Longitudinal surveys are different from cross-sectional surveys because they examine individuals over time, and collect data regarding trends (Creswell, 2012). If your research question pertains to changes over time, then you might consider implementing a longitudinal survey. Longitudinal studies require surveying your sample at regular points in time, and are often conducted over the course of many years. However, they can also take place over the course of a shorter period of time, like a school semester or a singular year.
There are two types of longitudinal studies: repeated cross-sectional studies and panel studies.
Repeated cross-sectional studies collect data from different samples each time they are implemented (Carr et al., 2018). They can be used if a researcher is not concerned with individual-level change, but rather general trends over time in a specific population or cohort (Rafferty et al., 2015).
Panel studies collect data from the same sample, and revisit the same participants time after time. Because the same individuals are surveyed repeatedly, panel studies are able to track individual-level changes over time (Rafferty et al., 2015).
While cross-sectional surveys can be incredibly useful, longitudinal surveys have deeper analytical potential. By tracking changes on an individual, group, or even population level over time, longitudinal studies can be used as the basis for causal inference. Surveying over multiple time periods allows researchers to disentangle certain compounding effects, such as age/cohort or age/period.
Deakin University. (n.d.). Repeated Cross-Sectional Data Analysis. Deakin University Libguides. https://deakin.libguides.com/quantitative-study-designs/cross-sectional
Ethics Check
Tracking a panel over time requires the researcher to retain personally identifying information. Due to Institutional Review Board requirements, you should be careful to create a system to separate responses from the personal identifiers of those who gave them. You can:
- Separate incoming survey data into two sections: One includes information that could be used to identify someone’s identity, such as birthdate, place of residence, specific job title, etc. The other includes responses to all other questions.
- Create unique identifiers for each response that do not refer to anything specific about each individual, and use the identifier as the “respondent ID” for each response.
- Store responses and identifying information in separate places. You may use different drives, or store responses in the cloud and identifiers on a hard drive or thumb drive.
- Encrypt and protect the folder(s) where the data are stored with passwords.
There are also some drawbacks to using longitudinal studies, as they are both costly and time-consuming. In panel studies, researchers will need to track participants over time, updating changes to names and/or addresses accordingly. These types of studies may also suffer from participant drop off, as over time, individuals may no longer wish to participate. Repeated cross-sectional studies are not affected by respondent drop off, as responses are intentionally taken from different samples each time.
| Pros | Cons |
|---|---|
| Effective for tracking changes in experiences over time | Relatively higher cost to administer |
| Beneficial for examining either changes in same participants or in the general population | Higher potential for bias if participants drop off over time |
| Allow more room for disentangling variables that affect results | Harder to manage logistically because you must follow up with and keep track of participants |
| Allow for stronger inferences based on findings (causes and correlations) | Cannot disentangle all cohort/age or age/period effects |
Regardless of their duration, all surveys take time to plan. No matter which time frame you choose, planning early will improve the quality of your survey.
Equally as important as determining your survey’s time frame is deciding upon your method of data collection. The next few pages will introduce you to the four main modes of questionnaire design to help you determine which is best suited for your research question.
