Abstract: Using the Chinese General Social Survey data in 2017, combined with empirical analysis methods such as Probit model, Random forest model and extended regression model, the impact and mechanism of Internet use on rural females’ non-agricultural employment decisions in the digital era were analyzed. The results showed that Internet use could promote non-agricultural employment of rural females, and the marginal effect was 6.8%; the higher the frequency of modern media use, the higher the probability of non-agricultural employment of rural females, especially the impact on self-employment was greater than employment; the influence of information access channel was not significant. Heterogeneity analysis showed that Internet use had a higher probability of increasing non-agricultural employment for rural females with junior middle school education level, low family care intensity and low family economic level. Internet use might affect non-agriculture employment through three channels of increasing human capital level and experience accumulation, extending social relations and improving the traditional gender perception. It showed that in the digital economy era, the use of the Internet played a positive role in the participation of rural females in non-agricultural employment, and also provided a new perspective for exploring the digital dividend under the existing labor constraints, optimizing the labor allocation of resources, and promoting the sound and rapid economic development.

Key words: digital economy, Internet use, rural females, non-agricultural employment, Random forest, influence mechanism