What factors correlate with suicide rates?

LiveStories cityscape plain blue.jpg

Suicide is the tenth leading cause of death in the United States, according to data from the CDC. The suicide rate has risen steadily since the mid-2000’s. In 2016, nearly 45,000 Americans took their own lives—more than the number of deaths from automobile accidents or opioid overdoses.

The CDC breaks down deaths by self-harm into basic demographic categories: age, sex, and race/ethnicity. A more complex picture of suicide emerges when correlations with other demographic and health-related factors are considered.

Suicide Correlations

We compiled state-level suicide data, along with data for 16 other indicators (primarily from the CDC’s Behavioral Risk Factor Surveillance System), covering the years 2011 to 2016. The relation between state suicide rates and these other indicators can be visualized on a scatterplot chart. To begin, choose an indicator from the menu below:

There are several ways to quantify correlation; here we’ve calculated Pearson’s correlation coefficient for each indicator. Visually, you can think of the correlation coefficient as a scale that tells you: “how much do these data points form a diagonal line?” A perfect upward-sloping diagonal line has a coefficient of 1—for example, the scatterplot you see when “suicide rate” is correlated with itself. A diagonal line sloping downward, on the other hand, would be a -1 — a perfect negative correlation.

Before we get deeper into the data, an important caveat:

Correlation is not causation. In other words, just because two factors are correlated does not mean we can say one causes the other. Correlation is simply a way of quantifying the relationship between two factors. The relationship between these factors and the suicide rate may point to some deeper connection—or they may be entirely incidental. 

Correlation Coefficients for the Suicide Rate

With that caveat in mind: the chart below shows the correlation coefficients for each of the 16 indicators, relating each to the suicide rate. 

Some of these relationships are well-known. States with larger percentages of veterans tend to have higher suicide rates—a relation known to the Department of Veteran's Affairs, which has made suicide prevention one of its top priorities. States with high percentages of people reporting depression diagnoses also have a somewhat strong correlation with suicide. 

Negative economic indicators have surprising relations to suicide. It is widely understood that unemployment, especially long-term unemployment, can exact a serious mental and physical toll on workers. However, this indicator (which excludes retired people and homemakers) has a negative correlation with suicide deaths. Low household income has almost no correlation with suicide.

Suicide, Marriage, and Divorce

The strongest correlation with suicide, of the 16 indicators we looked at, is divorce. This, too, may not be surprising—divorce can be a psychologically traumatic experience.

One might expect that a strong relationship with divorce might imply that marriage has an inverse relationship with suicide rates; conventional wisdom and some research holds that married people live longer and healthier lives. However, marriage is also strongly correlated with suicide. In fact, its correlation coefficient (0.44) is stronger than that of depression (0.38). The prevalence of "never married" people, in contrast, has almost no correlation with the suicide rate. 

As noted above, great care must be taken when drawing conclusions from data about correlations. Suicide is complex, and it is unlikely that any single factor can adequately explain an individual’s suicide. If you have comments, questions, or suggestions for other indicators to include in this or future analyses, let us know.

Individuals, the media, and public health departments all have important roles to play in reducing the risk of suicide. The CDC released a technical package (pdf) of suicide-prevention policies, programs, and practices. Individuals interested in learning more about preventing suicide can visit the National Suicide Prevention Hotline's website, Be the One to Save a Life.

About the Data

The suicide death rate and opioid overdose death rate are from CDC Wonder, queried with the following ICD-10 codes for underlying cause of death: X60-X84, Y87.0.

The veteran percentage indicator is from the American Community Survey.

All other indicators are from the CDC's Behavioral Risk Factor Surveillance System (BRFSS).