Regardless of the type of research you conduct, under a cross-sectional design you collect data over a short and fixed period of time. In relation to survey research, Dr. Ruth A. Palmquist at the University of Texas notes that "Cross-sectional surveys are used to gather information on a population at a single point in time."
For instance, if you conduct a survey asking people their opinion of the President, you only know how that population feels when you asked; you have no information regarding how attitude or opinion changed over time. In other types of work, such as medical research, you assess "outcomes and exposures... at a moment in time, without either forward or backward timing," as noted at the Yale University School of Medicine website.
In order to understand cross-sectional research design fully, it helps to look at another common research design. Longitudinal research, for example, allows you to gather data from your population over the course of time. As Professor Michelle A. Saint-Germain at California State University-Long Beach (CSULB) explains, a longitudinal design allows you "to measure change in variables over time." If you want to know how opinion of the President changed over time, you might sample your population just after the President was elected and at one or more points thereafter.
Professor Saint-Germain lists several advantages of cross-sectional research. She claims that it allows you to collect large amounts of data from a large number of people on a wide variety subjects. These features often mean that the data is of use to researchers from various disciplines. Saint-Germain also states that cross-sectional research works well for exploratory studies. This makes it useful for students who have yet to develop the skills or obtain the time and resources necessary to execute more sophisticated research designs. Exploratory research is also a good way to test a research topic prior to investing large amounts of time and money in it.
While a basic cross-sectional research design can be cost-effective, Saint Germain warns that as the size of your sample or the number of variables increases, so does the cost of executing your project. In addition, cross-sectional research does not allow you assess "facts about 'time order' of exposure and effect, in other words whether one preceded the other," according to Yale's website. Simply put, you cannot uncover details about "cause and effect," according to Saint Germain.
As Saint-Germain points out, the U.S. Census is an example of cross-sectional research. It reflects the composition of the American population at one moment in time. A massive undertaking, the Census aims to survey the entire population of the country. You likely do not have time or the millions of dollars needed to conduct a census; however, obtaining a sample from your population of concern can yield useful and accurate results, if part of an overall well-designed research study.