This study analyzes the mobility of workers initially employed in low-paid occupations who moved to moderately paid occupations, based on 18,347 observations of 1970 Census data, compared to 1965 data. The study relies on the concept of labor segment, which provides an antidote to the individualistic perspective. Two broad segments, a low-paid and a mainstream stratum, are defined in order to operationalize mobility of the type that might lift a family or individual out of poverty. To permit investigation of effects due to race, gender, and industrial sector, the low-paid stratum is subdivided into eight labor segments. In its linear form, the mobility model specifies that for each labor segment the probability of upward mobility is the sum of three effects: age, years of schooling, and current occupations. The estimates indicate that the chances for upward mobility differ across low-paid occupations. In addition, there are large effects for race and gender, and a substantial effect of industrial sector on black males. The mobility flows examined through the large data base show the operation of segmentation factors. The results support the segmentation view of the labor market, i.e., a person's life chances are factors other than his or her productive potential. Socioeconomic inequality is exacerbated because gender, race, industry, and occupation are determinants of mobility. (KC)