Canadian Crime Rates in the Penalty Box

by   Simon Demers, et al.

Over the 1962-2016 period, the Canadian violent crime rate has remained strongly correlated with National Hockey League (NHL) penalties. The Canadian property crime rate was similarly correlated with stolen base attempts in the Major League Baseball (MLB). Possible implications and avenues for future research are discussed.



page 1

page 2

page 3

page 4


Racial Disparities in the Enforcement of Marijuana Violations in the US

Racial disparities in US drug arrest rates have been observed for decade...

Report from the NSF Future Directions Workshop, Toward User-Oriented Agents: Research Directions and Challenges

This USER Workshop was convened with the goal of defining future researc...

Dimension-Free Anticoncentration Bounds for Gaussian Order Statistics with Discussion of Applications to Multiple Testing

The following anticoncentration property is proved. The probability that...

Multiscale cross–correlations and triangular arbitrage opportunities in the Forex

Multifractal Detrended Cross-Correlation methodology is applied to the f...

Guessing Based on Compressed Side Information

A source sequence is to be guessed with some fidelity based on a rate-li...

Visible Insights of the Invisible Pandemic: A Scientometric, Altmetric and Topic Trend Analysis

The recent SARS-COV-2 virus outbreak has created an unprecedented global...

Identifying strongly correlated groups of sections in a large motorway network

In a motorway network, correlations between the different links, i.e. be...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.

1 Introduction

Uncovering strong empirical correlations between crime rates and other trends has the potential to be useful because it provides new opportunities to test, validate or refute criminological theories. In the United States, for example, there appears to be a strong negative correlation between measures of consumer sentiment and rates of not only robbery and property crime (Rosenfeld and Fornango, 2007) but also homicide (Blumstein and Rosenfeld, 2008) and felony homicide in particular (Rosenfeld, 2009). This suggests that changing economic conditions can have an impact on crime. Similarly, divorce rates and crime rates also seem to move together over time (Greenberg, 2001), supporting the idea that major social changes and family strains can engender crime.

This short research note highlights surprisingly strong correlations between Canadian aggregate crime trends and certain professional sport trends over the 1962-2016 period. Beyond any entertainment value, these statistical relationships have the potential to open new avenues of investigation for criminological research.

2 Data

The analysis relies on police-reported data compiled by Statistics Canada as part of its Uniform Crime Reporting (UCR) Survey. The aggregate UCR Survey data permits historical comparisons back to 1962 (Allen, 2018). Following convention, crime numbers are converted into rates per 100,000 resident population to facilitate year-over-year comparisons that take into account population growth. The focus of the analysis is on violent crime, property crime, and homicide rates.

Penalty data for the National Hockey League (NHL) was extracted directly from its official website (, 2018)

. This NHL data includes the total number of penalties (Pen) and the associated number of penalty infraction minutes (PIM) for each season. Two NHL lockouts shortened the 1994-95 and 2012-13 regular seasons to 48 games instead of 82 games per team. The totals for those seasons are manually adjusted (normalized) to account for lost games and make year-over-year comparisons possible. For its part, the 2004 NHL lockout led to a completely cancelled season. In that case, fictitious 2004 data is imputed by simply interpolating between 2003 and 2005 data points.

Batting data for the Major League Baseball (MLB) was downloaded from the website (2018). This MLB data includes various statistics intending to reflect offensive productivity. The analysis focuses on stolen base attempts or stolen base opportunities, which is the combination of stolen bases (SB) and “caught stealing” events (CS). Most MLB seasons over the analysis period were disputed over 162 games per team but three significant MLB strikes resulted in lost games during the 1972, 1981, 1994, and 1995 seasons. However, the MLB data is averaged on a per game basis, which removes the need to normalize the data or simulate full seasons in order to account for lost games.

3 Findings

3.1 Violent crime in the NHL penalty box

First, it turns out that the Canadian violent crime rate is positively correlated with the number of NHL penalties and penalty minutes in a surprisingly strong manner, as measured by Pearson’s correlation coefficients ( and 0.925, respectively). Fig. 1 illustrates the quality of the linear fit. The ordinal association is also quite strong, as measured by Kendall’s rank correlation coefficients ( and 0.750, respectively).

Figure 1: Canadian violent crime rate and NHL penalties, 1962-2016. and .

This statistical relationship is notable because NHL penalties are typically assessed by referees for excessive toughness and physical violence and are therefore the closest thing to a violent crime in the context of professional ice hockey (Heckelman and Yates, 2003), besides actual on-ice criminal assaults (CityNews Staff, 2016).

Possible underlying factors that could drive such a strong correlation between violent crime and NHL penalties include all factors that professional hockey players would have been exposed to with essentially the same timing and same intensity as the rest of the general population. This could plausibly encompass, for example, early childhood lead exposure (Nevin, 2000, 2007) and a social civilizing or pacifying process (Spierenburg, 2001; Stille, 2003; Restrepo, 2015).

It is not possible at this stage to completely eliminate the possibility that NHL penalties might actually stimulate violent crime at the national level, through some kind of copycat or emulation effect for instance (Moser, 2004; Adubato, 2011; Card and Dahl, 2011; Kirby, Francis and O’Flaherty, 2014).

We note in passing that the correlation between violent crime and NHL penalties is unlikely to be explained away by annual weather or temperature variations (Burke, Hsiang and Miguel, 2015) because NHL games are generally played in a climate-controlled environment, unlike baseball (Reifman, Larrick and Fein, 1991; Larrick et al., 2011) or American football (Craig et al., 2016) games for instance.

3.2 Property crime around the baseball diamond

Second, it turns out that the Canadian property crime rate is positively correlated with the rate of stolen base attempts in the MLB. Fig. 2 illustrates the quality of the linear fit.

Figure 2: Canadian property crime rate and stolen base attempts in the MLB, 1962-2016. and .

Stolen base attempts reflect the propensity of baseball players to opportunistically try to move forward around the bases at the expense of the defensive team, in order to possibly score more runs as a result. Stolen base opportunities can therefore properly be seen as the professional baseball equivalent of a lawful property crime.

Plenty of explanations have been professed to explain the relative rise of base stealing in the mid-1970s (McMurray, 2015), its peak throughout the 1980s (James, 1982), and its decline in the mid-1990s (Kurkjian, 1994). However, identifying the underlying factors that could explain why stolen bases in the MLB could be related to property crime in Canada remains an open problem.

As a side note, Farrell, Hodgkinson and Andresen (2018) already highlighted that there was a remarkable similarity between historical property crime and homicide rates in Canada, once attempted murders and actual homicides are aggregated into a single measure of total homicides (). Canadian homicides and attempted murders, like property crimes, are also correlated with stolen base attempts ( and ). This is shown by Fig. 3.

Figure 3: Canadian homicide rate and stolen base attempts in the MLB, 1962-2016. and .

The Canadian homicide rate is actually correlated most strongly with failed stolen base attempts where the runner was caught stealing (), while the rate of attempted murders is actually correlated most strongly with successful stolen bases (). This may look incongruent at first glance but this is only because stolen bases are defined from the point of view of the runner. From the point of view of the defensive team, a runner who is caught stealing represents a successful out.

4 Conclusion

Correlation does not imply causation and it is entirely possible that the statistical relationships presented in this short research note may be coincidental or spurious. However, they could very well be examples of life imitating sports (Steen, 2014), or vice versa (Goldstein, 1989). Earlier studies have revealed a correlation between sports teams success and reductions in suicide (Joiner Jr., Hollar and Van Orden, 2006) and even homicide (Fernquist, 2000) rates.

The violent crime rate in Canada is correlated with NHL penalties to such an extent that it is difficult to see how the statistical relationship could be coincidental. However, more research would obviously be required to assess to what extent both trends might be influenced by other underlying factors or whether it might be fueled endogenously through a copycat effect, among many other possible pathways.

Canadian property crime rates and homicide rates, for their part, are positively correlated with stolen base attempts in the MLB. Farrell, Hodgkinson and Andresen (2018)

speculated that similarities between historical homicide and property crime rates may have been driven by the progressive implementation of security improvements, especially in the second half of the 1990s. While it is not immediately obvious whether (or how) this security hypothesis may also apply to stolen base opportunities in the MLB, more research into this area could be enlightening (and probably entertaining).


  • Adubato (2011) [author] Adubato, BethB. (2011). Fanning the Flames: Televised, Professional Football Games and Domestic Violence, PhD thesis DOI: 10.7282/T347498K.
  • Allen (2018) [author] Allen, MaryM. (2018). Police-reported crime statistics in Canada, 2017 Statistics Canada, Canadian Centre for Justice Statistics Retrieved from:
  • (2018) [author] Baseball-Reference. com (2018). Major League Baseball Batting Year-by-Year Averages Retrieved from:
  • Blumstein and Rosenfeld (2008) [author] Blumstein, AlfredA. Rosenfeld, RichardR. (2008). Factors Contributing to U.S. Crime Trends. In Understanding Crime Trends: Workshop Report (Arthur S.A. S. Goldberger RichardR. Rosenfeld, eds.) 2, 13–44. The National Academies Press, Washington D.C. Available at:
  • Burke, Hsiang and Miguel (2015) [author] Burke, MarshallM., Hsiang, Solomon M.S. M. Miguel, EdwardE. (2015). Climate and Conflict. Annual Review of Economics 7 577–617. DOI: 10.1146/annurev-economics-080614-115430.
  • Card and Dahl (2011) [author] Card, DavidD. Dahl, Gordon B.G. B. (2011). Family Violence and Football: The Effect of Unexpected Emotional Cues on Violent Behavior. The Quarterly Journal of Economics 126 103–143. DOI: 10.1093/qje/qjr001.
  • Craig et al. (2016) [author] Craig, CurtisC., Overbeek, Randy W.R. W., Condon, Miles V.M. V. Rinaldo, Shannon B.S. B. (2016). A Relationship Between Temperature and Aggression in NFL Football Penalties. Journal of Sport and Health Science 5 205–210. DOI: 10.1016/j.jshs.2015.01.001.
  • Farrell, Hodgkinson and Andresen (2018) [author] Farrell, GrahamG., Hodgkinson, TarahT. Andresen, Martin A.M. A. (2018). Homicide in Canada and the Crime Drop. Crime Science 7 1–3. DOI: 10.1186/s40163-017-0076-y.
  • Fernquist (2000) [author] Fernquist, Robert M.R. M. (2000). An Aggregate Analysis of Professional Sports, Suicide, and Homicide Rates: 30 U.S. Metropolitan Areas, 1971-1990. Aggression and Violent Behavior 5 329–341. DOI: 10.1016/S1359-1789(98)00012-3.
  • Goldstein (1989) [author] Goldstein, Jeffrey H.J. H. (1989). Violence in Sports. In Sports, Games, and Play: Social and Psychological Viewpoints 10, 279–297. Lawrence Erlbaum Associates, New Jersey.
  • Greenberg (2001) [author] Greenberg, David F.D. F. (2001). Time Series Analysis of Crime Rates. Journal of Quantitative Criminology 17 291–327. DOI: 10.1023/A:1012507119569.
  • Heckelman and Yates (2003) [author] Heckelman, Jac C.J. C. Yates, Andrew J.A. J. (2003). And a Hockey Game Broke Out: Crime and Punishment in the NHL. Economic Inquiry 41 705–712. DOI: 10.1093/ei/cbg038.
  • James (1982) [author] James, BillB. (1982). So What’s All The Fuss? Sports Illustrated. Retrieved from:
  • Joiner Jr., Hollar and Van Orden (2006) [author] Joiner Jr., Thomas E.T. E. J., Hollar, DanielD. Van Orden, KimberlyK. (2006). On Buckeyes, Gators, Super Bowl Sunday, and the Miracle on Ice: “Pulling Together” is Associated With Lower Suicide Rates. Journal of Social and Clinical Psychology 25 179–195. DOI: 10.1521/jscp.2006.25.2.179.
  • Kirby, Francis and O’Flaherty (2014) [author] Kirby, StuartS., Francis, BrianB. O’Flaherty, RosalieR. (2014). Can the FIFA World Cup Football (Soccer) Tournament Be Associated with an Increase in Domestic Abuse? Journal of Research in Crime and Delinquency 51 259–276. DOI: 10.1177/0022427813494843.
  • Kurkjian (1994) [author] Kurkjian, TimT. (1994). Baseball. Sports Illustrated. Retrieved from:
  • Larrick et al. (2011) [author] Larrick, Richard P.R. P., Timmerman, Thomas A.T. A., Carton, Andrew M.A. M. Abrevaya, JasonJ. (2011). Temper, Temperature, and Temptation: Heat-Related Retaliation in Baseball. Psychological Science 22 423–428. DOI: 10.1177/0956797611399292.
  • McMurray (2015) [author] McMurray, JohnJ. (2015). Examining Stolen Base Trends by Decade from the Deadball Era through the 1970s. Baseball Research Journal. Retrieved from:
  • Moser (2004) [author] Moser, Carrie A.C. A. (2004). Penalties, Fouls, and Errors: Professional Athletes and Violence Against Women. Sports Law Journal 11 69–75.
  • Nevin (2000) [author] Nevin, RickR. (2000). How Lead Exposure Relates to Temporal Changes in IQ, Violent Crime, and Unwed Pregnancy. Environmental Research 83 1–22. DOI: 10.1006/enrs.1999.4045.
  • Nevin (2007) [author] Nevin, RickR. (2007). Understanding International Crime Trends: The Legacy of Preschool Lead Exposure. Environmental Research 104 315–336. DOI: 10.1016/j.envres.2007.02.008.
  • (2018) [author] NHL. com (2018). Statistics - Teams - Season By Season - Regular Season - Penalties Retrieved from:
  • Reifman, Larrick and Fein (1991) [author] Reifman, Alan S.A. S., Larrick, Richard P.R. P. Fein, StevenS. (1991). Temper and Temperature on the Diamond: The Heat-Aggression Relationship in Major-League Baseball. Personality and Social Psychology Bulletin 17 580–585. DOI: 10.1177/0146167291175013.
  • Restrepo (2015) [author] Restrepo, PascualP. (2015). The Mounties and the Origins of Peace in the Canadian Prairies. DOI: 10.13140/RG.2.1.4725.5121.
  • Rosenfeld (2009) [author] Rosenfeld, RichardR. (2009). Crime is the Problem: Homicide, Acquisitive Crime, and Economic Conditions. Journal of Quantitative Criminology 25 287–306. DOI: 10.1007/s10940-009-9067-9.
  • Rosenfeld and Fornango (2007) [author] Rosenfeld, RichardR. Fornango, RobertR. (2007). The Impact of Economic Conditions on Robbery and Property Crime: The Role of Consumer Sentiment. Criminology 45 735–769. DOI: 10.1111/j.1745-9125.2007.00096.x.
  • Spierenburg (2001) [author] Spierenburg, PieterP. (2001). Violence and the Civilizing Process: Does It Work? Crime, Histoire & Societes / Crime, History & Societies 5 87–105. DOI: 10.4000/chs.740.
  • CityNews Staff (2016) [author] CityNews Staff (2016). Some notable on-ice incidents that led to criminal charges. CityNews Toronto. Retrieved from:
  • Steen (2014) [author] Steen, RobR. (2014). Floodlights and Touchlines: A History of Spectator Sport. Bloomsbury Sport ISBN: 9781408152157.
  • Stille (2003) [author] Stille, AlexanderA. (2003). Did Knives and Forks Cut Murders? Counting Backward, Historians Resurrect Crime Statistics And Find the Middle Ages More Violent Than Now. The New York Times B7. Retrieved from: