This course will introduce the unique design features of large-scale assessment data and provide guidance in data analysis strategies, including the selection and use of appropriate plausible values, sampling weights, and variance estimation procedures (i.e., jackknife approaches). The course will provide participants with training virtually in analyzing public-use NAEP or TIMSS data files using the R package EdSurvey, which was developed for analyzing national and international large-scale assessment data with complex psychometric and sampling designs. Participants will learn how to perform:   

• data manipulation,
• descriptive statistics
• cross tabulation and plausible value means,
• achievement levels,
• percentiles, and
• linear and logistic regression. 

The knowledge and analytic approach learned from this course can be applied to analyzing other large-scale national and international data with plausible values. This course is designed for individuals in government, universities, private sector, and nonprofit organizations who are interested in learning how to analyze large-scale assessment data with plausible values. Participants should have at least basic knowledge of R software (e.g., took an entry level training on R programming) as well as statistical techniques including statistical inference and multiple regression. Having working knowledge of Item Response Theory and sampling theory would help but not required. Participants need to have a computer preloaded with the latest version of the R and RStudio software to practice the analysis.

Instructors:

Emmanuel Sikali, National Center for Education Statistics (course director)

Paul Bailey, American Institutes for Research

Ting Zhang, American Institutes for Research

Martin Hooper, American Institutes for Research

Michael Lee, American Institutes for Research

Yuqi Liao, American Institutes for Research