A home security company recently published a blog post listing Hope, Arkansas as the ninth most dangerous city in Arkansas. The same article has Texarkana, Arkansas ranked as the 5th most dangerous city in Arkansas. Texarkana, Texas Police Department recently addressed this in early September when the company listed Texarkana, Texas as the second most dangerous city in Texas. In a statement on the TTPD Facebook page at the time they said, “The UCR is truly only good for showing crime trends within a particular community over a period of time – not for saying one city is safer than another.”
What follows is a response from the Hope, Arkansas Police Chief.
Recently I became aware of information posted on the website of Home Security Shield entitled “15 Most Dangerous Cities in Arkansas.” Hope was listed as the ninth most “dangerous” city by this company based on a composite score of selected property and violent crimes listed on the FBI Uniform Crime Report site.
Obviously, I am not writing to dispute the numerical data provided to the FBI Uniform Crime Report. This data is provided by the Hope Police Department and the numbers are what they are. I am writing to challenge the inference drawn from the data that characterizes cities on the list as “Most Dangerous”. Home Security Shield’s conclusion is misleading, incomplete and inappropriate.
Statistics, in practice, can be divided into 3 areas: Data Analysis; Data Production; and Statistical Inference. “Statistical Inference moves beyond the data in hand to draw conclusions about some wider universe” (David S. Moore; The Basic Practice of Statistics). The FBI UCR website clearly warns, “UCR data are sometimes used to compile rankings of individual jurisdictions and institutions of higher learning. These incomplete analyses have often created misleading perceptions which adversely affect geographic entities and their residents. For this reason, the FBI has a long-standing policy against ranking participating law enforcement agencies on the basis of crime data alone. Despite repeated warnings against these practices, some data users continue to challenge and misunderstand this position… the FBI cautions and, in fact, strongly discourages, data users against using rankings to evaluate locales or the effectiveness of their law enforcement agencies.”
Based purely on UCR violent crime data measuring Murder, Rape, Robbery, and Aggravated Assault for 2012 and 2013, citizens of Hope had a statistical chance of .95% and .82% (less than 1%) respectively of being a victim of a violent crime. For 2013 this is a 25/100ths of one percent difference from the mean for all departments in Arkansas submitting data to the UCR program. Taking into account that approximately 73% of all violent crime in Hope, Arkansas is committed by a family member, friend, or acquaintance, this means that a citizen of Hope has a .22% chance (less than 1 quarter of one percent) of being a victim of a violent crime by a stranger. The data suggests that your chance of becoming a victim of a violent crime is more strongly correlated to whom you marry, befriend or associate with as opposed to your geographic location. Even greater correlations may exist with education, employment, poverty, etc. These figures do not even attempt to address repeat offenders involving the same victim, such as is more common with domestic matters. Does this justify labeling a city as “Most Dangerous”?
Considering property crime data measuring burglary, theft, auto theft and arson for 2012 and 2013, citizens of Hope had a statistical chance of 6.7% and 5.4% respectively of being a victim of a property crime. For 2013, the mean for all Arkansas submitting agencies was 4.59%; this is a .81 % difference. Based on this standard alone, Hope would rank as 10th and 21st respectively. The largest data set in the property crime data category is theft. Hope reported 420 and 340 thefts for 2012 and 2013 respectively. Did you notice the approximate 20% difference in the two years? When you take into account that approximately 75% of all thefts and 50% of all property crime in the City of Hope are misdemeanor crimes (22% of which are associated with shoplifting or theft of motor fuel), the chance that a citizen will be the victim of a felony property crime is approximately 2.73%. Misdemeanor crime by its very descriptive nature is “minor” or “not serious.” These are crimes for which a person cannot be imprisoned. I think any serious thinking person would be hard pressed to call a city “Most Dangerous” based on this information.
Data users must consider the common sense interpretation of data and be careful drawing conclusions. For example, the City of Hope had no homicides for a 4 year period, 2008-2011. In 2012 we had 1 homicide. Would an article entitled, “Hope Homicide Rate Increases 100% for 2012” properly convey to data users useful information? Of course not! While the number is true mathematically, it does not convey useful information to the average reader. This often occurs in the world of small numbers.
It has been my experience, and I am sure it is your experience, that citizens of Hope can go anywhere in this city without fear of being assaulted, molested, or abused in any way. Do we have crime? Yes. What increases our statistical chance of becoming a victim: our association with disreputable people, our choice to remain in abusive family relationships, specific locations and time, lack of education, poverty, etc.
In the above paragraphs we have looked at UCR crime data. Another data set, known as NIBRS (National Incident Based Reporting), measures 23 offense categories made up of 49 specific crimes. Hope would rank 17th among all cities reporting this information in Arkansas for 2012. Had this data set been used in the same misleading manner, cities such as Caddo Valley, Paragould and Alma would be listed above Hope. How absurd to use these numbers in this way. All of us common sense folk can understand this.
I like to think of UCR and NIBRS data as a quick look at the health of a city. Much like a doctor looks at numbers associated with blood work and talks with patients about their symptoms. Data can help establish baselines and bring attention to areas of concern for diagnosis and treatment. You may have a cough, feel bad and have some abnormalities in your blood work, but there is a big difference in diagnosing these symptoms as a cold, flu, respiratory infection or cancer. If you provide a diagnosis of cancer, you certainly want to be sure to some scientific degree of certainty that your diagnosis is accurate.
In the case of Home Security Shield, it is my belief they have diagnosed the situation incorrectly and unfortunately caused confusion and worry on the part of some of our citizens. I hope I have been able to shed some light on these numbers to help you see them in a more balanced way.
As citizens, we have a responsibility and moral obligation to work together to bring about a better, more prosperous, and generous city for all. I personally am very proud to live in Hope, Arkansas and work toward this goal. Is there room for improvement, certainly! But make no mistake, we live in a safe community with many opportunities.