What is longitudinal research?

Longitudinal studies collect information repeatedly over time about the same group of people, allowing for analysis of change over time.  Compared to cross-sectional research, longitudinal designs have a number of significant advantages, including allowing for examination of patterns of change and the influence of earlier life circumstances on later outcomes. Longitudinal research can also provide insight into causal mechanisms and processes that is beyond the scope of cross-sectional studies. According to CLOSER, the UK is home the wold’s largest and longest-running portfolio of longitudinal studies, with studies featuring large sample sizes, wide breadths of measures and exceptionally long follow-up times.

Types of longitudinal studies

Longitudinal study design can vary depending on the priorities and goals of researchers, and choosing the appropriate study for your research will depend on your research question and the data available. While longitudinal study designs can vary on a wide range of factors including sampling units, time scale and research focus, four broad types of studies are described here.

• Cohort studies
  Cohort studies follow the lives of a population subgroup that has experienced the same life event during a given time period (a cohort). The most well-known examples of this type of study are birth cohort studies which follow groups of people born within the same time period. The 1958 and 1970 British Birth Cohorts, for example, recruited a random sample from all children born in a particular week in the UK. Examples of non-birth cohorts include people working for a particular organisation at one time, such as the Whitehall Study, or a population studying at school at a given time, such as the Next Steps or REACH studies. Cohort studies are often set up to observe long-term change and individual developmental processes and focus on collecting information about specified individual cohort members. Some long-running cohort studies also collect information about cohort members’ parents, families and children.
  
  
• Household panel surveys
  Rather than focusing on collecting data over time about a sample of individuals, household panel surveys collect information about the household at each wave. Household panel surveys generally collect information about every member of a household as well as information about the household itself, such as the type of dwelling, housing costs and the relationship between the household members. These studies generally aim to remain representative of the population of households as a whole, requiring a more complicated design than cohort studies, and often include complex following-rules so that new entrants, individuals and families can be incorporated into the sample at each wave. Understanding Society, The UK Household Longitudinal Survey is an example of a household panel survey.
  
  
• Repeated cross-sectional studies
  Some cross-sectional studies are repeated at regular intervals and include a large number of repeat questions at each data collection point. While studies of this type are not longitudinal because they collect data from different samples at each time point, the use of repeated measures allows researchers to examine population or group change over time. For this reason, a number of repeat cross-sectional studies are included in the catalogue. The recruitment of a new sample at each data collection point allows these studies to observe a representative sample at each wave. If representative samples are present in multiple waves, population changes in behaviour or circumstances can be examined using repeated cross-sectional data. Repeated cross-sectional surveys cannot be used to look at individual change, however. The European Social Survey, which biennially recruits and interviews fresh representative samples from over 20 European countries, is an example of a repeated cross-sectional survey.
  
  
• Accelerated cohort studies
  Most age-based cohort studies study just a single cohort of individuals over time. In contrast, accelerated longitudinal designs recruit multiple cohorts, with each cohort entering the study at a different age. This approach has the advantage of allowing a study to span the age range researchers are interested in within a shorter period of time than would be possible using a single cohort design. As each participant’s measurement schedule covers only part of the age range of interest, a disadvantage of this approach is that data will not be available for all participants at all ages.


• Linked cohort studies
  A dataset composed of routinely collected or administrative data, such as medical or census records. The data relating to a group of individuals are linked together to follow them over time. An example of a linked cohort on the Catalogue is the Social and Economic Predictors of the Severe Mental Disorders Study (SEP-MD), which connects individual census data to clinical records and death registrations.

For more information on using longitudinal data visit the CLOSER learning hub or our Online methods training page.

The information provided on this page is drawn from P.J. Lavrakas (Ed.). (2008) Encyclopedia of survey research methods. Thousand Oaks, CA: SAGE