Nearly all quantitative analyses in higher education draw from incomplete datasets-a common problem with no universal solution. In the first part of this paper, we explain why missing data matter and outline the advantages and disadvantages of six common methods for handling missing data. Next, we analyze real-world data from 5,905 students across…

As the United States becomes more racially and ethnically diverse and draws more students from across the globe, more representative data are needed to understand at-risk and underrepresented populations in higher education, particularly in the science, technology, engineering, and mathematics (STEM) fields. The authors argue that the current…

The purpose of this paper is to review recent developments related to the development of indicators of educational progress in the context of the Post 2015 deliberations to generate a new international architecture for educational investment through to 2030. There have been a plethora of suggestions and several parallel consultation processes…

Many do not consider the effect that missing data have on their survey results nor do they know how to handle missing data. This chapter offers strategies for handling item-missing data and provides a practical example of how these strategies may affect results. The chapter concludes with recommendations for preventing and dealing with missing…

This study examines the gaps related to gender and race/ethnicity in entrance, persistence, and attainment of postsecondary science and engineering (S&E) education. After reviewing selected prior research and examining potentially relevant variables in two National Center for Education Statistics (NCES) surveys, several variables were selected…