Cross-Sectional Design

Conceptual Overview and Discussion

The broad meaning of cross-sectional design—data collected at one time period—can present some definitional challenges. The distinction between data collected at one time period and data collected over time can be a difficult one. The definition of a single time period can vary from a single visit to a site on one day, to visits conducted over a whole year or more. In the case of longer time periods, the distinction is primarily in the intent of the study. If change over time is an aspect of the study—for example, if changes in nursing practices are explored over one year as a hospital implements new policies—then the study is employing a longitudinal design. If, on the other hand, the intent is to explore nurses' reactions to the variety of cases in a hospital emergency room during one year, not the changes in their reactions, the study would be considered primarily cross-sectional.

From this broad perspective, cross-sectional design is applicable to all types of case studies. Using Robert Yin's classifications, cross-sectional designs can be applied to both single and multiple case studies, including those that are holistic and embedded. For example, a single holistic case study of a marketing firm would be considered cross-sectional if it is conducted over one defined time period and includes a cross-section of employees from all levels in the firm. This approach can be extended to single embedded case studies and all forms of multiple case studies. In specific reference to the multiple case study, John Gerring uses the term cross-sectional design to refer to situations where many cases are being used for cross-case analysis without temporal variation.

From a developmental perspective, cross-sectional case studies should include participants selected from different stages in their trajectories (e.g., ages, grades, or years of experience). In a single embedded case study this could take the form of a whole-school case with embedded groups at each grade level. In a multiple case study this could be accomplished through defining each class as a case and choosing classes from different grade levels.

In these developmental situations, the cross-sectional design is often used as a substitute for a longitudinal design. Rather than comparing the same students, for example, as they change over time, changes are explored by looking at groups of students of different ages. The primary reasons for conducting a cross-sectional study rather than a longitudinal one are the savings in cost and time associated with collecting data. The drawbacks, however, include the inability to distinguish between true age-related changes and age differences due to other factors, including cohort factors, and the inability to understand individual change. Individuals may still be defined as the unit of analysis, but individual change cannot be analyzed.

...