After declining for over two decades, homicides in the United States increased sharply in 2015 and 2016. This slowed a little bit in the years that followed, until another dramatic increase in homicides occurred in 2020. In fact, the 30% increase from 2019-2020 is the largest ever recorded. By 2021, homicides rose another 5%. This uptick was not as striking as the one seen in 2020, though the numbers were still higher than pre-2019.
One common misconception is that less restrictive gun control and higher levels of gun ownership in Republican states lead to more homicides. Problematically though, many of the studies claiming this don’t actually rely on measures of gun crime to generate these conclusions — they rely on gun death rates, which are comprised of homicides and suicides. When taking a closer look at the data on gun ownership and only gun homicides, there is no clear association between the two. Nonetheless, the media continues to purport false claims that red states have suffered from more murders than blue states.
As this news has come out, many are blaming Republican leadership as the reasoning for the homicide increase, citing poor policymaking and lack of gun control. In recent months, The Third Way (a non-profit thinktank) has also attempted to prove these sweeping claims, and their article has now been circulated widely and cited by many policymakers. But like the other studies discussed above, does the article stand up to scrutiny? In this post, I will take a closer look at the data to answer this question.
The data analyzed in the Third Way report is based on 2019 and 2020 state-level data, collected from yearly crime reports released by state governments as well as reputable local news sources (when official data were unavailable). Whether a state was considered Republican or Democratic was based on their presidential vote in the 2020 election.
The article claims that “the alarming murder increase in the U.S. is a problem that affects Republican-run jurisdictions as much or more than the Democratic-run jurisdictions.” They include the table below to show the top ten states with murder rate increases.
But a deeper dive into this data found the results to be not statistically significant. For this post, I took the raw data provided by the authors via their website, and ran some significance tests on it. For context, a p-value of 0.05 or lower would indicate a statistically significant or “real” finding.
First, I compared the percent change in murder rate for red states with blue states using an independent samples t-test. Because three states were missing values for 2019 (Alabama, Mississippi, and Missouri), they were not included in this calculation. The results showed no significant differences regarding the change in the murder rate comparing red states to blue states (two-sided p-value = 0.14).
Because political pary is often correlated with urbanicity, I explored whether urbanicity was a factor in the homicide increase. Using urbanicity data from the U.S. Census Bureau, I coded states as “non-urban” and “urban” based on ratios of urban to nonurban gerographical areas in each state. States that had ratios higher than the median were coded as “urban” while states that had ratios lower than the median were coded as “non-urban.” I then ran an independent samples t-test, and found that non-urban areas did see a greater increase in the murder rate relative to non-urban states (35.13% vs. 26.83%), but this was not statistically significant (two-sided p-value = 0.25).
To look at the impact of political affiliation and urbanicity together, I ran a multivariate linear regression model. When both variables were included in the model, neither had a significant impact on the murder rate increases (two-sided p-values = 0.23 for political party and .44 for urbanicity).
The dramatic homicide increase that occurred from 2019-2020 was noteworthy in nearly all states. On its face, it looks like areas with Republican leadership saw greater increases in homicides relative to areas with Democratic leadership. Similarly, it appears that non-urban areas also saw greater increases relative to urban areas. However, when looking at the data more closely, it appears that neither political party nor urbanicity impacted the percent change in homicides. That said, it is overly simplistic to claim that Republican policies are to blame.
More research is needed to better understand why homicides have increased so much. Future research might examine year-to-year changes in homicide rates, and whether that correlates with political party, urbanicity, or other factors (e.g., gangs, drug use, proactive policing). In addition, future research could explore differences within red states and within blue states to better understand some of the factors affecting the homicide increase, apart from political party.