SOCIETAL STRATIFICATION

Societal stratification phenomena are the relatively enduring, hierarchically ordered relationships of power among the units of which society is composed. The smallest units are adults, gainfully employed men and/or women, nuclear families, or sometimes extended families or households. Such units are ordered from highest to lowest in terms of power: political power, acquisitional power, the power of prestige, and the power of informational standing. Everybody experiences stratification every day, although a person often notices it only in the sense that some people seem better or worse off than he or she is. Social thinkers, powerful people, and revolutionaries have always been especially concerned with stratification.

Secure knowledge of the varying forms stratification structures may take is important because of the effects those structures have on many aspects of human experience, such as people's dreams of a better life, efforts to improve their situations, strivings for success, fear of failure, sympathy for the less fortunate, envy of others' good fortune, and even feelings about revolution.

A complete understanding of stratification requires several kinds of knowledge: first, what stratification structures consist of and how they vary; second, the individual and collective consequences of the different states of those structures; and third, the factors that make stratification structures change. This article reviews current thinking on the first of these elements.

HISTORY: CLASSICAL THEORY

Two different lines of thought inform modern theory on societal stratification. One is classical theory; concerned with political power and privilege, it employs historical evidence. The other is the empirical tradition, which deals with systematic data on stratification as it exists contemporarily. Present-day theory of the behavior of stratification phenomena can be traced to Karl Marx's challenge to the manufacturing and financial elites of his day. Behind his concerns and those of the working class for which he was Europe's chief spokesman for many years lay the great economic and political upheavals of the eighteenth and nineteenth centuries.

The American and French revolutions and their aftermath culminated in legislation that made adults in many countries equal before the law. The related wave of emancipation of slaves and serfs in Europe and the Americas was also part of the intellectual environment of that day. Of more direct relevance to Marx's thinking was the rise of trade and the factory system, along with the growth of cities and the expansion of wealth. Marx saw urban populations dividing into two opposed classes. The capitalist class employed the workers; owned the workplaces, machines, and tools; and had ready access to large amounts of money for investment. The capitalists were opposed by their employees, the working class, who had nothing to offer but their time and energy. In Marx's view, these two classes differ in terms of power and privilege: power because capitalists give orders that workers must accept, privilege because capitalists take the surplus (whatever is left after paying the cost of production) for themselves and their investments, leaving for workers only the wages that the market for labor forces capitalists to pay. Actually, Marx was interested in how these classes came into being and the conflicting interests they expressed. He did not write specifically on societal stratification as it is understood today.

Later writers on stratification, attempting to elucidate or contradict Marx, spelled out more complex sets of stratification dimensions. Weber (1946, 1947) saw power as the general factor basic to the enduring inequalities referred to as stratification. Sometimes, like Marx, he used categories whose underlying dimensions had to be elucidated by others. Party, class, and status groups were his key concepts. When these concepts are dimensionalized (reconstituted as variables), "party" is seen to be legitimate political influence, "class" is seen to express a hierarchical order of economic status, and the variable underlying "status groups" is seen to be their hierarchical order according to the degree of social honor. In other writings, Weber saw education as a stratification variable. In still others, he often wrote about authority, or legitimate superordinate and subordinate relations of power. Weber said nothing about how people are distributed in these dimensions or, of course, about how and why such distributions vary.

More thoroughly and precisely than Marx or Weber, Sorokin (1927) crafted the bases of modern theories of societal stratification. He distinguished political stratification, economic stratification, and occupational stratification. The first is a dimension of political power, and the second a dimension of the power of income and wealth. He left the dimensionality of occupational status unclear, sometimes implying that it was authority, sometimes privilege, and sometimes intelligence. Much of Sorokin's theory of societal stratification remains intact. First, he noted that all societies are stratified to some degree, a position widely accepted today. Second, empirical researchers continue to refine and elucidate his concepts of occupational status and occupational mobility. Third, in this connection he asked why occupational stratification exists and concluded that organized communal life requires mechanisms and people to coordinate essential activities and that such coordination demands and rewards unusual ability. This view, now called the functionalist hypothesis, has been elaborated and disputed ever since. Fourth, he held that the degree of stratification varies from society to society and over time within given society: Stratification, he said, is in "ceaseless fluctuation." Sorokin specified several ways in which stratification structures may vary. The whole structure may rise or fall; the top may rise or fall, changing the degree of inequality; and the "profile," or the shape of the distribution, may vary. Similarly, the rate of individual upward or downward mobility may vary, and whole strata may rise or fall.

Sorokin thus presented a theory that specified (1) the general dimensions by which people are stratified within a society, (2) some ways in which the distributions of people on those dimensions may vary, and (3) why stratification exists. Also, he held such structures to be in ceaseless change.

The latest work in the classical tradition is that of Lenski (1966). His key dimensions are power, privilege, and prestige, in that order of importance. Beyond this, Lenski offers three main ideas. First, both functional theory and conflict theory, its opposite, are partly right. Society's needs demand coordination, implying the existence of strata based on power or authority and implying a degree of consent on the part of many of those whose activities are organized by others. However, conflict results from that control: Authority is often abused and, even when it is not, may be misunderstood. Second, inequalities are mostly those of power, with inequalities of privilege and prestige following mostly as consequences of them. Third, the degree of inequality, which is seen as a single phenomenon encompassing the rate of mobility and the distance between strata, increased with the growing comprehensiveness and complexity of society until the Industrial Revolution, after which it declined. According to Lenski, the main forces driving change in the degree of inequality are the size of the surplus of production and, undergirding this, the march of technological efficiency.

Lenski is clearly in the classical tradition in his concern with power and privilege and dependence on historical evidence. To some extent, he echoes Sorokin's concern with variations in stratification structures through an emphasis on the degree of inequality. He provides a compelling treatment of the issue of conflict versus societal necessity in regard to the existence of stratification. He uses historical evidence effectively and systematically to mark variations of inequality in agrarian and horticultural societies. However, Lenski's emphasis on two main, all-encompassing aspects of stratification—power (his key criterion variable) and inequality (used to denote the way in which power and its concomitants are apportioned)—forces too many separately varying stratification phenomena into too few molds. This problem becomes critical in industrial societies, where stratification dimensions vary independently of one another.

HISTORY: THE EMPIRICAL TRADITION

As has been noted, this tradition of research on stratification is concerned with the here and now. This line of research has developed excellence in the measurement of the hierarchical positions of small demographic units within larger stratification structures. Although newer than the classical tradition, it has a long history. Several more or less independent status-measurement devices were formulated in the 1920s and 1930s. Most were concerned with either the prestige of the breadwinner's occupation or the quality of the home. They tended to share certain assumptions: that stratification consists of a single hierarchy, in the early days usually called social class; that one or two different scales are sufficient to test hypotheses concerning social class; that social class positions can be distinguished by direct observation and/or interviews with someone who knows the status holders; that routines can be devised that allow one to assign valid and reliable numerical scores to each status holder on each of the scales used to measure social class; that the unit to be scored is the household, which can be one person or several persons living in a single home; and that it is the whole unit that is to be scored, whether with data on the home or data on the head of the household. Many of these devices became obsolete because they had to be recalibrated for each new community or type of community to which they were applied. Those that survived—education and occupational status—did so because they provide comparable scores across large populations, such as nations.

Of the two main survivors, educational attainment is easy to measure: the exact number of school years successfully completed from none through sixteen, seventeen, eighteen, and so on. Measurement of occupational status is another matter. Two systems are currently in use. Occupational prestige ratings assume that each person in a given occupation shares the prestige most people attribute to that occupation. Occupational prestige scales have been constructed for many countries (Treiman 1977). Occupational socioeconomic status indices (SEI) are scales that use education and income to measure the status level of each occupation and then attribute to an individual the resulting score of her or his occupation. In the United States, Treiman's prestige scale and the SEI provide highly correlated occupational scores (Featherman and Hauser 1978).

Regardless of the original intent of such scales—to measure positions in what once was believed to be the only stratification hierarchy—the two variables educational attainment and occupational status are also appropriate for use with the classical theorists' multidimensional view of stratification.

A SYNTHESIS

The current synthesis was carried out by stratification theorists who were both sensitive to the concerns of classical theorists with power and privilege and steeped in the empirical tradition. Thus, they brought the classical theorists' concern with political power, economic power, and social honor (Weber), including Sorokin's occupational status and Lenski's prestige, together with the empiricists' concern with education and occupational status (overlapping Weber, Sorokin, and Lenski) and with quantitative measurement and analysis.

Svalastoga's Social Differentiation (1965) appears to be the first statement of the synthesis. Svalastoga indicates the centrality of four dimensions of status: political, economic, social (mostly occupational), and informational (mostly educational). He calls attention to structural variations through his "parameters": the degree of inequality, the correlation among dimensions, and the degree of permeability (intergenerational circulation mobility or movement up and down the hierarchies). Duncan (1968) both accepted and clarified Svalastoga's synthesis. His list of "scales of reward or status" provides a good outline of the large number of variables that should be measured to achieve a full-scale determination of people's levels on each status dimension. Also, he divides three of Svalastoga's four dimensions into two categories each. He, like Svalastoga, then lists three ways in which the structure of stratification variable may vary. The first is the degree of inequality. The second is called "rigidity of inequality" or "status crystallization," which is the same as Svalastoga's "correlation." The third is "rigidity of stratification," which is Svalastoga's "permeability" turned upside down.

Like Sorokin's and others' positions, Haller's (1970) statement of the synthesis assumes that stratification to one degree or another exists in all societies at all times. Revised slightly in this article, this form of the synthesis holds that there are two classes of dimensions of stratification. The first are "content"— or power—dimensions, after Weber: the capability of a given unit to elicit from others behavior promoted by the first unit, with such power having been routinized by coercion or consent. Agreeing conceptually but not always terminologically with the classical writers, this expression of the synthesis posits political power, economic power, and the power of prestige as universal dimensions of power. For civilized societies, it adds the power of years of formal education.

This position thus posits legitimatized political influence (including authority) as the dimension underlying Weber's "party," Sorokin's "political stratification," Lenski's "power," and Svalastoga's "political status." It posits Weber's "class," Sorokin's "economic stratification," Lenski's "privilege," and Svalastoga's economic status as referring to the same set of hierarchical phenomena: access to goods and services—the economic dimension of a stratification structure. From Weber, it takes the variable of social honor; from Sorokin and modern occupational status researchers, that of occupational stratification; from Svalastoga, that of social status; and from Lenski, that of prestige. From the empirical tradition, it takes the measurement of occupational power ("status"). All these elements are seen as referring to a third homogeneous set of hierarchical phenomena: the power of respect or deference attributed to a unit because of that unit's participation in a social category (such as an occupation) that has a specific level of evaluation by a society—the prestige dimension of societal stratification. As has been indicated, from Svalastoga and Duncan, with much support from the empiricists and also some from Weber, it takes informational power as a content dimension of a stratification structure, with education as its main indicator.

At the general level, each power dimension is of course presumed to be applicable in some form to all human societies as far back as human communal life can be traced. It is the exact expression of each dimension and the relationship among the dimension that vary across time and place. For entire contemporary societies, the main expressions of each dimension seem to be the following: for the political power dimension, political power, a variable researchers cannot yet measure despite its centrality in classical theory; for the economic dimension, income (occasionally wealth), a variable of concern to those in the empirical tradition; for the prestige dimension, occupational status in either of its two main forms of occupational prestige ratings (Treiman 1977) and occupational socioeconomic index scores (Featherman and Hauser 1978); and for the informational power dimension, educational attainment level in terms of years of formal schooling successfully completed. Thus, in recent years it has become apparent that for today's societies, the main variables of the empirical tradition have central places among the content dimensions of the classical tradition. Income, occupational status, and education are the theoretically defensible variables most readily available to measure three of the four classical content dimensions.

Like Sorokin's, Svalastoga's, and Duncan's, Haller's formulation of the synthesis specifies several structural dimensions, with each one held to be applicable to every appropriate measure of each content dimension. The three structural dimensions of Svalastoga and Duncan are included: degree of inequality, status crystallization, and degree of status inheritance. Two others from Sorokin are included, although they are modified to fit today's understanding. One is the general level or central tendency, and the other is a division of Sorokin's concept of profile into two concepts: mode structure and skewness. Although calculated from data on small units, each structural dimension applies to the society as a whole. Although logically they are partly dependent on one another, each one makes a unique contribution to an understanding of stratification. Each appears to be amenable to statistical description. Each is applicable to every indicator of the standing of every small unit (say, family) in the society. Valid measures of each content dimension taken at one point in time on a generalizable sample of the population of small units of that society would provide a complete description of the stratification structure of that society at that time. Successive measures would provide a complete description of the evolution of that society's stratification structure over time, thus providing a general idea of the variations in the degree of stratification in that society. Each applies to comparisons over time or among societies.

General Level. As Sorokin realized, the levels of structural dimensions may rise and fall as wholes. That is, the average economic, political, prestige, and informational standing of small units changes over time. These rises and falls may be seen in changes in the central tendency—say, the arithmetic mean or the median value—of the standing of small units. The rises and falls of the central tendency of any one of these dimensions do not necessarily follow the same pattern as those of another. Average economic, prestige, and informational power may increase, for example, while average political influence falls. This could happen in a society where a development-oriented dictatorship reduces citizen political participation while increasing levels of income, raising prestige by upgrading the occupational structure, and increasing access to education. Indeed, the economic, prestige, and educational levels of the populations of the more developed democracies have increased almost consistently since World War II, though this may not always be said for dictatorships. Also, raising the level of the occupational structure is exactly what some researchers mean by upward structural mobility, the case in which almost everyone is carried upward by changes in the economy that eliminate low-skill jobs while adding specialized jobs.

Degree of Inequality. The distances among the small units of a society may increase or decrease over time. This, so to speak, stretches the positions on the power dimensions apart or squeezes them together. The statistical term for this is the degree of dispersion. A number of measures of dispersion exist, such as the standard deviation (or its square, the variance), the range, the semi-interquantile (or quintile, decile, etc.) range, the share distributions, and the Gini, Theil, and Kuznets coefficients. There are two basic types of inequality: absolute and relative. Absolute conceptions assume that the metric on which the degree of inequality is measured is fixed so that as, say, real income per capita grows, the dollar difference between the mean of the top 10 percent and the mean of the bottom 10 percent of the small units may increase while each is rising above its previous level, with the slope of the top rising faster than is the slope of the bottom. For income, a proper description of these phenomena would be "changes in the size distribution of income." Absolute inequality and its changes are sometimes published. Much more often published are the share distributions of income. For any society at any time, share distributions take the total amount of, say, income as a constant 100 percent (or 1.00) and determine the degree to which the whole amount, regardless of its absolute size, is evenly or unevenly divided among the population. These distributions include the percentage of all income held by the top X percent and the bottom Y percent of the population. Or, as in the case of the Gini, Theil, and Kuznets coefficients, they use values ranging from 1.0 to zero, in which 1.00 is the maximum degree of inequality and zero is complete equality. Viewed at one point in time in a single society, measures of relative inequality are useful, but for comparison among societies or across time in the same society, they may be misleading. In fact, for many years the share distribution measures of the income of the American people remained essentially unchanged while the size distribution inequality increased dramatically (U.S. Department of Commerce 1980). This was the case because real per capita income was increasing rapidly. The greater the degree of inequality, the greater the degree of stratification.

Crystallization. It has long been recognized that a stratification structure may tend toward or away from monolithicity, in which the different power dimensions merge into a single hierarchy or tend to be in partially separate hierarchies. At one extreme, the position of a small unit on any one of the dimensions can be found by knowing its position on any other dimension. In other words, if the four content dimensions are perfectly correlated, those in lofty positions on one dimension also will be in lofty positions on all the other dimensions, while those in humble positions on one will be in similarly low positions on any other. At the opposite extreme, a unit's position on a given content dimension is irrelevant to its position on any other. In the real world, any two or three might be highly interrelated, all might be moderately intercorrelated, and so forth. For obvious reasons, Svalastoga called this structural dimension "correlation." Others have called it "status crystallization." Note that crystallization levels and forms may be summarized better by a method called factor analysis than by the correlations themselves. Factor analysis can show which sets of content variables tend to vary together in a population and which do not. It also can help determine which are the dominant dimensions and which are of lesser importance in a given stratification structure. For example, it appears that the former Soviet stratification structure was dominated by the political dimension; the American, by the economic dimension. Factor analysis of the correlations of the content dimensions could indicate whether these beliefs are true. The greater the degree of crystallization, the greater the degree of stratification.

Status Inheritance. Status inheritance is the degree to which people's level on a given content dimension is controlled by that of their parents. It is exactly the obverse of circulation mobility: A high degree of power position inheritance implies a low degree of circulation mobility. The basic statistical summary of this phenomenon is either the correlation coefficient (r) or the coefficient of determination (r2) of the power dimension positions of offspring and their parents. (The r2 tells how much one variable is determined by the others.) The greater the degree of status inheritance, the greater the degree of stratification.

Sorokin's Profile. Every variable has a socalled distribution, a shape that appears when the number of scores (the frequency) is plotted against the value of scores. Much statistical theory today assumes that real-world distributions conform to certain mathematical shapes. The bell-shaped "normal" curve is the one most often used. For distribution of income, the "log normal" curve, with which the distribution of the natural logarithm of the individual amounts forms a normal curve, is often employed. Stratification researchers often take it for granted that the distributions of power dimensions are either normal or log normal, but there is no sociological reason to assume this. The shape of the distribution of a content dimension is precisely what Sorokin meant by his term "profile." Lacking the data and concepts to proceed further, he simply called the real-world shapes of these distributions their profiles. Today we can see that there are two aspects of each profile: mode structure and skewness.

In strikingly underdeveloped societies, almost everyone is concentrated at the very lowest values of economic, political, occupational, and educational power: extremely poor, utterly uninfluential, of low prestige, and illiterate. Above those people, their "betters" are arranged in rank order, with a wide range in which the few people who are above the bottom dwindle up the line to a handful of individuals of lofty standing. Each such distribution would have a very low mode (or distinct cluster of cases) and median (where half of the cases are higher and half are lower) and a higher arithmetic mean, with a sharply skewed tail. In somewhat more developed societies, such distributions, instead of yielding bell-shaped or log normal curves, might show multiple modes, with many people concentrated around a fairly low point, quite a few concentrated around a point a bit higher, a few concentrated toward the top, and after that a sharp skewing up to the very few at the top. The consequences of such forms for the lives of the people involved are no doubt great. For example, if in a certain society almost everyone is destitute, the few who are more or less well to do are highly visible. Even if the wealthy were really not far above the others, everybody would think of that society as being highly stratified. If in another society people are bunched together at several points along a hierarchy, thus forming multiple modes, or discrete classes, those in each mode might come to consider themselves members of a special class in opposition to those concentrated at another mode. Thus, the exact forms of profiles are essential to a description of a society's stratification structure. Theoretically, these forms have substantial consequences for many stratification-dependent behaviors.

Profile: Mode Structure. Mode structure refers to the number, size, and location of distinct modes on the distribution of each content variable. In polymodal structures, the more pronounced the modes, the greater the degree of stratification.

Profile: Skewness. Several statistical devices exist to mark the degree of skewness. The greater the level of skewness, the greater the degree of stratification.

ILLUSTRATIONS

Data by which to measure and compare stratification structures are exceedingly difficult to obtain. A complete description at a specific point in time requires well-measured, valid indicators of four power dimensions, one or more for each dimension. For each indicator, several measurements must be made: The average level, the degree of absolute and relative inequality, the degree and factor-analytic forms of the crystallization of the whole set of indicators of the power dimension, the degree of power position inheritance, and the distributions of each one must be plotted to indicate its mode structure and measure its skewness. Describing such an overall structure requires the construction of 24 or more different indicators of structural dimensions. These indicators have to be based on representative, societywide samples large enough to permit the recording of small differences, as in the case of the few people at the upper end of a skewed distribution. The study of variations in the structure of stratification demands that comparable measurements be taken on the same variable at different times and in different places. In itself, the requirement of comparability is extremely severe when one is making comparisons among societies with different cultures or over long periods of time within the same society.

Exploratory work of this sort has been conducted on data provided by Brazil. The data were collected on a national probability sample of households in 1973 and are available for all employed men and women in the households sampled. These people are the "small units" of the descriptive analysis presented below. Brazil is a particularly good place in which to conduct such exploratory research for two reasons: It is a large country whose regions are markedly different from each other in terms of development, and it has only one language and culture. The first factor makes it feasible to test for structural variations of stratification associated with development levels, treating regions as societies; the second eases the problem of comparability.

As was indicated earlier, it is not currently feasible to obtain measures of the political power dimension in Brazil or anywhere else. However, there is widespread agreement that income is a proper measure of the economic status dimension, that occupational status instruments based on the average education and income of each occupation are proper measures of the prestige dimension, and that education is a similarly appropriate measure of the informational status dimension. These data are available for some of the parameters that would have to be assessed to obtain a complete description of the regional-development variations of the Brazilian stratification structure in 1973.

Here one is comparing sharply different development regions. The stratification structures of three of Brazil's socioeconomic development (SED) macroregions in 1970 were delineated by obtaining multiple-item, factor-weighted SED scores on that nation's 360 official continental microregions and plotting their levels on the map of Brazil (Haller 1983). This showed the following five macroregions: the Developed South (median SED = 78 on a scale of zero to 100), the South's Developing Periphery (median SED = 54), the Undeveloped Amazonia (median SED = 32.5), the Unevenly Developed Northeast (median SED = 31), and the Underdeveloped Middle North (median SED = 13).

Obviously, this article cannot reproduce each one of the structural dimensions for each SED macroregion for men and for women. Instead, it provides a few key illustrations for three of the regions: the Developed, the Developing, and the Underdeveloped.

Variables routinely used as indicators were formulated to measure three of the four stratification content dimensions: education in years successfully completed, occupational status scores (composed of canonically weighted scores based on the education and income of each occupation), and annual income in 1973 U.S. dollars.

The illustrations are based on regularly employed men and women 15 to 65 years of age. All such persons who lived in the three regions under comparison and were part of the sample have been included. The numbers of sample members vary sharply by region and by sex. The Developed South is much more populous than the other two regions, and about three times more men than women are employed. The largest of the six gender-by-region subsamples thus consists of men in the South: over 40,000 (see Table 1). The smallest consists of women in the South's Developing Periphery: over 2,500.

Let us begin with the profiles (see Figure 1), graphs that have been sketched to show the shape of the stratification structures for men and for women as they appear in the three regions. There are two reasons for paying close attention to these curves. First, they show power relations among the people. The presence of multiple modes shows the existence of discrete and potentially opposed classes. Both the mode structure and the marked skewing indicate a high degree of stratification for each sample. Second, the fact that these distributions diverge sharply from normal or log normal curves shows that the numbers, that is, the data presented in Tables 1 and 2, are at best approximate because the shapes of the distributions affect their meaning.

The curves show the following:

1. Multiple modes are exhibited by both men and women in 11 of the 12 graphs pertaining to the developed and developing regions. The exception is distribution of women's income in the developing region.
2. For the two most developed regions, comparable curves show just about the same mode structure. In these regions, education tends to be bi- or trimodal and occupational status tends to be at least trimodal. Among men and among women in the developing region, income also exhibits multiple modes. In the underdeveloped region, the shape of the curves is markedly different form that of the others. The curves in this region show a heavy concentration of both men and women at the bottom of each indicator variable, though some of the region's six graphs show the formation of small second and sometimes third modes at high status levels. The apparent conclusion is that the underdeveloped area exhibits a relatively high degree of equality at the very bottom of the Brazilian stratification structure. This is precisely the opposite of the thinking among many observers of Brazil, who believe that inequality is greater in the underdeveloped region (perhaps because of the glaring visibility of the tiny stratum at the top).
3. Each curve shows a high degree of skewness. That is, the highest positions are held by a tiny proportion of the people, and on the whole, as the tail of the distribution lengthens, the higher the level, the tinier the percentage of the people.
4. In every case, the main modes are the one or two at the bottom, where most people tend to be concentrated.
5. Almost every graph shows a tendency for one or two smaller modes to appear toward the middle of the distribution. For education, this occurs at around grade 12. For occupational status it is about 50 units, or the level of office clerks, primary school teachers, and the like. For income, it is about $2,000 to $3,000 per year, or a monthly wage between $160 and $250.
6. There may be a tiny mode near the top of the educational and occupational status distribution in the more developed regions.
7. For occupational status and income, women are more concentrated toward the bottom than are men.
8. Clearly, the main regional variations in profile are between the two more developed regions and the underdeveloped region.
9. In terms of mode structure, the more developed areas seem more stratified than does the underdeveloped area.
10. In terms of skewness, it appears that the underdeveloped area is more highly stratified.

Data on the general levels and absolute inequality levels of the three content dimensions are presented in Table 1. For occupational status, the degree of circulation mobility also is presented. The general level rises with development for all three variables, except for the occupational levels of women in the developing region, whose status is slightly higher than that of women in the developed region. Again, with two exceptions among women in the developing region, the higher the level of development, the greater the degree of absolute inequality. Echoing what was gleaned from the graphs, the general level and the absolute inequality levels of the underdeveloped region are markedly lower than those in the other areas. Finally, among men, the higher the level of development, the lower the degree of circulation mobility. Women show no trend in this regard.

Evidence regarding structural crystallization is presented in Table 2. Among men, the higher the level of development, the higher the degree of crystallization. Among women, the same trend may be present in the data, but with one small inconsistency.

CONCLUSION

This article has attempted to describe the contemporary synthesis of classical and empirical traditions of sociological thought concerning societal stratification, with special emphasis on what may be learned about ways to describe variations in stratification structures. It also presents some illustrations showing how indicators of some of the structural dimensions vary across development regions among employed men and women in

Brazil. In general, these indicators show that the more highly developed a region is, the more stratified it is.

Measuring stratification variations among societies is an immense task because of the number of variables that must be studied and because of differences in culture, language, and social organization among peoples. Still, at both the individual and societal levels, the effects of structural differences of stratification are among the most perplexing of this age and perhaps of all ages. For this reason, understanding how and why stratification structures vary and specifying the consequences of such differences are worth the considerable effort required.

references

Duncan, Otis Dudley 1968 "Social Stratification and Mobility: Problems in the Measurement of Trend." In Eleanor Bernert Sheldon and Wilbert E. Moore, eds., Indicators of Social Change: Concepts and Measurements. New York: Russell Sage Foundation.

Featherman, David L., and Robert M. Hauser 1978 Opportunity and Change. New York: Academic Press.

Ganzeboom, Harry B. G., Donald J. Treiman, and Wout C. Ultee 1991 "Comparative Intergenerational Stratification Research: Three Generations and Beyond." Annual Review of Sociology 17:277–302.

Grusky, David B. 1994 Social Stratification: Class, Race, and Gender in Sociological Perspective. Boulder, Colorado: Westview Press.

Haller, Archibald O. 1970 "Changes in the Structure of Status Systems." Rural Sociology 35:469–487.

—— 1983 The Socioeconomic Macroregions of Brazil. Lanham, Md.: Bernam-Unipub.

Lenski, Gerhardt 1966 Power and Privilege: A Theory of Social Stratification. New York: McGraw-Hill.

Sorokin, Pitrim A. 1927 Social Mobility. New York: Harper.

Svalastoga, Kaare 1965 Social Differentiation. New York: David McKay.

Treiman, Donald J. 1977 Occupational Prestige in Comparative Perspective. New York: Academic Press.

U.S. Department of Commerce 1980 Social Indicators II. Washington, D.C.: U.S. Government Printing Office.

Weber, Max 1946 "Class, Status, and Party." In Hans Gerth and C. Wright Mills, eds., From Max Weber. New York: Oxford University Press.

—— 1947 "Social Stratification and Class Structure." In A. M. Henderson and T. Parsons, eds., The Theory of Social and Economic Organization. New York: Free Press.

Archibald O. Haller