Panel Data

Changyong Feng & Wan Tang & Xin Tu

Panel data, also known as longitudinal data, are important in many areas of research, including epidemiology, psychology, sociology, economics, and public health. Data from longitudinal studies in clinical trials and cohort studies with long-term follow-ups are a primary example of panel data. Unlike data from traditional cross-sectional studies, panel data consist of multiple snapshots or panels of a study group or a cohort of subjects over time and, thus, provide a unique opportunity to study changes in outcomes of interest over time, causal effects, and disease progression, in addition to providing more power for assessing treatment differences and associations of different outcomes. Such data also present many methodological challenges in study designs and data analyses, the most prominent being correlated responses and missing data. As a result, classic models for cross-sectional data analysis such as multiple linear and logistic regressions do not apply to panel data. In cross-sectional studies, observations ...

Users without subscription are not able to see the full content on this title. Please, subscribe or login to access all content on this website.