A large part of my quantitative and survey methodology training comes from working at large-scale social surveys. 

​
​China Family Panel Studies (CFPS)

Before starting my doctoral training, I spent two years working as a data analyst at the CFPS, a nationally representative and longitudinal survey of Chinese households.​​

• I was mainly in charge of cleaning the 2014 and 2016 child datasets. Key procedures include clarifying valid samples, checking if the data's skip logic is correct, rectifying abnormal values and filling in missing values based on interview records and interviewers' notes etc.
• I also participated in generating variables such as household income, child test scores, and occupational prestige scores.
• Before the data became publicly available, I ran analyses to check data quality and cross-year consistency

​India Human Development Survey (IHDS)

After starting my doctoral training at the University of Maryland, I joined the IHDS team as a research assistant. IHDS is a nationally representative and longitudinal survey of Indian households. 

• I participated in questionnaire design for the third wave of data collection. 
• I also participated in programming and testing the Computer Assisted Personal Interviewing (CAPI) system.
• ​I have worked with IHDS team members on several research projects.

​American Time Use Survey (ATUS) & IPUMS Time Use

While doing my PhD at the University of Maryland, I also worked as a research assistant of the ATUS and IPUMS Time Use. The ATUS is a nationally representative U.S. time diary survey for period since 2003. IPUMS Time Use harmonizes these data and provides a data extract builder that allows users to create custom time use variables and data extracts for analysis.

• I conducted research using the ATUS data in collaboration with ATUS team members.
• ​I participated in data preparation and organization of conferences and workshops.

---

Working for large-scale social surveys has given me special insights when using survey data for analysis. I am particularly interested in how the design of survey questions and the context/environment of conducting interviews would influence data quality. 
I wrote two reports on how the design of survey questions on paid work would influence data quality on labor force participation:

• Stone, Eric, and Xu Yan. 2020. Measurement Brief: Paid Work. Women’s Empowerment: Data for Gender Equality 
• Ji, Yingchun, and Xu Yan. 2020. Measuring Paid Work in China: Lessons Learnt from The China Family Panel Study (CFPS). Women’s Empowerment: Data for Gender Equality.