The longitudinal design within developmental research – a long list of problems

by Research Methods & Statistics

The longitudinal research design refers to measuring the same variable in the same sample of individuals throughout a period of time. The participants are repeatedly observed over a period of time, this can be fairly brief for example, 6 months until a year or it can prolong lasting the individual’s entire life span progressing from infancy, throughout childhood and adolescence and finally into adulthood.

The individuals being studied are typically cohorts (born at the same time with similar circumstances). Researchers are investigating one or many aspects of human development. A single, or several, measurements of a variable is made in the same set of individuals at several different points in their life in order to investigate the link between age and the particular variable. A longitudinal study is a research example of a within-subjects design. No treatment is administrated; instead the treatment is the age. Longitudinal studies are often described as a number of observations followed by a period of aging and development then another set of observations.

As with all aspects of research methods there are several advantages and disadvantages associated with longitudinal studies. Firstly, the advantages include that it is possible for researchers to assess the stability and continuity of several attributes in a sample by repeatedly investigating and testing the same participants (Kagan & Moss, 1962).1 Secondly, it is possible for researchers to identify and establish normative developmental trends and processes by searching for commonalities such as the point at which children and adolescences experience and undergo and series of changes (Newman et al, 1997).2 Additionally, a greater understanding of individual differences within developmental psychology will be established by tracking several participants overtime. Additionally, there are no cohort effects as the researchers focuses on examining on group of people throughout a period of time rather than making comparisons between groups with different ages. Finally  this research combines both quantitative and qualitative research traditions building a bridge between the concepts of the two types of data (Ruspini 1999).*

There have been many successful longitudinal studies within developmental psychology, especially within the area of early social development (Howes and Matheson, 19923; Rutter et al, 20074). For example, Rutter et al’s longitudinal study into the effects of institutionalisation on emotional development. Tizard and Hodges (1989)5 identified an unusual pattern of behaviour they referred to as disinhibited attachment in children raised without attachments in institutions. Rutter et al examined this particular pattern of behaviour in their longitudinal study of Romanian orphans adopted by families in the United Kingdom. Researchers compared the development of Romanian orphans adopted by UK families and UK children who were also adopted. The Romanian children had been put in orphanages with poor living conditions when one-to-two weeks old. The children were brought to the UK and

  • Fifty-eight were adopted before they were 6 months old
  • Fifty-nine were adopted between the age of 6 months and 24 months old
  • Forty-eight were adopted between the ages of 2 and 4 years and were so called the late adopted group

These were the naturally occurring independent variables. The children were followed up and investigated at several points of their childhood such as at the ages of 4, 6 and 11. The evidence was gathered through a collection of research method procedures such as interviews, questionnaires and observations. Rutter et al found evidence of disinhibited attachment in the children at the age of 6 and this was most common in the late adopted Romanian orphan group. This behaviour type was rare in UK children and those Romanian orphans adopted at an earlier age. When reviewed again at the age of eleven, the late adopted Romanian orphan group still showed signs of disinhibited attachment also, many of the children with disinhibited attachment were also receiving special educational help or mental health services.

Specific to this experiment, a longitudinal study design allowed researchers to keep a track and progress on the development of disinhibited attachment in the different groups of children, in that, they could check for consistency as the children grew older. Also, the longitudinal study allowed for a wide range of measurements of the individuals such as questionnaires, interviews and observations. This type of research design aided in increasing psychological understanding of the effects on social development and highlighted the need for interventions to stop children developing this type of attachment.

For the majority of the blog so far longitudinal studies can be seen in a favourable light. However, although the portrayal of this research design so far has been positive it is incredibly important to remember the limitations of using such a research design and the disadvantages associated with this particular type of methodology. The obvious limitation of using this type of study is that they are very expensive as the researchers have to track down and persuade people to come back for further testing when necessary. Also, if the experiment ends up lasting many years then this adds further costs as researchers repeatedly need training in order to conduct the study. The longitudinal design is extremely time consuming for both the researcher (as it requires a huge amount of commitment to participate and continue throughout the study) and the experimenter (the researcher must stay interested in the research for many years to see the final results). Furthermore, and more specific to the developmental psychology, the focus and theories within the field of developmental science are constantly changing and therefore the research questions at the beginning of a longitudinal study may seem exciting however, by the end of the research they become somewhat trivial.

There is a cluster of limitations surrounding the longitudinal research strategy; these include what is known in research as the practice effect, selective attrition, a non-representative sample and the cross generational problem. Practice effects threaten the validity of a longitudinal study. As longitudinal studies include being interviewed, questioned and observed repeatedly over a long period of time the participants may become test wise or more familiar with the test therefore show performance improvements that are not related to normal patterns of development. Another significant problem researcher’s face when using longitudinal studies is selective attrition. The design is subject to high dropout rates due to a number of potential reasons. The children being investigated, over time may move away or become uninterested and become bored with being tested and therefore decide to withdraw from participation. Parents may also decide they do not wish for their child to participate any longer and force them to withdraw. There is also a possibility that the participants may die during the time span of the study. The end result of participants withdrawing from the experiment may lead to a smaller and possibly a non-representative sample. This sample will provide the researchers with significantly less information regarding the developmental questions being researched. A non-representative sample may also limit the final conclusions of the research specific to those children who did not move away and who co-operated. Those participants who dropped out may have been systematically different to those group members that stayed.

Another limitation of the longitudinal study design is referred to as the cross generational problem. The children in the project are usually selected from one cohert and are very likely to have different experiences to other children from other generations or eras. For example, in today’s modern society there is a huge increase in the number of children who attend different forms of day care such as nursery. The development of children and child rearing practices has dramatically changed from older eras therefore we cannot be certain that children from different generations developed in similar or the same ways.

To conclude, whilst there have been many noteworthy longitudinal studies following the development of individuals that have established and assessed several aspects of development the host of limitations surrounding this research method such as practice effects, selective attrition, non-representative samples and the cross-generational problem make us question the validity of such research.

On a final note, the best features of a longitudinal design and another research method known as the cross-sectional design have been combined together to advance developmental research into a study called the sequential design. This is what we shall be discussing next week…