Numerous studies have sought to test which theory provides a better description of the attainment process, but the results have been in-conclusive. In this article, the contentions of the two theories are reviewed, and an alternative theory-a signaling theory of promotions-is developed which combines each of the elements. Hypotheses about the effects of early jobs and earnings, on which the theories suggest different predictions, are then proposed. The hypotheses are tested using data on a 3-year entry cohort whose attainments are followed longitudinally over a 13-year period. The analyses are able to examine the effects of early earnings, early job status, and even specific job titles on later earnings and status attainment. The results suggest patterns of career mobility which conflict with the "trade-off hypothesis" suggested by human capital theory and the "ahistorical-effects hypothesis" suggested by structural theory. The results most closely support the signaling theory.
Human Capital versus Social Structural Theories
The issue of whether careers are structured by early social forces and barriers is seen theoretically in the conflict between human capital theory in economics (and, more generally, neoclassic economic theory) and social structural theory in sociology. Human capital theory posits that labor markets offer open opportunity and that individuals' attainments are largely a function of their effort, ability, education, and training. It also posits that individuals' ability and effort create productivity, which is the basis for their compensation. In addition, it asserts that individuals create their own productive capacity by making investments in themselves.
In contrast, social structural theory contends that individuals' attainments are determined less by their personal attributes than by structural aspects of labor markets. It posits that the structure of existing jobs controls the "wage contour" for various kinds of employees (Doeringer and Piore 1971; Thurow 1975). As a result, individuals' earnings will not be directly determined by their personal attributes, efforts, or investments in themselves, but by the jobs to which they are allocated. Jobs are assigned wage rates according to normative considerations, and individuals are paid according to the jobs into which they are allocated (Thurow 1975; Granovetter 1981). As a result, the structural model contends that attainments are highly structured, and the effects of personal qualities on earnings are mediated by this structure.
When social structural theory is applied to organizational careers, it is commonly conceived as an internal labor market (Doeringer and Piore 1971). An internal labor market is characterized by a series of discrete job ladders which allow entry only at the bottom and whichdetermine upward movement in response to the development of knowledge or skill (Doeringer and Piore 1971, p. 21; Spilerman 1977; Althauser and Kalleberg 1981, p. 130). Internal labor markets presume that early selections determine the kind of training the organization provides to employees and the ultimate career paths employees will follow.
A Signaling Theory of Promotions
Although human capital and social structural theories are the two main theories to be applied to describing individuals' careers in organizations, a third theory, signaling theory, is also applicable, and this article presents qualitative and quantitative evidence which supports this theory. According to signaling theory as it was originally proposed (Arrow 1973; Spence 1973; Stiglitz 1975), employers seek to make their employee selections conform to human capital theory; but, contrary to the assumptions of human capital theory, employers have difficulty knowing which individuals have the most ability. The expense and difficulty of getting information about employees' ability induces employers to use certain social attributes and accomplishments of individuals as signals of their abilities.
Signaling theory is usually applied to explaining hiring decisions, since that is the prototypical case in which employers have little information about individuals. However, the problem continues to exist after employees are hired. Particularly in large organizations, large numbers of employees must be evaluated and compared; and these individuals often occupy very different jobs with incomparable products, and they are evaluated by different supervisors. In addition, most employees considered for promotions are in white-collar jobs for which job performance is difficult to evaluate. While extensive efforts have been made to develop assessment tests and procedures, these have generally been found to have limited validity (Campbell, Dunnette, Lawler, and Weick 1970) and are regarded with mistrust by managers (Shaeffer 1972). A few large corporations have devised elaborate assessment programs (notably UniRoyal and AT&T) (see Bray, Campbell, and Grant 1974); but only the largest companies can afford such programs, and even they cannot afford to implement them for more than one stage of the promotion process. Clearly, promotion committees face the problem of information scarcity, which suggests the relevance of signaling theory.
Wacancy theory may be considered a variant of social structural theory, although it generally follows chains of vacancies rather than the sequence of career moves of individuals (see White 1970a; Stewman 1975).
According to most accounts, supervisors' ratings are the main determinant of promotions. But complications prevent ratings from being the only criterion. Promotion committees must have some way of comparing high ratings by different supervisors and of comparing good job performance in jobs of different difficulty. Promotion committees need an objective, comparable indicator of job difficulty and of the value of the individual's contribution.
The organization considered in this study, like many large organizations, determines job compensation by a detailed job evaluation system which assesses the difficulty, challenge, and skill requirements of jobs (see also Treiman 1979; Grandjean 1981). The job evaluation system in this corporation devotes considerable effort and expense to assure that the job evaluation process is rigorous and that these rankings will be socially accepted as legitimate and fair. Within job statuses, salaries are set by a "merit pay" system which seeks to set pay-in-grade according to individuals' contributions. Given the extensive effort devoted to these compensation systems and to buttressing their legitimacy, job statuses and earnings would seem to be ideal signals of the ability requirements of jobs and of the ability that individuals demonstrate in their jobs. While not formally a part of the promotion process, these two compensation systems permit comparative evaluations of the abilities of individuals occupying different jobs. Are they indeed used as such?
Managers report that they are. In interviews with this researcher, the managers report that because of the difficulties promotion committees have in evaluating candidates' abilities, they sometimes infer individuals' abilities from their attainments as indicated by their job status and earnings. Since job status and earnings are objective indicators, and since they have strong salience and importance in the ways individuals are evaluated in this organization on a daily basis, they tend to have particularly strong weighting in these deliberations. What is note-worthy in these accounts is not just that these attainments affect promotion decisions, but also that they do so because they are inferred to be "signals" of individuals' abilities. This inference seems to be based on the assumption that the organization's selection system is meritocratic; consequently, those who have risen the furthest have done so because they are the most able. This rationale is most crisply stated in the frequently cited maxim, "the cream rises to the top." In effect, this maxim seems to be a shared informal "theory" about how to detect ability.
Signaling theory represents a middle course between human capital and structural theories. If promotion committees really do use job status and earnings as ability signals, then signaling theory predicts many of the consequences suggested by social structural theory even though it posits that managers are trying to pursue the aims of human capital theory. It contends that employers pursuing the goals of human capital theory will take structural factors as signals of individuals' ability, but it contends that job attainments affect later career attainments, not by imposing structural constraints, but by conferring information about employees' ability.
