48 percent of the youth in the sample reported technology-delivered dating aggression, 44 percent experienced dating violence (either as aggressors or victims), 55 percent reported involvement with physical nondating violence and nearly 96 percent reported community violence exposure.  The patterns of TDA, physical violence, and community violence exposure were similar across gender.
  • TDA was prevalent among high-risk urban youth.
  • There was a strong association between TDA and neighborhood violence exposure—for example, hearing gunshots, seeing drug deals, seeing someone get shot or stabbed.
  • There was a strong association between TDA and physical dating violence as well.
  • Mindfulness ("present-moment awareness of the sensations and stimuli that prompt these behaviors") was associated with less TDA.

"For individuals who would have perpetrated dating aggression regardless of the medium, technology may just be another way to monitor, control and threaten their dating partners," Epstein-Ngo said.  "However, there is also research to support the idea that individuals are more likely to say things online or over email that they wouldn't say if they were face-to-face with someone. So having access to technology may also make it easier to say mean things about others or do things that we wouldn't normally do if we were to look someone in the eye." 
"This study shows that TDA is related to physical dating aggression and may be a precursor to or a symptom of serious physical violence among dating partners. Not everyone who committed TDA also perpetrated physical violence, but nearly everyone who perpetrated physical violence also committed TDA. More research is needed to better understand this association."

Read more: resources:
Interpersonal Power in Intimate Relationships: A Data-Driven Learning Guide (

Monday, November 10, 2014

Could President Obama's Current Approval Ratings Inform The 2076 Election?

The New York Times' The Upshot recently reported on a new model of presidential voting created by Andrew Gelman, a political scientist and statistician at Columbia University, and Yair Ghitza, chief scientist at Catalist, a Democratic data firm.

The model is based on the idea that political preferences are shaped by a lifetime of events and life experiences, but more particularly those happening between the ages of 14 and 24: according to the model, "events at age 18 are about three times as powerful as those at age 40."  Gelman and Ghitza used Gallup's presidential approval ratings as a proxy for those events and estimated how people's political preferences change at different stages of their lives.  The model focuses on White voters because of the data limitations of older survey designs.

"Knowing how formative events are at different ages, along with the president’s approval rating, allows Mr. Ghitza and Mr. Gelman to estimate a group’s presidential voting tendencies over time, including during childhood. These preferences are not necessarily how the group voted, because they do not take into account short-term shifts, like Mr. Obama’s wide popularity in 2008. Think of them, instead, as estimates of how a group would vote in an average presidential election."

"It is important not to think about a single election or of a single defining political event,” Mr. Ghitza and Mr. Gelman write. “Rather, generations appear to be formed through a prolonged period of presidential excellence."

Read more: resources:
Voter Turnout in the United States: A Data-Driven Learning Guide (
Identity Politics and the Latino vs. Hispanic Debate: A Data-Driven Learning Guide (
Voting Behavior: The 2012 Election (

Tuesday, November 4, 2014

2014 Midterm Elections and The Turnout Gap

As American citizens line up to vote today, the Pew Research Center and the Washington Post offer two windows on the midterm elections.

The Pew's "6 Facts About The Electorate On Midterm Day" presents a snapshot of the electorate, an overview of the mood and opinions of the general public and those who are likely to vote.  The Pew notes that voters appear pessimistic about the economy despite signs of recovery.  They're also disillusioned about President Obama and their Representatives in Congress.  Republicans are seen as the party that could do a better job on key issues, while Democrats are seen as the party more willing to work with the opposition and more concerned with "the needs of people like me."  If the Pew's analyses are correct, the voters most likely to turn out for the midterm elections are "Steadfast Conservatives (staunch critics of government), Business Conservatives (prefer limited government but more moderate on social issues) and Solid Liberals (solidly Democratic)."

The Washington Post gives us an interactive tool to visualize the midterm turnout gap -- the lower voter turnout in midterm elections compared with presidential election years.  The Washington Post notes that this trend "typically favors Republican candidates. Voters who fall into demographic groups that have traditionally been Democratic strongholds — young adults, African Americans and women — have wider turnout gaps than other groups, a challenge that Democratic candidates must negotiate every midterm election year."  The interactive tool allows users to track the turnout gap since 1996 for the entire electorate, as well as by gender, race, age, education, and marital status.

To access the interactive tool, please go to:

Read more: resources:
Voter Turnout in the United States: A Data-Driven Learning Guide (
Voting Behavior: The 2012 Election (

Monday, November 3, 2014

New Interactive Tools from The Economist

The Economist has released new interactive economic indicators that allow users to customize and share their own tables to tell their own stories.

The  full economic and financial indicators table is updated twice daily.  Data can be organized by country and category.  Users can also "choose between two tables: one covering output, prices and jobs; the other offering trade, exchange and interest rates. Select regional and economic country groupings, such as the G20 and the BRICS beneath the 'All' tab. Sort columns by country or rank by category. Highlight and track up to five focus countries by selecting individual rows."

Tables can then be saved and shared.  For more information, go to:

Read more:

Tuesday, October 28, 2014

Should People Be Able to Choose The Manner In Which They Prefer to Die? Americans Are Divided

A new report ("Dying in America: Improving Quality and Honoring Individual 
Preferences Near the End of Life") from the Institute of Medicine's Committee on Approaching Death calls for broad changes in the way this country handles end-of-life care.  The 21-member nonpartisan panel composed of doctors, nurses, insurers, religious leaders, lawyers and experts on aging calls for a "major reorientation and restructuring of Medicare, Medicaid and other health care delivery programs" and the elimination of "perverse financial incentives that encourage expensive hospital procedures when growing numbers of very sick and very old patients want low-tech services like home health care and pain management" (quoted in this New York Times article).

Americans seem to agree: two thirds of respondents to a Pew Research Center survey believe that there are circumstances in which doctors should not do everything possible to save a patient's life and the patient should be allowed to die.  But the minority who say that medical professionals always should do everything possible to save a patient’s life is growing (from 15 percent in 1990, to 31 percent in 2013).  A deeper look at the data reveals that views about euthanasia and doctor-assisted suicide are influenced by race/ethnicity, religion, political ideology, as well as by the circumstances (does the person suffer from an incurable disease?  Is she suffering?  Is she a burden to her loved ones?).

When it comes to doctor-assisted suicide, the public is evenly divided according to the Pew survey: 47% approve and 49% disapprove of laws that would allow a physician to prescribe lethal doses of drugs that a terminally ill patient could use to commit suicide.

Dr. Victor J. Dzau, the Institute of Medicine’s president, said that "patients don't die in the manner they prefer."  But the Pew survey results indicate that attitudes about end-of-life care and decisions are complicated.  If this country is to develop a "modernized end-of-life care system," the law may need to accommodate the wide range of end-of-life choices that an increasingly diverse American population wants.

Euthanasia can be classified as either active or passive and as either voluntary or involuntary. In active euthanasia, specific steps are taken to cause the patient's death, such as injecting her with poison, or giving her an overdose of pain-killers. In contrast passive euthanasia refers to the withdrawal of medical treatment or the withholding of food and fluids with the deliberate intention of causing the patient's death. Voluntary euthanasia is when the patient requests that action be taken to end her life, or that life-saving treatment be stopped. Involuntary euthanasia is when a patient's life is ended without the patient's knowledge and consent, usually because she is unconscious, or too weak to communicate. Euthanasia differs from assisted suicide, where a physician provides lethal medications but the patient decides whether and when to ingest them.

Read more: resources:
Euthanasia: A Data-Driven Learning Guide (