This study explores the bidirectional causality between fertility preferences and outcomes in China, using the latest two waves of the Chinese General Social Survey (CGSS 2018 and 2021). The findings highlight significant implications for two contrasting phenomena observed in the data: 1) A progressive shift from son preference to daughter preference overall across cohorts; 2) A persistent gender discrepancy between fertility outcomes and preferences - actual fertility outcomes are more skewed towards boys than ideal fertility preferences suggest. The shift towards the daughter preference can be attributed to the “girl-updating” mechanism: having girls on average reduces the gender gap in preferences by approximately 0.5 children, while the effects of having boys are less consistent. On the other hand, the gender discrepancy between fertility outcomes and preferences can be explained by two behavioral patterns: 18% of the discrepancy comes from people with a son preference taking more aggressive measures to ensure male offsprings, whereas 82% is attributed to the “implicit son preference” among people with a balance preference, who opt for a boy when limited to only one child. Moreover, a dynamic Difference-in-Difference study suggests that during the early stage of birth control policies, urban residents with a balance preference suffer from increasingly stricter constraints of having only one child than their rural counterparts, exacerbating the skewed sex ratio.