In this sample of working men, the intercorrelations of education, occupation, and income are found to be moderate to fairly high. The relationships are also, in general, monotonic except that skilled manual workers tend to earn higher incomes than do white-collar clerical and sales workers. Despite the income variations however, the largest differences in subjective class identification are those separating manual from nonmanual workers. The importance of occupation to class identification is also indicated by multiple regression analysis, in which occupation and education show the largest regressionweights. income being considerably less important.
The relationships between a man's social origins and his current education and occupational status are lower than those found in previous Canadian and us studies. Part but certainly not all of the explanation for this difference probably lies in the more independent measure of occupational status employed in the present study. Neither the present data nor that from the Turrittin or Cuneo and Curtis studies support the contention that Canadian social structure is more ascriptive than us social structure (at least in so far as the influence of social origins on occupation is concerned).13
One explanation for the (relatively) weak influence of social origins on present occupation may be (as Cuneo and Curtis indicate) that we have relied on too narrow a range of social origin variables. Such variables as sex. immigrant status. religion. and region of birth may be much more predictive of later status than father's education and occupation alone.
An analysis of the effects of status inconsistency on subjective class identification shows no trend for people to identify with their highest or lowest rather than with their average status. The tendency to emphasize mean status is also shown by the finding that groups high in status inconsistency do not show greater variations in subjective class identification than groups with more consistent statuses.
The various aspects of status are only weakly related to a variety of (assumed) outcome measures paralleling the findings of Hodge (1970). Overall, the stratification variables are most highly correlated with number of organization memberships, feelings of powerlessness (negatively), and job morale. Education, income, and occupation are related in different directions to five of the nine dependent variables.
Looking at the effects of subjective class rating, we find that, with objective status controlled, the higher the subjective class identification, the more free-enterprise and antiunion the individual's orientations and thegreater his social participation. The direction of the causal relationships is. however, open to question.
The findings ofthis paper along with previous findings indicate the necessity for crosscrossnational Canadian studies relevant to some of the topics presented here. Until we have such data interpretations will remain speculative.
We should point out that the present data offer no argument against theories of the elitist nature of Canadian social structure. The finding that a man's social origins are far from predetermining later adult status for a total population does not contradict the finding of the low permeability of elite barriers. In fact, as elite theorists would nodoubt point out, social stability and lack of widespread protests against inequality might be partly a consequence of relatively low obstacles to mobility among the middle class of the nonelites. By definition only a relatively small (though perhaps a more influential) minority ever become personally aware of blockage to elite status.
Again, after examining correlations between social origins and current status and between status and a variety of possible correlations for a heterogeneous population, one can ask under what conditions or for which subgroups of the whole might correlations be larger? We have already suggested that status inheritance is higher than average among higher status groups. Moreover, it might be among men who are status congruent that status most influences attitudes and behaviour. Furthermore, SES-dependent variable relationships might be higher for those men who have held their current statuses for some time (e.g., older men, the nonmobile). It is thus in specification of relationships that the more interesting findings may lie.