This study investigates the participation patterns of employer-sponsored adult education and training by Swedish workers over the period of 1975 - 1995, and evaluates the importance of the determinants of such provision. The study also explores the ways participation experiences in employer-sponsored education and training influence subsequent participation, occupational mobility and economic outcomes for Swedish workers. Data collected from 1975 to 1995 in The Swedish Living Conditions Survey (ULF), both the cross-sectional data and its panel component, have been analyzed. Contained in the panel are 3,319 Swedish adults who have been followed up in the ULF from 1979 to 1995. The study employs measures of work and job characteristics such as indicators of occupational status, wage, union membership, length of employment, job type, job responsibilities, influence on decision-making at the workplace, learning opportunities at the workplace, enterprise ownership, as well measures of personal characteristics, such as age, gender, level of formal education. The approach taken in the evaluation of the influences of work, job and demographic characteristics on the likelihood of receiving employer-sponsored education and training has been to develop and estimate logistic regression models by means of which these effects during different periods (1975, 1979, 1986/7, 1994/5) can be assessed and compared. Another three models have been investigated, using the panel data, namely: (1) a logistic regression model predicting subsequent participation in employer-sponsored education and training by similar experiences at earlier career stages; (2) a multiple regression model predicting 1994/5 annual income with participation history as a predictor; (3) a discriminant function analytical model predicting 1994/5 occupational status with participation history as a discriminating variable. The departure point of this study is that separate analyses have been done with the public sector and private sector sub-data sets. The findings indicate that occupational status, level of education, age, gender and to a less extent, union membership, and other work and job characteristics are the more important predictors of the likelihood of participation in employersponsored education and training for the Swedish workers in the private sector. For those working in the public sector, institutional factors relating to management style and job responsibility as well as age play an important part in training decisions. The findings also indicate that youth and older age groups, particularly those with the private sector, have been consistently under-served by provision of employer-sponsored education and training throughout the period investigated. The results reveal that by the mid-1990s, gender was not a issue in the public sector in terms of the participation rate but female workers were still disadvantaged in the private sector. The findings verify a trend that the participation gaps between the well-educated and the undereducated, and those between professionals and non-skilled and semi-skilled have narrowed by a great magnitude, yet not adequately to close up the gaps. In 1975, the likelihood of participation for a professional worker was nine times as high as that for a non-skilled worker in the private sector. By 1995, the comparable figure was three times. The findings from panel data analysis show that, for workers in the private sector, their participation status as of the mid-1990s was significantly correlated with their participation status back in the mid-1980s and late 1970s, even after the statistical adjustment. For the public sector employees, their participation status as of the mid-1990s is related to participation status in the mid-1980s only. The results of income estimation models show that all three indicators of participation status are significantly associated with higher earnings, but a further analysis that separates the private sector from the public sector indicates that the result applied more in the case of the private sector employees. In contrast, none of the measures of participation status are significantly associated with higher earnings for employees with the public sector. The findings, as a result of discriminant function analysis, indicate that participation undertaken in the mid- 1980s, together with the earlier status of occupation and formal education background, form the first discriminant function that classify occupational status of 1994/5, explaining 54% of the explained variance.

This study investigates the participation patterns of employer-sponsored adult education and training by Swedish workers over the period of 1975 - 1995, and evaluates the importance of the determinants of such provision. The study also explores the ways participation experiences in employer-sponsored education and training influence subsequent participation, occupational mobility and economic outcomes for Swedish workers. Data collected from 1975 to 1995 in The Swedish Living Conditions Survey (ULF), both the cross-sectional data and its panel component, have been analyzed. Contained in the panel are 3,319 Swedish adults who have been followed up in the ULF from 1979 to 1995. The study employs measures of work and job characteristics such as indicators of occupational status, wage, union membership, length of employment, job type, job responsibilities, influence on decision-making at the workplace, learning opportunities at the workplace, enterprise ownership, as well measures of personal characteristics, such as age, gender, level of formal education. The approach taken in the evaluation of the influences of work, job and demographic characteristics on the likelihood of receiving employer-sponsored education and training has been to develop and estimate logistic regression models by means of which these effects during different periods (1975, 1979, 1986/7, 1994/5) can be assessed and compared. Another three models have been investigated, using the panel data, namely: (1) a logistic regression model predicting subsequent participation in employer-sponsored education and training by similar experiences at earlier career stages; (2) a multiple regression model predicting 1994/5 annual income with participation history as a predictor; (3) a discriminant function analytical model predicting 1994/5 occupational status with participation history as a discriminating variable. The departure point of this study is that separate analyses have been done with the public sector and private sector sub-data sets. The findings indicate that occupational status, level of education, age, gender and to a less extent, union membership, and other work and job characteristics are the more important predictors of the likelihood of participation in employersponsored education and training for the Swedish workers in the private sector. For those working in the public sector, institutional factors relating to management style and job responsibility as well as age play an important part in training decisions. The findings also indicate that youth and older age groups, particularly those with the private sector, have been consistently under-served by provision of employer-sponsored education and training throughout the period investigated. The results reveal that by the mid-1990s, gender was not a issue in the public sector in terms of the participation rate but female workers were still disadvantaged in the private sector. The findings verify a trend that the participation gaps between the well-educated and the undereducated, and those between professionals and non-skilled and semi-skilled have narrowed by a great magnitude, yet not adequately to close up the gaps. In 1975, the likelihood of participation for a professional worker was nine times as high as that for a non-skilled worker in the private sector. By 1995, the comparable figure was three times. The findings from panel data analysis show that, for workers in the private sector, their participation status as of the mid-1990s was significantly correlated with their participation status back in the mid-1980s and late 1970s, even after the statistical adjustment. For the public sector employees, their participation status as of the mid-1990s is related to participation status in the mid-1980s only. The results of income estimation models show that all three indicators of participation status are significantly associated with higher earnings, but a further analysis that separates the private sector from the public sector indicates that the result applied more in the case of the private sector employees. In contrast, none of the measures of participation status are significantly associated with higher earnings for employees with the public sector. The findings, as a result of discriminant function analysis, indicate that participation undertaken in the mid- 1980s, together with the earlier status of occupation and formal education background, form the first discriminant function that classify occupational status of 1994/5, explaining 54% of the explained variance. / Education, Faculty of / Educational Studies (EDST), Department of / Graduate