It's tough to live in America today and not be scared to death. We are reminded constantly that we live in the most violent nation in the Western world. Our rates of robbery, rape, assault and homicide are far higher than those of other Western nations, according to the most reliable indicators, and they have remained so for a long time.
For more than 30 years, since the apparent beginnings of the sharp increase in the crime rate in the early 1960s, Americans have been fed a steady diet of crime news. Americans under 30 make up a generation raised on the idea that society is besieged by crime. We get media accounts of individual criminal events, so much so that the local nighttime TV news seems to be a continuous string of drug and crime stories. We get summary figures of national crime statistics, such as the Uniform Crime Reports and the National Crime Victimization Surveys, released quarterly and annually. And we have our own experiences and the word-of-mouth accounts of crime victimization that we hear from our friends and neighbors, which tend to have longer-lasting impact.
The result of living in a society where crime is so plentiful and news of crime--official and unofficial--so abundant is that Americans have a more heightened awareness of crime than do the people of other countries, and we let our perceptions of crime, especially our fear of violent crime, color our daily lives to a degree unmatched by the people of other nations. Even people who live in low-crime areas alter their lives out of fear of crime, sometimes even more than people who live in higher-crime areas.
Recent studies indicate that Americans continue to express high levels of frustration and anger with the persistence of high levels of crime in this country. Public skepticism of government efforts to control crime--as in the recent Violent Crime Control and Law Enforcement Act, popularly known as the "Clinton Crime Bill"--is very high. Many people have settled into a gloomy, fearful, "nothing works" mood: their perception is that crime is an inevitable rising tide, and there is nothing they (or anyone) can do to stop it.
This view of the crime rate, which is prevalent among ordinary people who have never studied crime statistics seriously, is summarized in two points:
1. The crime rate, especially the rate of violent crime, is at an all-time high.
2. This rate can be expected to continue to climb into the indefinite future.
It is not difficult to find support for these dismal views. "As Murder Toll Rises, Clearances Dip," the Law Enforcement News reports. "Murder Toll, Up 2.5%, Reached New High in 1991," the Criminal Justice Newsletter relates. A "Record Year for L.A. Gang Killings," the Juvenile Justice Digest proclaims, while Trial reminds us that "Murder is Leading Cause of Women's Deaths on the Job." We are told that we are not safe anywhere ("Three Shot to Death at Office Christmas Party," and "Texas Massacre: Year Later, Survivors Cope with Cafeteria Rampage"), and that anyone, at any age, is suspect: "Florida Reports Continuing Surge in Violent Crimes Committed by Juveniles."
These violent crimes headlines were gleaned not from newsstand tabloids but from professional publications circulating in the criminal justice field. If this is what professionals are reading, in our own journals, what can we expect members of the public to think, with their diet of hard news supplemented by the often-lurid features of popular magazines, tabloids and talk shows?
How many people would we expect to encounter, out running loose in American society, who would say that they think the crime rate is going down, or even under control, or that they feel safer today than they did a few years ago? Anyone who makes optimistic noises about crime control is looked upon as being at best out of touch with reality, or at worst as some kind of apologist loony. Could any sensible person find anything comforting in recent crime trends?
When we were doing research for an unrelated article on capital punishment recently, we collected statistics on murder rates in the 50 states over the 20-year period from 1973 to 1992. We began to notice, in a cursory look at the homicide rates in the individual states, that the murder trends appeared to be very uncertain. Rather than up-up-up, the pattern might be up-down-up, or down-up-down, or up-down-up-down, or maybe some other pattern, depending on which state's statistics you happened to be looking at. You might end up with what you could call a national trend, by putting all the 50 states together, but that trend would not reliably reflect the trends you could identify in the states taken one at a time.
What we undertook here was a comparison of the murder rates in the 50 states over the period from 1973 to 1992. What has happened with the murder rate, not nationally, but in the states? Should everyone be more-or-less equally scared to death, or should some be much more scared than others? In which states are people entitled to be most fearful?
Our purpose was to see if the murder rate is getting worse everywhere, and if so, how much worse it is now that it was 20 years ago. We also wanted to see what trends a state-to-state comparison would reveal, and what commonalities might be identified among states with high or low murder rates, or states with rates that went up or down markedly over the two decades.
The statistics we compiled are taken from the Uniform Crime Reports. We took the FBI's estimates of homicide rates (which includes all degrees of murder and non-negligent manslaughter per 100,000 population) for each year from 1973 to 1992. The UCR figures are subject to the usual criticisms--homicides miscounted, population estimates in error--but they are at least consistent over time. Murder remains the crime most likely to be reported accurately, and the methodology of counting has not changed.
Table 1 is our basic statistical summary. For each state, it shows the homicide rate (per 100,000) at the beginning (1973) and end (1992) of the 20-year period, and the change over time. It shows the highest year and rate, and the lowest year and rate. A final column shows total homicides in each state, the mean rate over 20 years, and the state's rank among the 50 states over the study period.
We would like to make three general observations, in considering the statistics in Table 1, before we move on to other graphics:
1. The murder rate varies widely from one state to another, and from one part of the country to another.
2. The murder rate fluctuates sharply from one year to the next within each state.
3. In only one-third of the states did the murder rate increase over the 20-year study period; in two-thirds it decreased or remained constant.
What is most noticeable, about the murder rate in America, is how much it varies from one place to another. For 20 years, the murder rates in most of the states of the deep South and along the south Atlantic coast, plus the highly-urbanized states of California, Illinois and New York, have been on the average three to five times higher, and sometimes ten or more times higher, than the murder rates of several Northern and Midwestern states.
If you lived in Louisiana in 1992, for instance, you would be about ten times as likely to be murdered as you would be if you lived in North or South Dakota, New Hampshire, Maine or Iowa. So, if reducing your vulnerability as a potential murder victim was of utmost importance to you, the most fundamental move you could make would be north. When it comes to avoiding murder, "Go north" is excellent advice--unless you're heading for New York City, Detroit or Chicago. Perhaps we should have said, "Go northwest."
If the murder rate climbs too high, and you can't move to a colder climate, the best advice is: "Wait til next year." Looking at the detailed year-to-year statistics, which are not included here, murder rates often dropped sharply from one year to the next, and increased again the same way. There tended to be not so much gentle curves of change as jagged peaks and valleys, and one particular year, whether it be high or low, might mean nothing at all in its statistical implications.
If you look at the states that tended to have higher murder rates--exactly half the states had rates that climbed above 10.0 some time during the 20 years--in these states the lowest rate tended to be half or greater than the highest rate (20 of 25 states). In those states with peak rates below 10.0, the fluctuation tended to be greater; in 12 of these 25 states, the lowest rate was less than half the highest rate. So on the whole, the greater annual fluctuations tended to be in the states that had lower overall homicide rates. The best example would be Vermont, which hit its peak rate of 5.5 in 1976, and then immediately followed with its lowest rate of 1.4 in 1977 and again in 1979.
Homicide Rate (per 100,000) Highest Lowest 20-Year Totals
State 1973 1992 Change Year Rate Year Rate Homicides Mean Rate Rank
Alabama 13.2 11.0 -2.2 1975 16.0 1983 9.2 9,187 11.89 7
Alaska 10.0 7.5 -2.5 1982 18.5 1988 5.7 967 10.85 11
Arizona 8.1 8.1 None 1980 10.3 1989 6.7 4,823 8.36 22
Arkansas 8.8 10.8 +2.0 1974 11.2 1984 7.5 4,120 9.08 19
California 9.0 12.7 +3.7 1980 14.5 1973 9.0 57,202 11.31 9
Colorado 7.9 6.2 -1.7 1981 8.1 1990 4.2 3,732 6.29 26
Connecticut 3.3 5.1 +1.8 1989 5.9 1976 3.1 2,848 4.50 36
Delaware 5.9 4.6 -1.3 1974 10.3 1983 4.1 705 5.77 30
Florida 15.4 9.0 -6.4 1973 15.4 1984/1992 9.0 24,771 11.97 6
Georgia 17.4 11.0 -6.4 1974 17.8 1983 8.4 14,717 13.07 3
Hawaii 5.3 3.6 -1.7 1980 8.7 1982 3.1 1,053 5.41 32
Idaho 2.6 3.5 +0.9 1974 5.6 1991 1.8 690 3.74 39
Illinois 10.4 11.4 +1.0 1974 11.8 1985 8.0 22,565 9.90 16
Indiana 7.2 8.2 +1.0 1980 8.9 1983 5.2 7,518 6.90 24
Iowa 2.2 1.6 -0.6 1981 2.6 1992 1.6 1,215 2.11 49
Kansas 6.0 6.0 None 1974/1980 6.9 1988 3.4 2,578 5.38 33
Kentucky 9.7 5.8 -3.9 1976 10.6 1992 5.8 6,008 8.39 21
Louisiana 15.4 17.4 +2.0 1992 17.4 1985 10.9 12,234 14.63 1
Maine 2.1 1.7 -0.4 1981/1989 3.2 1991 1.2 557 2.44 47
Maryland 11.3 12.1 +0.8 1992 12.1 1985 7.9 8,675 9.89 17
Massachusetts 4.4 3.6 -0.8 1973/1974 4.4 1987 3.0 4,380 3.76 38
Michigan 12.1 9.9 -2.2 1973 13.0 1979/1982 9.1 19,552 10.64 12
Minnesota 2.7 3.3 +0.6 1975/1992 3.3 1983 1.7 2,093 2.52 45
Mississippi 16.1 12.2 -3.9 1973 16.1 1988 8.6 6,101 12.22 5
Missouri 9.0 10.5 +1.5 1979 11.2 1984 7.1 9,308 9.38 18
Montana 6.0 2.9 -3.1 1973 6.0 1988/1991 2.6 651 3.85 37
Nebraska 4.3 4.2 -0.1 1980 4.4 1982 2.0 1,064 3.37 41
Nevada 12.2 10.9 -1.3 1980 20.0 1989 8.2 2,181 12.87 4
New Hampshire 2.1 1.6 -0.5 1991 3.6 1984 1.0 472 2.47 46
New Jersey 7.4 5.1 -2.3 1973 7.4 1987 4.6 8,739 5.85 29
New Mexico 11.4 8.9 -2.5 1975 13.3 1989 8.6 2,863 10.62 13
New York 11.1 13.2 +2.1 1990 14.5 1985 9.5 41,563 11.63 8
North Carolina 13.0 10.6 -2.4 1973 13.0 1988 7.8 11,969 9.99 15
North Dakota 0.8 1.9 +1.1 1981 2.3 1989 0.6 166 1.26 50
Ohio 7.3 6.6 -0.7 1974 8.9 1984 5.1 14,525 6.74 25
Oklahoma 6.6 6.5 -0.1 1982 10.8 1976 6.4 4,939 8.07 23
Oregon 4.9 4.7 -0.2 1986 6.6 1990 3.8 2,554 4.92 34
Pennsylvania 6.3 6.2 -0.1 1975/1980 6.8 1984 4.5 14,077 5.92 27
Rhode Island 3.4 3.6 +0.2 1989 4.9 1976 2.4 711 3.68 40
South Carolina 14.4 10.4 -4.0 1974 16.2 1986 8.6 7,011 11.15 10
South Dakota 3.8 0.6 -3.2 1986 4.0 1992 0.6 297 2.13 48
Tennessee 13.2 10.4 -2.8 1974 13.4 1984/1989 8.4 9,336 10.20 14
Texas 12.7 12.7 None 1980 16.9 1987 11.7 41,414 13.87 2
Utah 3.2 3.0 -0.2 1979 4.8 1989 2.6 989 3.31 42
Vermont 2.2 2.1 -0.1 1976 5.5 1977/1979 1.4 271 2.63 44
Virginia 8.5 8.8 +0.3 1975 11.5 1983 7.0 9,278 8.40 20
Washington 4.0 5.0 +1.0 1975/1988 5.7 1973 4.0 4,120 4.87 35
West Virginia 5.7 6.3 +0.6 1975 7.4 1985 3.8 2,194 5.85 28
Wisconsin 2.6 4.4 +1.8 1991 4.8 1978/1984 2.5 3,102 3.26 43
Wyoming 6.8 3.6 -3.2 1975 10.2 1987 2.0 495 5.54 31
What made for these differences? Was the culture of Vermont, or its emergency medical care, or the availability of weapons, or the level of interpersonal violence, or anything else, subject to such remarkable change in the latter part of the 1970s to account for these fluctuations? These are the kinds of changes that drive law enforcement officials crazy, because there is no apparent pattern to be discerned, and no sensible explanation that can be provided the public in good faith.
The most important point the beleaguered officials of Vermont could make was the same point officials in 29 other states could also make: over the 20 years from 1973 to 1992, the murder rate went down, not up. In three states it was constant. In 17 states, led by California, New York, Louisiana and Arkansas, it went up.
Objectively, then, in only one of every three states was an individual citizen more likely to be a murder victim in 1992 than in 1973. He may not feel safer from murder, but in point of fact he is. And in Vermont he is about as safe as any place in the country.
Table 2 is a different way of arranging the data in Table 1. It ranks the 50 states from highest to lowest mean (or average) homicide rates, based on 20-year averages. At the top are Louisiana, Texas and Georgia; at the bottom North Dakota, Iowa and South Dakota. This table also shows the minimum or low rates and the maximum or high rates, and the range between them.
We wanted to make two points about Table 2 before moving on. First, seven of the ten states with the highest murder rates are in the South--Louisiana, Texas, Georgia, Mississippi, Florida, Alabama and South Carolina. Three non-Southern states--Nevada, California and New York--are among the ten highest. Nothing has changed much to alter the perception of Southern violence that has existed in criminological literature since Sheldon Hackney's "Southern Violence" was published in the 1960s. (See also Fred Hawley and Steven Messner's "The Southern Violence Construct," Justice Quarterly, 1989.)
The second point is the sharp decline in murder rates in two states, Nevada and Alaska, in the 1980s. Both of these lightly-populated states had very high murder rates in the early 1980s which declined in the latter part of the decade.
Nevada had the all-time record-breaking rate of 20 murders per 100,000 in 1980, and a low of 8.2 homicides in 1989. Could such changes as the proliferation of gambling into other states, and the changing nature of the Nevada tourist trade, be important here?
Alaska reported a high rate of 18.5 murders per 100,000 in 1982, and a low rate of 5.7 in 1988. Was this change mostly fortuitous, or might it be related to the economic impact of the downturn in the oil industry that began in the mid-1980s?
Table 3 groups the states into the nine Uniform Crime Reporting regions. You can compare each state with others in its region, and you can compare one region with another. In a longer version of this article, we created a series of graphs, one for each region, to show the year-by-year homicide rates for each state in the region. We have included the graph for the Southern States
--West South Central region, which includes Louisiana, to illustrate what these graphs look like. This graph is Chart 1. We also included, as Chart 2, an aggregate or national graph showing how the rates in the nine regions changed over 20 years, and in general how the regional rates compare with each other.
The general pattern that seems to emerge, in looking at the charts, is this: murder rates peaked in the period from 1973 to 1980 (this was true in 34 of the 50 states), dropped for several years to a low during the period from 1983 to 1990 (true for 35 states), then begin to climb again, but not by 1992 quite back to the point they were in 1973. This would lead to a general description of the 20-year murder rate as "up-down-up," but with the ending point generally lower on the scale than the starting point.
Eight of the nine UCR regions had at least one state showing murder rate increases from 1973 to 1992:
I. Connecticut, Rhode Island (Northeastern States--New England Division)
II. New York (Northeastern States--Middle Atlantic Division)
III. Illinois, Indiana, Wisconsin (Midwestern States--East North Central Division)
IV. Minnesota, Missouri, North Dakota (Midwestern States--West North Central Division)
V. Maryland, Virginia, West Virginia (Southern States--South Atlantic Division)
VI. Arkansas, Louisiana (Southern States--West South Central Division)
VII. Idaho (Western States--Mountain Division)
VIII. California, Washington (Western States--Pacific Division).
The two regions which had at least half of their states experience an increase in the murder rate were:
I. Midwestern States--East North Central Division, which is made up of the states of Illinois, Indiana, Michigan, Ohio and Wisconsin. Three of five states experienced an increase.
II. Southern States--West South Central Division, which is made up of Arkansas, Louisiana, Oklahoma and Texas. Arkansas and Louisiana reported increases.
The region that had the highest percentage of states with reductions in the murder rate was the Southern States--East South Central Division, made up of Alabama, Kentucky, Mississippi and Tennessee. All reported fairly significant reductions over the experimental period: Kentucky from 10 to 6; Tennessee from 13 to 10; Alabama from 13 to 11; and Mississippi from 16 to 12. This region seemed to become a safer area to live in, in terms of murder, while the region immediately adjacent to it, the West South Central, appeared to become a more dangerous place to live.
When all of the rates from the regions are summed and plotted over time (Chart 2), it is clear that on a regional basis, a majority of the totals in 1992 are close to the 1973 rates. The one area that had a noticeable increase was--you guessed it-- the Southern States--West South Central Division, which includes Louisiana.
Why the higher level of murders in the South? Is the explanation economic, cultural, or a combination of these and other circumstances? What similarities do the seven Southern states in the top ten share with Nevada, New York and California?
The debate on the high levels of southern violence continues. Some recent evidence, using an independent measure of cultural support for violence which is conceptually independent from the southern region, seems to point in another direction.
Early research relied on "region" as being equivalent to "cultural support for violence," thus leaping to the conclusion that living in a certain state translated into belonging to a subculture which supports the use of violence.
The Legitimate Violence Index, first developed by Larry Baron and Murray Strauss in 1988, measured cultural support for violence through the application of 12 indicators of noncriminal violence, such as watching television violence and using corporal punishment in schools. Combining these variables with poverty, homicide rates, and other factors, the results showed:
This hypothesis (States in the southern region are more likely to have a higher level of support for legitimate violence than states in other regions of the country) was not confirmed as the western region showed the greatest magnitude of support for legitimate violence, although the difference was not statistically significant. The finding casts doubt on the argument that the South represents a distinctive culture of violence. It also raises questions about the validity of using southern region as an indicator of cultural support for violence.
On the issue of an economic explanation of homicide rates, the authors conclude, based on a statistical analysis of the Confederate South, legitimate violence, poverty, economic inequality, and a number of control variables, that:
"More urbanized states and states with greater poverty and economic inequality also tend to have higher rates of homicide. Thus, cultural and economic factors are not at odds in the explanation of homicide, rather each accounts for a significant proportion of state to state variation."
When we used a statistical technique called cluster analysis on the 20-year data set, we found a grouping of states with similar patterns to the Baron and Strauss study using the Legitimate Violence Index. One of the clusters included the states of Nevada, Texas, Louisiana, New York, California, Georgia, Florida, Mississippi, South Carolina and Alabama, which would seem to support the earlier findings. Could these states, individually or collectively, be destined to be leaders in high murder rates, regardless of resources allocated to dealing with the violent crime problem? Is the violent crime rate, specifically the murder rate, an intractable or impossible-to-solve problem?
We can find reason to hope in the experiences of Nevada and Alaska. Both experienced short-term, significant and apparently lasting declines from their peak murder rates.
On the other hand, we have the experiences of California, New York, and Louisiana, which are all top ten murder states where the rate was increasing at the end of the study period. What commonalities might we find among these states that would explain these increases?
One suggestion is that each state has one or more urban metropolises with high murder rates, which tends to heavily weight the entire state's homicide rate. Recent research by Lois Fingerhut and others has established definite relationships between urbanization, age, drugs, race and firearms use in criminal violence. These patterns appear to be important also in several of the other top ten murder states --Texas, Georgia and Florida--where murder rates declined or remained stable during the study period but remained at an overall high rate. This is an important direction for future research.
It may be, in returning to our original question about whether people's fears for their safety are justified, that the best we can do is respond with a qualified "maybe not." It does appear that, for whatever reasons, the murder rate in most states was lower as the 1990s began than it was 20 years earlier. The down side of this bit of encouragement is that the rate, even in the Northern states, is still much higher than it would be in any other Western nation. If you live in the South, or in one of America's largest cities, your prospects of being murdered remain much higher than the national average of 1 chance in 10,000.
The murder rate in America continues at a level high enough that most people aren't really aware (or don't really believe) that for most of them the rate has stabilized or gone down over the last two decades. In this environment, fear takes on a life of its own, and it is not easily killed with statistics. The murder rate must not only go down for people to feel safer, but people must also act as if they believe that it has gone down.
The flattening out of the murder rate over the last 20 years is not enough of a change to convince anyone that we have rounded a corner, that violent crime is under control. As skeptical as the American public has become, the change would have to be much more profound and prolonged to bring about a reversal of public opinion that has been a generation in the making.