ers in the sample, 17 percent of those in professional and managerial jobs reported working 50 or more hours per week, and approximately 26 percent of those in nonprofessional/managerial occupations reported working less than 30 hours per week. These results were based on self-reports of hours worked in the previous week. Furthermore, Jacobs (1998) suggests that self-reports of work week length correlate highly with independent measures of working time, such as time between departure and return (with commuting time subtracted).

Another relevant dataset is the WorkTrends™ survey, conducted annually since 1984 by Gantz Wiley Research of Minneapolis. This multitopic survey was administered to samples of U.S. households, stratified according to key U.S. census demographics, including age, income, and geography. Surveys were completed by principal wage earners, with the stipulation that respondents were employed full-time at establishments of 100 employees or more. Analyses of the WorkTrends™ database for the years 1985, 1990, and 1996 were prepared for the committee for use in this report. In 1985, there were 5,000 surveys distributed and 2,667 completed and returned (53 percent); in 1990 and 1996, 10,000 surveys were distributed with returns of 4,573 (46 percent) and 6,978 (70 percent), respectively.

Two major concepts are generally used to assess the meaning of work for the individual: (1) job involvement or work commitment is the extent to which work is a central life interest and (2) work values are the extent to which people place importance on various aspects of work. The ability to draw conclusions about trends in these two aspects of the meaning of work is hampered considerably by the paucity of useful data on worker attitudes. One dataset that does contain some information on trends in the meaning of work for Americans is the General Social Survey. This survey is a nearly annual, multitopic survey administered to a sample of roughly 1,500 adult, English-speaking American men and women (for an introduction to the General Social Survey, see Davis and Smith, 1992).