A Review of Integrative Data
Analysis Using Across Independent
Institution：Psychological Application Research
Center, South China Normal University
Though, longitudinal designs and research findings provide the
best basis for understanding the developmental and aging-related
processes, and influences factors across the life span. Researchers
usually encountered a number of significant challenges in
longitudinal studies, including subject attrition, the changing of
measurement structures across groups and developmental periods,
and the need to invest substantial time and money.
Integrative data analysis (IDA) is a set of methodologies that
allow researchers to overcome many of the challenges of single
longitudinal study, when raw data are available. The article drops
the focus on a review of IDA using in pooling longitudinal
studies, discussing both the application and the theory.
2. Principle of IDA
IDA is a set of methodological techniques that can simultaneous
analysis multiple independent samples that have been pooled into
one data set.
IDA makes full use of exiting data to directly test the replication
of findings across independent studies, increase statistical power,
address new questions not answerable by a single study, and
provide opportunity to build more cumulative psychology.
IDA is great useful for experimental studies, developmental
psychological studies, and clinical studies, and can speed up the
understanding of cross-context, cross-age, and cross-culture
variation. There are some studies dropped focus on the principle
and practice of IDA using across longitudinal studies.
3. IDA Using Across independent longitudinal studies
3.1Challenges of IDA using in pooling longitudinal data
When adopting an IDA framework for multiple longitudinal data
integration, there are some challenges must be addressed before
broad adoption by developmental researchers. Key practical
issues associated with heterogeneity between-study.
Mainly sources of heterogeneity between longitudinal studies
include sample characteristic differences, measurement
characteristic differences, design and analysis characteristic
Identifying important sources of heterogeneity between multiple
longitudinal studies is an critical aspect of IDA.
3.2 Analytic strategies
Researchers adopted random-effects IDA or fixed-effects IDA as
general strategy to address the sources of heterogeneity in IDA;
Researchers also developed psychometric models like item
response model and confirmatory factor analysis to attain
commensurate measures across longitudinal studies, these
models can be directly applied within the context of IDA to place
scores on the same measurement scale across studies.
In practice of IDA in longitudinal studies, a two-stage approach
was developed, in which ability estimates for each person at each
occasion are estimated with an item response model. In the
second step, the ability estimates obtained from a first step can be
used as observed data and modeled with growth curve analysis.
(І) IDA simultaneously integrated analysis multiple
developmental data sets, was characterized by a host of
advantages, including greater statistical power for infrequent
event on change; broader psychometric assessment of theoretical
constructs; longer longitudinal windows of study; the opportunity
to test hypotheses not considered in the original studies.
(II) One emphasis of Further research should be focused on issue
of developing common psychometrics and models for IDA.
(Ⅲ) combinational use of IDA and Mata-analysis on pooling
longitudinal data would be a feasible way to reach well
understanding of developmental and aging process.