community to reduce the violence. Sometimes the only remedy prescribed is mobilization of the largest possible police or control force as early as possible.

None of these views is accurate. We found that:

A variety of different control forces employed a variety of tactics, often at the same time, and often in a confused situation;

■Substantial non-force control efforts, such as negotiations and on-the-street persuasion by "counterrioters," were usually underway, often simultaneously with forcible control efforts; counterrioter activity often was carried on by Negro residents of the disturbance area, sometimes with and frequently without official recognition;

No single tactic appeared to be effective in containing or reducing violence in all situations.

Local official force

In 20 of the 24 disorders, the primary effort to restore order at the beginning of violence was made entirely by local police. 68 In 10 cases no additional outside force was called for after the initial response.69 In only a few cases was the initial control force faced with crowds too large to control.70

The police approach to the initial outbreak of disorder in the surveyed cities was generally cautious.71 Three types of response were employed. One was dispersal (clearing the area, either by arrests or by scattering crowds), used in 10 cases." 72 Another was reconnaissance (observing and evaluating developments), used in eight cases.73 In half of these instances, they soon withdrew from the disturbance area, generally because they believed they were unable to cope with the disorder. The third was containment (preventing movement in or out of a cordoned or barricaded area), used in six cases.75

74

No uniform result from utilizing any of the three control approaches is apparent. In at least half of the 24 cases, it can reasonably be said that the approach taken by the police failed to prevent the continuation of violence.76 To the extent that their effectiveness is measurable, the conclusion appears to hold for subsequent police control responses as well." There is also evidence in some instances of over-response in subsequent cycles of violence.78

The various tactical responses we have described are not mutually exclusive, and in many instances combinations were employed. The most common were attempts at dispersal in the disturbance area and a simultaneous cordon or barricade at the routes leading from the disturbance area to the central commercial area of the city, either to contain the disturbance or to prevent persons outside the area from entering it, or both.79

In 11 disorders a curfew was imposed at some time, either as the major dispersal technique or in combination with other techniques.80

In only four disorders was tear gas used at any point as a dispersal technique.81

Only Newark and New Haven used a combination of all three means of control-cordon, curfew, and tear

gas.82

Supplemental Official Force

In nine disturbances-involving a wide variation in the intensity of violence-additional control forces were brought in after there had been serious violence which local police had been unable to handle alone.83 In every case, further violence occurred, often more than once and often of equal or greater intensity than before.84

The result was the same when extra forces were mobilized before serious violence. In four cities where this was done,85 violence nonetheless occurred, in most cases more than once,86 and often of equal or greater intensity than in the original outbreak.87

In the remaining group of seven cities, no outside control forces were called,88 because the level and duration of violence were lower. Outbreaks in these cities nevertheless followed the same random pattern as in the cities which used outside forces.89 Negotiation

In 21 of the 24 disturbances surveyed, discussion or negotiation occurred during the disturbance. These took the form of relatively formal meetings between government officials and Negroes during which grievances and issues were discussed and means were sought to restore order.90

90

Such meetings were usually held either immediately before or soon after the outbreak of violence.91 Meetings often continued beyond the first or second day of the disorder and, in a few instances, through the entire period of the disorder.92

The Negro participants in these meetings usually were established leaders in the Negro community, such as city councilmen or members of human relations commissions, ministers, or officers of civil rights or other community organizations.93 However, Negro youths participated in over one-third of these meetings. In a few disorders both youths and adult Negro leaders participated,95 sometimes without the participation of local officials.96

94

Employees of community action agencies occasionally participated, either as intermediaries or as participants. In some cases they provided the meeting place.97

Discussions usually included issues generated by the disorder itself, such as the treatment by the police of those arrested.98 In 12 cases, prior ghetto grievances, such as unemployment and inadequate recreational facilities, were included as subjects.99 Often both disorder-related and prior grievances were discussed,100 with the focus generally shifting from the former to the latter as the disorder continued.

How effective these meetings were is, as in the

It is sometimes assumed that the rioters were criminal types, overactive social deviants, or riffraff-recent migrants, members of an uneducated underclass, alienated from responsible Negroes, and without broad social or political concerns. It is often implied that there was no effort within the Negro community to attempt to reduce the violence.

We have obtained data on participation from four different sources: 111

Eyewitness accounts from more than 1,200 interviews in our staff reconnaissance survey of 20 cities;

■ Interview surveys based on probability samples of riot area residents in the two major riot cities-Detroit and Newarkdesigned to elicit anonymous self-identification of participants as rioters, counterrioters or noninvolved;

■ Arrest records from 22 cities; and

■ A special study of arrestees in Detroit.

Only partial information is available on the total numbers of participants. In the Detroit survey, approximately 11 percent of the sampled residents over the age of 15 in the two disturbance areas admittedly participated in rioting; another 20 to 25 percent admitted to having been bystanders but claimed that they had not participated; approximately 16 percent claimed they had engaged in counterriot activity; and the largest proportion (48 to 53 percent) claimed they were at home or elsewhere and did not participate. However, a large proportion of the Negro community apparently believed that more was gained than lost. through rioting, according to the Newark and Detroit surveys.112

Greater precision is possible in describing the characteristics of those who participated. We have com

THE PROFILE OF A RIOTER

The typical rioter in the summer of 1967 was a Negro, unmarried male between the ages of 15 and 24. He was in many ways very different from the stereotype. He was not a migrant. He was born in the state and was a lifelong resident of the city in which the riot took place. Economically his position was about the same as his Negro neighbors who did not actively participate in the riot.

Although he had not, usually, graduated from high school, he was somewhat better educated than the average inner-city Negro, having at least attended high school for a time.

Nevertheless, he was more likely to be working in a menial or low status job as an unskilled laborer. If he was employed, he was not working full time and his employment was frequently interrupted by periods. of unemployment.

He feels strongly that he deserves a better job and that he is barred from achieving it, not because of lack of training, ability, or ambition, but because of discrimination by employers.

He rejects the white bigot's stereotype of the Negro as ignorant and shiftless. He takes great pride in his race and believes that in some respects Negroes are superior to whites. He is extremely hostile to whites, but his hostility is more apt to be a product of social and economic class than of race; he is almost equally hostile toward middle class Negroes.

He is substantially better informed about politics than Negroes who were not involved in the riots. He is more likely to be actively engaged in civil rights efforts, but is extremely distrustful of the political system and of political leaders.

THE PROFILE OF THE COUNTERRIOTER

The typical counterrioter, who risked injury and arrest to walk the streets urging rioters to "cool it," was an active supporter of existing social institutions. He was, for example, far more likely than either the rioter or the noninvolved to feel that this country is worth defending in a major war. His actions and his attitudes reflected his substantially greater stake in the social system; he was considerably better educated and more affluent than either the rioter or the noninvolved. He was somewhat more likely than the rioter, but less likely than the noninvolved, to have been a migrant. In all other respects he was identical to the noninvolved.113

CHARACTERISTICS OF PARTICIPANTS

Race

Of the arrestees 83 percent were Negroes; 15 percent were whites.114 Our interviews in 20 cities indicate that almost all rioters were Negroes.

Age

The survey data from Detroit, the arrest records, and our interviews in 20 cities, all indicate that the rioters were late teenagers or young adults.115 In the Detroit survey, 61.3 percent of the self-reported rioters were between the ages of 15 and 24, and 86.3 percent were between 15 and 35. The arrest data indicate that 52.5 percent of the arrestees were between 15 and 24, and 80.8 percent were between 15 and 35.

Of the noninvolved, by contrast, only 22.6 percent in the Detroit survey were between 15 and 24, and 38.3 percent were between 15 and 35.

Sex

In the Detroit survey, 61.4 percent of the self-reported rioters were male. Arrestees, however, were almost all male-89.3 percent.116 Our interviews in 20 cities indicate that the majority of rioters were male. The large difference in proportion between the Detroit. survey data and the arrestee figures probably reflects either selectivity in the arrest process or less dramatic, less provocative riot behavior by women. Family Structure

Three sources of available information-the Newark survey, the Detroit arrest study, and arrest records from four cities-indicate a tendency for rioters to be single.11 The Newark survey indicates that rioters were

single-56.2 percent-more often than the noninvolved-49.6 percent.

The Newark survey also indicates that rioters were more likely to have been divorced or separated-14.2 percent than the noninvolved-6.4 percent. However, the arrest records from four cities indicate that only a very small percentage of those arrested fall into this category.

In regard to the structure of the family in which he was raised, the self-reported rioter, according to the Newark survey, was not significantly different from many of his Negro neighbors who did not actively participate in the riot. Twenty-five and five-tenths percent of the self-reported rioters and 23 percent of the noninvolved were brought up in homes where no adult male lived.118

Region of Upbringing

Both survey data 119 and arrest records 120 demonstrate unequivocally that those brought up in the region in which the riot occurred are much more likely to have participated in the riots. The percentage of self-reported rioters brought up in the North is almost identical for the Detroit survey-74.4 percent-and the Newark survey-74 percent. By contrast, of the noninvolved, 36 percent in Detroit and 52.4 percent in Newark were brought up in the region in which the disorder occurred.121

Data available from five cities on the birthplace of arrestees indicate that 63 percent of the arrestees were born in the North. Although birthplace is not necessarily identical with place of upbringing, the data are sufficiently similar to provide strong support for the conclusion.

Of the self-reported counterrioters, however, 47.5 percent were born in the North, according to the Detroit survey, a figure which places them between selfreported rioters and the noninvolved. Apparently, a significant consequence of growing up in the South is the tendency toward noninvolvement in a riot situation, while involvement in a riot, either in support of or against existing social institutions, was more common among those born in the North. Residence

Rioters are not only more likely than the noninvolved to have been born in the region in which the riot occurred, but they are also more likely to have been long-term residents of the city in which the disturbance took place. 122 The Detroit survey data indicate that 59.4 percent of the self-reported rioters, but only 34.6 percent of the noninvolved, were born in Detroit. The comparable figures in the Newark survey are 53.5 percent and 22.5 percent.

Outsiders who temporarily entered the city during the riot might have left before the surveys were conducted and therefore may be underestimated in the

Looting in Milwaukee, August 1967

survey data. However, the arrest data,123 which is contemporaneous with the riot, suggest that few outsiders were involved: 90 percent of those arrested resided in the riot city, 7 percent lived in the same state, and only 1 percent were from outside the state. Our interviews in 20 cities corroborate these conclusions. Income

In the Detroit and Newark survey data, income level alone does not seem to correlate with self-reported riot participation.124 The figures from the two cities are not directly comparable since respondents were asked for individual income in Detroit and family income in Newark. More Detroit self-reported rioters (38.6 percent) had annual incomes under $5,000 per year than the noninvolved (30.3 percent), but even this small. difference disappears when the factor of age is taken into account.

In the Newark data, in which the age distributions of self-reported rioters and the noninvolved are more similar, there is almost no difference between the rioters, 32.6 percent of whom had annual incomes under $5,000, and the noninvolved, 29.4 percent of whom had annual incomes under $5,000.

The similarity in income distribution should not, however, lead to the conclusion that more affluent Negroes are as likely to riot as poor Negroes. Both surveys were conducted in disturbance areas where incomes are considerably lower than in the city as a whole and the surrounding metropolitan area. 125 Nevertheless, the data show that rioters are not necessarily the poorest of the poor.

While income fails to distinguish self-reported rioters from those who were not involved, it does distinguish counterrioters from rioters and the noninvolved. Less than 9 percent of both those who rioted and those not

involved earned more than $10,000 annually. Yet almost 20 percent of the counterrioters earned this amount or more. In fact, there were no male self-reported counterrioters in the Detroit survey who earned less than $5,000 annually. In the Newark sample there were seven respondents who owned their own homes; none of them participated in the riot. While extreme poverty does not necessarily move a man to riot, relative affluence seems at least to inhibit him from attacking the existing social order and may motivate him to take considerable risks to protect it.

Education

Level of schooling is strongly related to participation. Those with some high school education were more likely to riot than those who had only finished grade school.126 In the Detroit survey, 93 percent of the selfreported rioters had gone beyond grade school, compared with 72.1 percent of the noninvolved. In the Newark survey the comparable figures are 98.1 and 85.7 percent. The majority of self-reported rioters were not, however, high school graduates.

The counterrioters were clearly the best educated of the three groups. Approximately twice as many counterrioters had attended college as had the noninvolved, and half again as many counterrioters had attended college as rioters. Considered with the information on income, the data suggest that counterrioters were probably well on their way into the middle class.

Education and income are the only factors which distinguish the counterrioter from the noninvolved. Apparently, a high level of education and income not only prevents rioting but is more likely to lead to active, responsible opposition to rioting.

Employment

The Detroit and Newark surveys, the arrest records from four cities, and the Detroit arrest study all indicate that there are no substantial differences in unemployment between the rioters and the noninvolved.127

Unemployment levels among both groups were extremely high. In the Detroit survey, 29.6 percent of the self-reported rioters were unemployed; in the Newark survey, 29.7 percent; in the four-city arrest data, 33.2 percent; and in the Detroit arrest study, 21.8 percent. The unemployment rates for the noninvolved in the Detroit and Newark surveys were 31.5 and 19.0 percent.

Self-reported rioters were more likely to be only intermittently employed, however, than the noninvolved. Respondents in Newark were asked whether they had been unemployed for as long as a month or more during the last year.128 Sixty-one percent of the self-reported rioters, but only 43.4 percent of the noninvolved, answered, "yes."

Despite generally higher levels of education, rioters were more likely than the noninvolved to be employed in unskilled jobs. 129 In the Newark survey, 50 percent of the self-reported rioters, but only 39.6 percent of the noninvolved, had unskilled jobs.

Attitudes About Employment

The Newark survey data indicate that self-reported rioters were more likely to feel dissatisfied with their present jobs than were the noninvolved. 130

Only 29.3 percent of the rioters, compared with 44.4 percent of the noninvolved, thought their present jobs appropriate for them in responsibility and pay. Of the self-reported rioters, 67.6 percent, compared with 56.1 percent of the noninvolved, felt that it was impossible to obtain the kind of job they wanted.131 Of the selfreported rioters, 69 percent, as compared with 50 percent of the noninvolved, felt that racial discrimination was the major obstacle to finding better employment.132 Despite this feeling, surprising numbers of rioters (76.9 percent) responded that "getting what you want out of life is a matter of ability, not being in the right place at the right time." 133

Racial Attitudes

The Detroit and Newark surveys indicate that rioters have strong feelings of racial pride, if not racial superiority.134 In the Detroit survey, 48.6 percent of the self-reported rioters said that they felt Negroes were more dependable than whites. Only 22.4 percent of the noninvolved stated this. In Newark, the comparable figures were 45 and 27.8 percent. The Newark survey data indicate that rioters wanted to be called "black" rather than "Negro" or "colored" and were somewhat more likely than the noninvolved to feel that all Negroes should study African history and languages.135

To what extent this racial pride antedated the riot or was produced by the riot is impossible to determine from the survey data. Certainly the riot experience seems to have been associated with increased pride in the minds of many participants. This was vividly illustrated by the statement of a Detroit rioter:

Interviewer: You said you were feeling good when you followed the crowds?

Respondent: I was feeling proud, man, at the fact that I was a Negro. I felt like I was a first-class citizen. I didn't feel ashamed of my race because of what they did.

Similar feelings were expressed by an 18-year-old Detroit girl who reported that she had been a looter: Interviewer: What is the Negro then if he's not American? Respondent: A Negro, he's considered a slave to the white

folks. But half of them know that they're slaves and feel that they can't do nothing about it because they're just going along with it. But most of them they seem to get it in their heads now how the white folks treat them and how they've been treating them and how they've been slaves for the white folks. *

Along with increased racial pride there appears to be intense hostility toward whites. 136 Self-reported rioters in both the Detroit and Newark surveys were more likely to feel that civil rights groups with white and Negro leaders would do better without the whites. In Detroit, 36.1 percent of the self-reported rioters thought that this statement was true, while only 21.1 percent of the noninvolved thought so. In the Newark survey, 51.4 percent of the self-reported rioters agreed; 33.1 percent of the noninvolved shared this opinion.

Self-reported rioters in Newark were also more likely to agree with the statement, "Sometimes I hate white people." Of the self-reported rioters, 72.4 percent agreed; of the noninvolved, 50 percent agreed.

The intensity of the self-reported rioters' racial feelings may suggest that the recent riots represented traditional interracial hostilities. Two sources of data suggest that this interpretation is probably incorrect.

First, the Newark survey data indicate that rioters were almost as hostile to middle-class Negroes as they were to whites.137 Seventy-one and four-tenths percent of the self-reported rioters, but only 59.5 percent of the noninvolved, agreed with the statement, "Negroes who make a lot of money like to think they are better than other Negroes." Perhaps even more significant, particularly in light of the rioters' strong feelings of racial pride, is that 50.5 percent of the self-reported rioters agreed that "Negroes who make a lot of money are just as bad as white people." Only 35.2 percent of the noninvolved shared this opinion.

Second, the arrest data show that the great majority of those arrested during the disorders were generally charged with a crime relating to looting or curfew violations.138 Only 2.4 percent of the arrests were for assault and 0.1 percent were for homicide, but 31.3 percent of the arrests were for breaking and enteringcrimes directed against white property rather than against individual whites.

Political Attitudes and Involvement

Respondents in the Newark survey were asked about relatively simple items of political information, such as the race of prominent local and national political figures. In general, the self-reported rioters were much better informed than the noninvolved.139 For example, self-reported rioters were more likely to know that one of the 1966 Newark mayoral candidates was a Negro. Of the rioters, 77.1 percent-but only 61.6 percent of the noninvolved-identified him correctly. The overall scores on a series of similar questions also reflect the self-reported rioters' higher levels of information.

Self-reported rioters were also more likely to be involved in activities associated with Negro rights.140 At the most basic level of political participation, they were more likely than the noninvolved to talk frequently about Negro rights. In the Newark survey, 53.8 per