¹ In addition to the works cited by Chance and Taylor, see Carlos Assadourian, S., Beato, Guillermo, and Chiaramonte, Juan, Argentina de la conquista a la independencia (Buenos Aires, 1972), pp. 309–49;Google Scholarde la Plaza, Salvador, “La formation de las clases sociales en Venezuela,” Cullura Universitaria, v. 87 (1965);Google ScholarMiranda, José, España y Nueva Espaiia en la é'poca de Felipe II (Mexico, 1962), pp. 94–103.Google Scholar That there were other hierarchies as well is pointed out in Colmenares, Germán, Historia económica y social de Colombia, 1537–1719 (Bogotá, 1973), pp. 408–460.Google Scholar

² Grubessich, Arturo, “Social History of a Colonial Hispanic American City: Valparaiso (Chile) During the Last Quarter of the Eighteenth Century” (M.A. thesis, University of Minnesota, 1978), p. 34.Google Scholar Given the novelty of this kind of information, the deviations are reported here in detail.

The chart shows that some two-thirds of the variations were between proximate groups in the socioracial hierarchy, such as “Spaniard-Don” and “Mestizo-Spaniard;” however, 30.5 percent of the variations, some 15 percent of all geographically stable male household heads, manifest large leaps or falls in their status. This kind of data may be generated only through minute linkage and systematic analysis of information on the lives of ordinary people from a wide variety of sources. See Grubessich, pp. 30–35.

³ Konetzke, Richard, Colectión de documentos para la historia de la formatión social de Hispano America (Madrid, 1953–1962), 3: 406–07.Google Scholar

⁴ Grubessich, pp. 41–50.

⁵ The choise of a random model is made on substantive and statistical grounds. First, the estimates have an immediately understandable, intuitive meaning. Second, they facilitate further statistical elaboration (impossible where the expected value$.are 0 or 100 percent as in the implicit model used by Chance and Taylor). Most importantly, this approach advances comparative research independent of the actual racial structures and even of the stratification criteria being applied. It should be apparent that the expectation of random mating is a statistical hypothesis for purposes of measurement and not a substantive one that the researcher actually hopes to confirm.

⁶ They incorrectly refer to their measure as a rate while it is in fact a ratio, that is the proportion of grooms for each racial group who married brides of other groups. Actual rates, based on the marriageable population by race and sex, would be extraordinarily interesting for exposing the real marriage propensities and determining the relative importance of demographic (sex-ratio) and social (stratification) forces. Census data at their disposal, properly tabulated and related to the marriage acts, could be made to yield this important information. In communities where marriage probabilities by race vary greatly, failure to incorporate this feature could lead to spurious conclusions. This problem plagues all analyses of endogamy regardless of the substantive issue. Most researchers proceed with a warning and with little hope of obtaining the appropriate denominator. This is one of the few cases in which the relevant information for the population at risk may exist. See note 10 below.

⁷ Strauss, David J., “Measuring Endogamy,” Social Science Research 6 (1977): 225–27.CrossRefGoogle Scholar The statistic K*, which measures the relative incidence of endogamy, is computed as follows:

Where,

P_o =Sum of Observed Endogamous Marriage/Total Marriages

P_e =Sum of Expected Endogamous Marriages (E_ii)/Total Marriages

E_ij = (Brides_i/Total Marriages) x Grooms_j

P* = Sum of Minimum Number of Brides or Grooms for Each Group/Total Marriages

(NOTE: Data may show some rounding effects; all calculations were performed with unrounded figures through the final step.) See also Bishop, Yvonne M. M., Feinberg, Steven E., and Holland, Paul W., Discrete Mullivariate Analysis: Theory and Practice (Cambridge 1975), pp. 393–97.Google Scholar

⁸ We prefer the percentage difference statistic because of its direct substantive significance and ease of interpretation. The K* statistics, like most measures of association, are more difficult to interpret meaningfully.

⁹ The expected values are derived by a process of proportional iterative fitting. See Strauss, pp. 239–41, and Bishop et al., pp. 206–09,398–99. Bernard, Richard M. and Sharpless, John B., “Analyzing Structural Influence on Social History Data,” Historical Methods 11 (Summer 1978): 113–121, use a similar but more refined approach to estimate the effects of demographic constraints but they neglect measures for analyzing social processes.CrossRefGoogle Scholar

¹⁰ Marriage figures must be related to population data derived from a timely census indicating the number of unmarried adults by race and sex. A very rough indication of the variations can be derived from comparing the occupational data by race for adult males with marriages as follows:

This suggests that Indian men married twice as frequently as peninsulars. In fact, aside from mulattoes, there is a pronounced inverse correlation between race and frequency of marriage. Similar but more refined data for both sexes are required to develop this line of analysis with confidence. Finally, one should not overlook the opportunity to analyze occupational and marriage data by broad age groups. Some differences may be due to structural differences in ages for each group; others may be due to life-cycle differentials by race.

¹¹ The authors lament the lack of occupational information for the Indians but nevertheless attempt to dissect the economic structure of the city without them (1543 were presumed to live there but their occupations were not specified). This omission means that a goodly proportion of the bottom half of the occupational pyramid must be ignored (1543 versus 3198 with known occupations). A related problem is that the survey reveals far too few common laborers, porters and the like. In a city with seventy-two high royal officials there were only ninety-three male servants listed. Similarly, 215 merchants required only fifteen muleteers. In an economy with many artisans there were virtually no laborers, odd-jobbers, carriers or simply unskilled hands to aid them, if the enumeration is accepted without adjustment. Moreover, unemployed males (102), farmers and farm laborers (172) and miners (14) are (unintentionally?) excluded from consideration (9 percent of listed workers; see p. 473, note 20). There is also a peculiar imbalance between merchants (215) and shopkeepers (18). Finally, Chance and Taylor classify all merchants as members of the elite (exactly two-thirds of the total) which seems excessive for the time and place. The result of these decisions and problems is to inflate the elite and make the lowest stratum of the working population disappear. In these circumstances analysis of the social structure is seriously distorted.

¹² This observation is drawn from Creider, Peggy, “A Study of the Casta Population of Buenos Aires based upon the Census of 1778” (summa cum laude thesis, University of Minnesota, 1968),Google Scholar done under the direction of Stuart B. Schwartz. It should be noted that after 1778, when attempts were made in Buenos Aires to create guilds, the efforts to form a shoemakers' guild broke down exactly over the issue of race. In 1793 the criollo shoemakers finally received permission to create a guild that excluded both castas and foreigners. See also Johnson, Lyman, “The Silversmiths of Buenos Aires: A Case Study in the Failure of Corporate Social Organization,” Journal of Latin American Studies 8.2 (11 1976): 181–213.CrossRefGoogle Scholar

¹³ See Somers, Robert H., “A New Asymmetric Measure of Association for Ordinal Variables,” American Sociological Review 27 (1962): 788–811;CrossRefGoogle Scholar or Loether, Herman J. and McTavish, Donald G., Descriptive Statistics for Sociologists (Boston, 1974), pp. 221–29.Google Scholar

¹⁴ See, for example, Lockhart, James, Spanish Peru, 1532–1560 (Madison, 1968),Google Scholar Cf. Martin, Norman, Los vagabundos en la Nueva España (Mexico, 1957).Google Scholar