U.S. Marijuana Arrests

Part One: County Level Arrest Data 1995-1997
By Jon Gettman, Ph.D.

Introduction

The purpose of this report is to provide detailed data on marijuana arrests that can expedite further study and analysis of local differences in arrest rates, trends, costs and criminal justice policies. The purpose of the commentary below is to provide a general context for reviewing the data as well as to explain the nature and composition of its presentation.

Marijuana arrests in the United States increased 109% from 1990 to 1998, according to the Uniform Crime Report (UCR) data compiled and released by the Federal Bureau of Investigation (FBI). In 1990 the National Household Survey (NHS) estimated that 18.9 million people used marijuana in the US on an annual basis. Despite some variance throughout the decade this estimate has remained somewhat consistent, and in 1998 the estimate of annual marijuana users was 18.7 million. Arrests have increased dramatically because law enforcement agencies have actively sought to increase marijuana arrests.

Arrests have also increased because national discussion of the marijuana laws has become narrowly focused on the issue of medical marijuana use. As discussion narrows to whether or not people who use marijuana medically should be arrested it is practically conceded by many that arrests for non-medical marijuana use are somehow justified. This report makes no such concession. Indeed a closer examination of marijuana arrest data begs the question: What justifies these arrests?

Commentary

One possible explanation for such a dramatic increase in arrests for marijuana offenses would be that new scientific findings had emerged proving, finally, that it was more dangerous than previously believed. Yet no such findings have emerged. Indeed one of the most respected professional association of pharmacologists in the United States issued a position paper in 1997 noting the contradiction existing laws and the contemporary scientific record:

"Over the years, blue-ribbon panels and government commissions have [been] consistent in recognizing that [marijuana] is one of the less harmful drugs of abuse. . . However, marijuana, hashish, and THC (the active substance in the hemp plant) are by no means harmless. . .

"Despite . . . significant adverse effects, questions have been raised by various investigative commissions about whether the social costs associated with the prohibition of marijuana are warranted by its actual harm to individuals and society, and especially whether imprisonment for mere possession unaccompanied by other crimes -- the law in some states -- is appropriate. It can be argued that placing marijuana in the same category as heroin and cocaine also sends a counterproductive message because it erases distinctions among drugs with very different degrees of hazard." Policy Statement on National Drug Policy[1], College on the Problems of Drug Dependence, March, 1997.

Under federal law marijuana is a schedule I drug. Under the Controlled Substances Act marijuana is assumed to have the same high potential for abuse as LSD and heroin. Despite marijuana's explicit or implicit decriminalization in numerous states, many law enforcement agencies justify their marijuana arrest policies by noting that the drug is a federal schedule I substance with a high potential for abuse. However, as indicated above, there are substantial scientific arguments that marijuana is not as dangerous as LSD and heroin. Thus, there is a wide gap between the basis for the marijuana laws and the contemporary scientific record.

It is the stated objective of national marijuana prevention programs to convince the general public that marijuana really is dangerous enough to justify its illegality. Often, the premise of these campaigns stress that the American public, especially today's parents, have been convinced that marijuana is harmless. National prevention campaigns employ government survey data in attempts to measure a state or region's perceptions of marijuana's "harmfulness." In states or regions of the country that have formally or informally decriminalized marijuana possession, citizens (and inherently local politicians and law enforcement) may have developed policing policies that reflect the public's perception that marijuana, while not harmless, does not warrant great governmental efforts and expense to maintain blanket marijuana prohibition. Differing perceptions of marijuana's harmfulness on a city or regional basis may well explain differences in local marijuana arrests.

Purpose of Report

Discussion and analysis of marijuana arrest figures is subject to the level of available data. Generally the most available data are the national totals produced by the Uniform Crime Report. This data provides for comparison with prior years and with other offenses, but little more. It does not enable one to compare arrests in different areas of the country or attempt to explain differences in arrests by way of variations in other data. Most importantly national data does not provide information on who bears the costs of marijuana arrests. Marijuana arrests have increased from nearly 327,000 in 1990 to nearly 683,000 in 1998. Where did this take place? Why do arrest figures vary dramatically from place to place? Which specific localities bear greater costs from state marijuana laws than others? How do local arrest trends compare to areas with similar populations nationwide? Where do the fewest marijuana arrests take place, and where do the most arrests take place? What local attributes can be associated with higher and lower marijuana arrest trends? Do differences in local marijuana laws explain differences in local arrest rates? Definitive answers to these questions are beyond the scope of this report, however, the purpose of this report is to provide the local level data necessary for individuals to discuss, debate, and research answers to these and other questions.

Uniform Crime Report Data

National estimates of marijuana arrests are derived from the annual report published by the Uniform Crime Report division of the FBI, Crime in the United States[2]. Not every arrest by every law enforcement agency is reported to the UCR database. The national estimates are based on reported data and estimations of missing data. National estimates of marijuana arrests are usually published as percentages of all drug arrests. These percentages are used along with the published estimate of all drug arrests to produce the national total for marijuana arrests.

County and state level data on marijuana arrests, or any other UCR offense category, are available from the Geospatial and Statistical Data Center[3] at the University of Virginia. This interactive service can provide arrest data for any requested locality in chart form or as downloadable data. While an excellent source of data on individual localities this does not provide a convenient basis for comparison of different regions in the country. This is the best alternative source for individuals to look up local arrest data for marijuana or any other criminal offense.

The complete UCR county level data set is available from the Inter-university Consortium for Political and Social Research[4](ICPSR) at the University of Michigan. The ICPSR data are difficult to work with because the information provided is unformatted.

However, this is the best source for projects involving statistical comparison or analysis of UCR data, and was the primary source for data for this report.

With minor exceptions the data provided in this report, while obtained from ICPSR, is the same data provided by the Geospatial and Statistical Data Center. County level data for Illinois and Florida are not included in the UCR data and were obtained independently. The original UCR data includes information on the degree of estimation involved in the reporting of a county level arrest total, as well as data that distinguishes zero from missing data. To simplify presentation this report provides data without regard to the degree of estimation. Also both zero values and missing data are generally indicated with the code "NR", meaning "None Reported". It is important to note that this report presents the best available data rather than an absolute count of every marijuana arrest in the country. Despite these statistical difficulties and other technical details this report provides a detailed and reliable basis for the comparison of marijuana arrest trends in different areas and regions. This report provides 1997 marijuana arrest data for 48 states; detailed data were not available for Kansas, Vermont, and the District of Columbia, and estimates were derived from the best available data for use in the national tables and maps. Otherwise 1997 data are provided for 2,951 of the nation's 3,140 counties, accounting for 95.5% of total estimated marijuana arrests for the year. (Data from Illinois and Florida were obtained directly from state law enforcement agencies; all other data is from the Uniform Crime Reports.)

Data Interpretation 

Where do the most marijuana arrests take place in the U.S.? Where are marijuana arrest rates the highest in the United States? Where do marijuana arrests consume a disproportionate share of local criminal justice resources? Where have marijuana arrest rates increased the most in the U.S.? This report provides the data necessary to answer these questions nationally and within each state in the country.

This report provides marijuana arrest data at the county level for a three-year period, from 1995 to 1997 (the most recent year's data available). Three indicators are provided for each county for each year data are available. These indicators are TL (total marijuana arrests), AQ (Arrest Quotient), and RATE (marijuana arrest rate). The arrest rate will be expressed in terms of arrests per 100,000 population. The percent change in arrest rates from 1995 to 1997 will also be provided in the tables. Counties will also be ranked by 1997 arrest rate.

The AQ indicator describes local marijuana arrests as a percentage of all local arrests, arrived by comparing the local percentage with the national percentage of marijuana arrests. This indicator is produced by dividing the local percentage by the national percentage; it is borrowed and adapted from regional economic analysis where it is used to identify local economic specialization. Here AQ will indicate areas where marijuana arrests are a greater percentage of all local arrests than the national standard. If marijuana arrests are 8% of all local arrests and 4% of all arrests nationally, the AQ value for the local area will be 2.0. This value indicates that the local percentage of marijuana arrests is twice the national average. In these data tables AQ values greater than 1.0 designate areas where marijuana arrests produce a greater burden for the local criminal justice system than the national standard. (Marijuana arrests were 4.09% of national arrests in 1995, 4.4% in 1996, and 4.65% in 1997. Accordingly in a region with an AQ of 2.0 in 1997 marijuana arrests would have been 9.3% of all local arrests. )

The total number of arrests for a county provides valuable information but can be misleading without consideration of the county's population. The rate of arrest compensates for population differences and allows for more accurate comparison of different regions. Significantly, the AQ value provides information on whether marijuana arrests are high, for example, because all arrests are high (indicated by an AQ value less than 1.0) or high because local law enforcement make a lot of marijuana arrests.

 Introduction to County and State Level Arrest Data

Hudspeth, Texas and Daggett, Utah had the highest 1997 marijuana arrest rates of all counties and reporting regions in the United States, 6,431 and 5,290 per 100,000 respectively. The national arrest rate was 256. Law enforcement in Hudspeth arrested 198 people for marijuana offenses with a local population of 3,079. Law enforcement in Daggett arrested 42 people with a local population of 794. The AQ values are 12.42 and 3.85 respectively, indicating that in both areas marijuana arrests place a considerably larger burden on the local criminal justice systems when compared to the national standard. The exceptionally high arrest rates in these locations suggest that many of the local arrests were not from the local population, but rather a high degree of arrests consisted of non-residents. The obvious question is what is it about these two particular places that result in so many marijuana arrests?

Hudspeth, Texas is on the US border with Mexico; it is just east of El Paso County along the Rio Grande river. Non-residents involved in smuggling marijuana from Mexico into the United States likely account for Hudspeth's exceptionally high arrest rate.

Daggett is in northeastern Utah. The county contains Ashley National Forest and the southern end of the Flaming Gorge National Recreation Area. Daggett County also lies between Dinosaur National Monument park in the next county to the east and Grand Teton and Yellowstone national parks in Wyoming to its north. As a central vacation location Daggett also attracts many non-residents, and this likely accounts for its exceptional marijuana arrest rate.

These examples suggest that regional characteristics may account for some differences in marijuana arrest rates. Marijuana arrest rates in the Southwest are the highest in the country. Texas has five counties in the top 10 and twelve counties in the top 25. Vacation centers are also likely places for high marijuana arrest rates. For example Cape May County, New Jersey, where a ferry crosses the Delaware Bay to connect the Jersey shore with the Eastern Shore of Delaware and Maryland, has the highest marijuana arrest rate in New Jersey (705, compared to a state arrest rate of 285). Worcester County, Maryland, also along the Eastern Shore, has the highest arrest rate in its state -- 2,267. The state arrest rate for Maryland for 1997, by comparison, is 314. In Cape May the AQ for 1997 is 1.15 indicating that the percentage of local marijuana arrests is only 15% higher than the national standard. In Worcester County, though, the AQ is 2.84, considerably lower than along the Rio Grande River but still nearly three times the national standard. In Worcester County marijuana arrests are a greater burden on the local criminal justice system than in Cape May County, but both counties have a greater concentration of marijuana arrests than the national standard.

Arrest data for the nation and for individual states is provided in alphabetical order and ranked by the 1997 arrest rate. These tables contain the complete UCR national estimate as well as the totals for the UCR data used for the county and state level presentations. The 1996 and 1997 national arrest rates from sources are essentially the same. In order to allow for comparison of areas with similar population levels additional tables present the highest 100 ranked counties in each of seven population ranges.

The table below provides the top ten counties with populations over 250,000, the largest and many of the most nationally known areas in the database. Fulton County, Georgia (Atlanta) had the highest marijuana arrest rate of this group in the country, 776 per 100,000. Douglas County, Nebraska (Omaha) was second (770) and Guilford County, North Carolina (Greensboro) was third at 697. The New York metropolitan area accounts for the 4th (New York), 6th (Bronx), 7th (Queens), 8th (Richmond) and 9th (Kings) ranked counties in this group nationally. The remaining counties in the top ten are East Baton Rouge, Louisiana (Baton Rouge) and Jefferson, Texas (Beaumont) on the Gulf Coast at the Texas/Louisiana border. All have rates over double the national standard of 256. Guilford County has a relatively low AQ of 1.07. With two exceptions the AQs of the remaining counties in this top ten indicate that marijuana arrests are about one third higher than the national rate. The exceptions are Douglas, Nebraska and Jefferson, Texas where arrests are twice the national standard. These ten counties account for nearly 9.5% of all the arrests reported in 1997. The top ten counties with populations over 250,000 account for nearly 38% of all the reported arrests.

From 1995 to 1997 Fulton County, Georgia increased the marijuana arrest rate from 508 to 776, an increase of 53%. This percentage increase ranked 17th out of the top 100 counties in this population size. The leading counties in terms of percentage increases in arrest rates were Summit, Ohio, including Akron, +170% from 88 to 238, El Paso, Colorado, including Colorado Springs, +165% from 82 to 218 and Clark, Nevada, including Las Vegas, +118% from 164 to 359. Also, Nevada (+61%), North Dakota (+58%), and North Carolina (+43%) led the states in terms of percent increase from 1995 to 1997.

The 1997 arrest rates of states that have decriminalized possession of small amounts of marijuana are listed in the table below. Ironically, the states with the four highest arrest rates in the nation also have decriminalized possession of small amounts of marijuana -- Alaska (418), New York (405), Nebraska (384) , and Mississippi (379). North Carolina was 12th with a rate of 310. The remaining 6 decrim states ranked 20th to 42nd of the 48 states reporting arrest data. "Decrim" states with AQs over 1 are New York, Nebraska, Mississippi, Maine, and Minnesota. Arrest rates rose from 19% to 43% in the period 1995 to 1997 in this group of states, but with two exceptions. The arrest rate in California, ranked 42nd in the country, only rose 11% from 1995 to 1997. In Oregon the arrest rate in 1997 was 2.5% lower than in 1995, ranking the state 38th in terms of change over this period. While it is true that marijuana arrests are substantially lower in six of the states where marijuana has been decriminalized, in the other five, interestingly, arrest rates are among the highest in the country.

Concluding Remarks

The purpose of this commentary has been to explain the source of the data presented in this report and provide a context for its analysis and for comparison of marijuana arrest trends in different areas. The hypotheses presented above are just that -- theories to be tested by further analysis. The primary objective of this report is to provide the necessary data to enable these and other explanations for differences in marijuana arrest rates to be derived, discussed, and tested with empirical data.

While total marijuana arrests appear to be leveling off they remain at the highest levels in US history, both in absolute numbers and in terms of arrest rates. The greater the level of arrests, the more important it has become for the government to justify the current policy and its economic and related social costs. It is clear from this data that marijuana arrest trends vary widely throughout the United States, and thus the cost of maintaining marijuana prohibition varies widely throughout the country as well.

Government agencies and anti-drug groups consistently insist that no one gets arrested for marijuana offenses any more. Unfortunately, The truth is just the opposite. More people in the United States are arrested for marijuana offenses today than ever before. Since 1965, 85% of the arrests for marijuana have been for possession only. Sadly, some areas of the country will be considerably proud of their status as an area with one of the highest marijuana arrest rates in the country. Marijuana consumers may want to employ this research in determining where they are educated, reside or travel. But marijuana arrests have doubled in the United States over the last decade despite considerable public opposition to the continued criminalization of marijuana's use and sale. Arrests have increased despite considerable scientific evidence challenging the decades old assumption that marijuana has a sufficiently high potential for abuse to justify its federal schedule I status. It is time for public discussion of marijuana laws to focus more closely on the costs and benefits of these historic arrest levels and whether or not they meet the standards for criminal sanctions in a just and free society.