Mental health matters

«Depression: let’s talk»—this was the front page message of the World Health Organization (WHO) web page on Thursday April 7th. Every year, the WHO calls attention to a global health issue through the World Health Day. This year, the focus is on mental health.

Mental health is a vast topic. According to, a branch of the US Department of Health, mental health can be loosely defined as «including our emotional, psychological, and social well-being. It affects how we think, feel, and act. It also helps determine how we handle stress, relate to others, and make choices» (1).

It is important that the WHO calls attention to mental health. In 2014, there were 42 773 suicides in the United States only, i.e. there were twice as many suicides as homicides. This rate has increased by 24% between 1999 and 2014 (2), and this makes suicide the 10th leading cause of death in the country. Some of these 10 causes can hardly be prevented, like pneumonia or alzheimer’s disease. Suicide, on the other hand, can be.

Some elite universities, such as MIT and Harvard, have above average suicide rates for college campuses (3) and privileged areas also have surprisingly high statistics. Palo Alto leads the country’s youth suicide rates (4)(5).

But suicide rates are not uniform across the country. The National Institute of Mental Health has collected suicide statistics per state and compared the numbers across the country using the data (6).

We made our own map of suicide rates with the data provided by the NIMH.

There are huge disparities across the United States. The rate of suicide in the state of Nevada is roughly three times higher than in the state of California. Alaska and the western states—with the exception of California—have much higher rates than the rest of the country. But why? How could one explain such a geographic divide for state level suicide rates?

Several sociological perspectives on suicide

Sociologists have studied the question of why these western states, often referred to as the “suicide belt”, have much higher suicide statistics, and have come up with a few possible explanations. Baller and Richardson (7) suggest that the West has lower family and religious integration, leading to more suicides. This idea is old and was described in Émilie Durkheim's book Suicide (1897), where Durkheim explained that integration into a community such as the church could decrease feelings of isolation and therefore decrease suicide rates. (8) Durkheim said that Romanatholics have lower suicide rates than other religions, suggesting that committing suicide would be against the Roman Catholic values.

Barkan, Rocque, and Houle (9) instead suggest that residential stability and population density are key explanations for the geographic divide in suicide rates. Areas with low residential stability have many newcomers and temporary residents who lack stable social ties and have weaker social institutions. Further, Houle et al. argue that low population density leads to less social interaction and fewer weak social ties, leading to overall less social integration. Relately, Robert Putnam explained in his book Bowling Alone that mobility has increased over time and he postulates that it has led to weaker social ties. This increase in mobility might explain part of the increase in suicide rates over time.

“[T]he inner mountain west is a place that is disproportionately populated by middle-aged and aging white men, single, unattached, often unemployed with access to guns. This may turn out to be a very powerful explanation and explain a lot of the variance that we observe,” said Matt Wray, a Professor of sociology at Temple University, in an interview with Stephen J. Dubner in the podcast episode “The Suicide Paradox” for the Freakonomics Radio podcast. (10)

Suicide is complex and does not have a single explanation. Suicide cannot be studied with randomized controlled trials, so one cannot robustly establish causal relationships. Moreover, the motives stay very often unknown. We try here to look at some state-level to illuminate aspects of the problem as a step towards a better understanding.

Our investigation

While we had hoped we could find some clear patterns of variables correlating with suicide rates, we quickly learned that even that—even just finding good correlations without any claims of causation—was difficult. We analyzed the following variables:

The different data sets are mapped below.

Suicide rate

Median household income

Income inequalities

Gun penetration

Share of population living in urban areas

Minorities share

Share of population under 18

Share of White population

Income Share

We established that it is difficult to find robust correlations between any of the variables and suicide rate by comparing these maps. Even if we did find any strong correlations, this would not imply causation. Moreover, suicide is driven by many different factors interacting, and it is highly unlikely that any single variable would have a significant influence on suicide statistics on its own.

There seems to be some relation between several variables and suicide rates. States with low rates of minorities seem to have lower suicide rates. Correspondingly, highly white states seem to partly have higher suicide rates,even if less blatantly than in the case of minorities share. For example, states such as Idaho, Montana, Wyoming and Utah have higher suicide rates and a largest share of their population is white. However, in the eastern part of the United States, the states of New Jersey, Delaware, Maryland, Connecticut, and Rhode Island are demographically very white but still have don’t have higher suicide rates.

Gun penetration also seems to correlate with suicide rates in the Western states of the country, in Alaska, Montana, Wyoming and South Dakota. However, in Alabama and Mississippi, suicide rates are low but gun penetration is high.

The share of the population living in urban areas, the share of the population under 18 years of age, income inequality, and median household income, do not seem correlated with suicide rates on our maps. Note that while finding correlations does not imply causation, the reverse is also true: Not finding correlation does not mean that the variable is irrelevant.

More analysis needed

Overall, we can’t draw a hard conclusion. Data enables us to see some patterns emerge, but these patterns would need to be further verified and cross checked.

If we want to better understand possible causes of higher suicide rates, we need to do analysis accounting for several variables at the same time. Further, need to study suicide rates at a much more granular level. For now, we can only confirm what we already knew: 1) suicide is incredibly complex, 2) suicide rates in the U.S. are increasing and non-uniform across the country, and 3) mental health deserves more attention.


  1. Retrieved from:
  2. Curtin, Sally C., Margaret Warner, and Holly Hedegaard. 2016. “Increase in Suicide in the United States, 1999-2014”, NCHS Data Brief No. 241, U.S. Department of Health and Human Services. Retrieved from:
  3. Rocheleau, Matt. 2015. “Suicide rate at MIT higher than national average”. Boston Globe, March 17th 2015. Retrieved from:
  4. Lee, Jacqueline. 2017. “CDC report: Youth suicide rate in Santa Clara county highest in Palo Alto, Morgan Hill”, The Mercury News, March 17th 2017. Retrieved from:
  5. Rosin, Hanna. 2015. “The Silicon Valley suicides. Why are so many kids with bright prospects killing themselves in Palo Alto?”, The Atlantic. Retrieved from:
  6. Retrieved from:
  7. Baller, Robert D. and Kelly K. Richardson. 2002. “Social Integration, Imitation, and the Geographic Patterning of Suicide.” American Sociological Review 67:873–88
  8. Émile Durkheim. 1897. Suicide
  9. Houle, Jason, Steven E. Barkan, and Michael Rocque. “State and Regional Suicide Rates: A New Look at an Old Puzzle.” Sociological Perspectives, Vol. 56, Issue 2, pp. 287–297. Retrieved from:
  10. Dubner, Stephen. 2011. “The Suicide Paradox”, New Freakonomics Radio Podcast. Retrieved from: