Home > Department of Media and Communications > Research > EU Kids Online > Best Practice Guide > FAQ 8: When is it best to use a longitudinal design?


FAQ 8: When is it best to use a longitudinal design?


Written by Ingrid Paus-Hasebrink, Philip Sinner and Fabian Prochazka, Austria

What's the issue?

The main aim of longitudinal studies is to analyse change over time. Childhood is about change; research on children is about development and socialisation processes. Therefore it seems to be necessary to use research designs which are able to describe individual changes within and beyond single life spans. In principle, cross-sectional designs are able to provide at least some evidence on changes when they ask for retrospective information. However most of them are limited to descriptions of the status quo.

Common practice

True longitudinal studies rely on panel data and panel methods where the same individuals are measured on more than one occasion checking the same variables. An alternative is an omnibus panel where the information collected varies from one point in time to another. Another alternative is the cohort study where people who belong to the same cohort are measured on more than one occasion.

Questions to consider

Studies relying on either true longitudinal design or repeated measures of similar groups seem to be quite rare in the field of media studies. A thorough overview of studies on children's use of online media in 18 European countries between 1999 and 2006 for example found only two examples of a longitudinal study (Staksrud, Livingstone, & Haddon, 2007). This is probably mostly due to the fact that these studies are often more complex and more expensive than cross-sectional studies.

It is recommended that research projects using repeated surveys as a method for measuring social change should aim at keeping changes in the research design between surveys to an absolute minimum. Duncan (Duncan, 1969) laid down this principle in simple terms by pointing out that "if you want to measure change, don't change the measure". This is perhaps one of the reasons why longitudinal designs are so little used for media research as it is very difficult to adhere strictly to this principle in studies where the nature of the object of study is constantly changing. This problem is especially evident when the time span of a research project stretches over several decades. Then the ideal of standardization will eventually come into conflict with the need to collect meaningful information from the respondents or participants in the study.

Further resources

Swedish Media Panel project (http://www.ssd.gu.se|)

Further reading
Bryman, A. (2004). Social research methods (2nd ed.). Oxford: Oxford University Press.
Peter, J., & Valkenburg, P. M. (in press). Adolescents' exposure to sexually explicit internet material and sexual preoccupancy: A three-wave panel study. Media Psychology.
Valkenburg, P. M., & Peter, J. (in press). The effects of Instant Messaging on the quality of adolescents' existing friendships: A longitudinal study. Journal of Communication.

Example of good practice: Swedish Media Panel project

An example of true longitudinal research is the Swedish Media Panel project (see http://www.ssd.gu.se). Founded by K. E. Rosengren and S. Windahl in 1975, it is a long term research programme focused on basic aspects of the use of mass media by Swedish children, adolescents and young adults, as well as on the causes, consequences and effects of that media use. Since 1995, the programme was directed by Ulla Johnsson-Smaragdi.

During a long period of continuous research the MPP group has produced a data bank in which a large mass of data related to individual media use, its causes, effects and consequences are stored, covering a number of cohorts and panels of children and adolescents passing through the school system and into work or continued studies during their early adulthood. In all, the bank contains data about: some 4400 children, adolescents and young adults; their family background, activities and relations; their relations to peers and their school experiences (including school grades, etc); their media use, life styles, present occupation and activities, as well as their plans for the future. Relevant data from their parents have also been collected on several occasions.
(Project summary, Swedish Social Sciences Data Service)

Example of good practice: Children and Television in Iceland

An example of a long term research project on children and media use is the Children and Television in Iceland study, in which information on media use for children aged 10-15 years has been recorded regularly since 1968, thus enabling comparison over time.

A researcher's experience: Socialisation and change in research with children

In order to really deal with socialisation processes, the dynamic character of the socialising factors, which determine how adolescents select media and acquire symbols useful for their daily lives over a long term period, has to be taken into account. Various research studies point out that children employ and assign significance to media depending on their socio-cultural conditions such as the societal stratum, the education level, the family form, the place and size of residence and the parental income (Austin, 1993; Livingstone & Bovill, 2001; Messaris, 1983; Warren, 2003). However, the media socialisation of children is influenced not only by objective, socio-economic conditions, but also by personal and interaction related processes, such as diverse family lifestyles, different forms of family, and the position of children within their peer-groups that determine the ways in which media content is acquired. All these factors are subject of change themselves. Thus, in order to get valid information on developmental changes, longitudinal designs are necessary. These designs enable us to draw a picture of the socialisation process of children and the role that media play in their lives (Paus-Hasebrink & Bichler, 2005; Paus-Hasebrink & Bichler, 2008). (Ingrid Paus-Hasebrink, Austria)