Sequential Designs

Another type of study design, sequential design or accelerated longitudinal design, attempts to resolve the confounding of age inherent in both cross-sectional and longitudinal designs. The confounding of age refers to the problem of differences attributed to age that are in actuality due to other variables such as the historical period in which the individual was born. In sequential studies, cross-sectional cohorts are assessed longitudinally over a fixed time period, typically a much shorter period of time than with a pure longitudinal design. Therefore, results refer to between- and within-cohort change. Data from this design can be analyzed using the same statistical techniques as longitudinal designs that assess differences in age (e.g., structural equation modeling, including growth curve analysis). Because sequential designs have the advantages of longitudinal and cross-sectional methods, it can be argued that this design is a good solution to many of the problems inherent in using longitudinal or cross-sectional designs alone, especially if the longitudinal component of the study is more than a short period of time. For example, Duncan et al. (1996) found no differences between accelerated longitudinal design and traditional longitudinal design in their ability to predict adolescent alcohol use, growth of alcohol use, and the current status of their alcohol use. Conse quently, this design may be a more efficient method for studies investigating developmental changes.


As probably is now evident, all methods of assessment have strengths and weaknesses (refer to Table 25.1 for a summary of all methods presented), and the method, design, and statistical procedures chosen by researchers should be based on the research question posed and the best ways to test proposed hypotheses. A multimethod approach to the study of development is often advisable, as it incorporates many methods and minimizes errors due to method or design flaws. For example, when examining the development of emotion regulation (the successful, socially appropriate management of negative and positive emotions), one might use adults' reports of children's regulation, behavioral assessments of children's regulation, self-reports of regulation (if the participants are adolescents), and physiological measures of regulatory processes (e.g., vagal tone). Emotional responding can be assessed with similar measures, with coding of facial reactions often being a key element. Which specific measures are selected will vary with the goals of the investigator. Moreover, it should be noted that different types of measures (facial, physiological, self-report) may tap different aspects of emotion regulation (or emotionality). Thus, researchers must provide clear descriptions of their measurement goals and procedures.

One reason measures used in developmental psychology differ so much is that investigators are interested in or choose to analyze different components and correlates of developmental constructs. For example, in regard to regulation, temperament theorists tend to use questionnaire measures that assess dispositional (relatively cross-situational) temperamental components of emotion regulation (e.g., effortful control, impulsivity, reactivity) that allow for generalizations across contexts and situations (although they also use observational measures). This is because they are interested in tapping underlying constitutionally based processes important in emotion regulation (Rothbart et al., 2001).

Morris, Robinson, and Eisenberg Summary of Methods Presented

Was this article helpful?

0 0

Post a comment