What is Longitudinal Research?

I teach Research Methods at a university, and I find that students are often confused by two different developmental methods: longitudinal and cross-sectional designs. However, a few simple explanations can help students to dinstinguish between the two. Here goes:

In cross-sectional designs, there are at least two groups involved, but there is only one time point. Let’s say I want to see if ten-year-olds are more friendly than fifteen-year-olds. In a cross-sectional design, I have a group of ten-year-olds and a group of fifteen-year-olds. I test their levels of friendliness at one point in time, and I compare the results to see who is more friendly.

Simple enough, right? But there’s a significant disadvantage here; I don’t know if people become more or less friendly over time, or if there is something in the history of each cohort that has made it more or less friendly. For example, older people are more frugal than younger people. Do people become more frugal as they age? Probably not. Most likely older people are more frugal because of the era they grew up in, not because age makes us more frugal. However, a cross-sectional study cannot distingish what we call cohort effects and age related changes. This is the MAJOR weakness of cross-sectional research. (Remember that–it will probably be on your test, especially if you are in my class!)

A longitudinal study involves one group of people and at least two time points. If I compare ten-year-olds and fifteen-year-olds on friendliness in a longitudinal study, I would take a group of ten-year-olds and test their loneliness. Then I would wait five years and test them again when they are fifteen. This is a longitudinal study because I’m only looking at one group of people, but I am looking at them at more the one time point. If I find that fifteen-year-olds are more friendly, I know that the group has become more friendly as they aged. I can say that this is an age related change rather than a cohort effect.

In many ways, a longitudinal study is superior to a cross-sectional study. However, cross-sectional studies are quite common. Why? With a cross-sectional study, you have results right away. You can finish that thesis or publish that article. Because cross-sectional only involves one data collection point, it’s also cheaper. You also save money by not having to track people across time.

A cross-sectional study also minimizes attrition. In longitudinal studies, people may drop out across time. If you test people at five year intervals, it’s likely your sample will get smaller every five years. People may die, move, or simply not want to participate any longer.

In summary, remember that cross-sectional designs involve two or more groups of people at one time point. Longitudinal designs involve one group of people at more than one time point. If you can keep this in mind, you are well on your way to understanding the differences between these two developmental designs.