Finally, we know that the interaction-emotional labor component of blue-collar jobs has increased dramatically. Team-based work and the need to coordinate with customers and technical personnel requires that workers motivate, placate, encourage, and please others as central requirements of their jobs.

This section considers the changing nature of service work—jobs in which face to face or voice to voice interaction is a fundamental aspect of work (MacDonald and Sirianni, 1996). Service work is historically categorized into three occupational groups: (personal) service, clerical and administrative support, and sales. For simplicity, we use the terms service workers, service jobs, or service occupations in discussing common aspects of these three broad occupational groups. Together these service jobs comprise roughly 41 percent of the workforce. Service jobs grew from 11 to 14 percent of the workforce between 1950 and 1996; clerical jobs grew from 12 to 15 percent; and sales, from 7 to 12 percent. Although women and minorities historically have filled many of these jobs, men have increased their relative numbers. In personal service, male workers rose from 38.3 percent of the workforce in 1983 to 41.9 percent in 1996; in clerical, males rose from 19.4 to 21.2 percent of the workforce; and in sales, from 42.5 to 45.3. This is in contrast to technical and professional work, in which the percentage of men fell by 4 percentage points in technical and 2 percentage points in professional occupations (Current Population Survey merged annual earnings files).

There is considerable variation in the content of work in these occupations along the dimensions we have identified. For example, many clerical occupations circa 1975 were still located in small, local establishments. Business office or customer service staff had moderate levels of autonomy, task scope, and cognitive complexity, more relational interactions with customers, but very low levels of technology use. By contrast, large service organizations included telephone operators, data processors, and typing pools, all of whom had very low levels of autonomy, task scope, and cognitive complexity; their interactions were transactional in nature, and their jobs were heavily mediated by technology and