The Impact of GSAT Score on Wage: Evidence from Taiwan Education Panel Survey

• Main Authors: Po-Yu Lin, 林柏聿
• Other Authors: 胡登淵
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/p76j98

碩士 === 淡江大學 === 產業經濟學系碩士班 === 104 === Previous studies of human capital and Labor economics discussed the relation between education and salary. This paper investigates the relation between GSAT score and salary for the young people aged 25 years old using quantile regression. The data are drawn from Taiwanese education panel survey, which surveyed from 2001 to 2010. To deal with sample selection problems, the Heckman two-stage estimation is used, and it is extended to solve the problem of multiple sample selection. To correct for the potential endogeneity problems, we use the variable regarding whether or not participated at arts class during the period from the first grade to the first year of high school (for males) and Body Mass Index (BMI) in 11th grade (for females) as instruments. The results show that GSAT score significantly affects salary only from 0.5 to 0.9 quantile. The positive correlation of GSAT score and salary is greater for females than for males. GSAT score is the significant variable in 0.5 to 0.9 quantile for young males and young females, at least, at ten-percent level. The estimates of the coefficient of GSAT score implies that salary increases for young males and young females by more than 0.9% to 1% with every additional score of GSAT. The results imply that GSAT score significantly affects salary only for high-income groups. The quantile regression models include three inverse mills ratios in the estimation, but it is only significant in the lower quantiles of young males. The effects of sample selection for young males are significant but not for young females. Although we control the regarding whether or not participated at arts class during the period from the first grade to the first year of high school for males as instruments. The results show that GSAT score still insignificant in linear model, while he effects of GSAT score for young females are greater than before.