Studying Individual Patterns of Development Using I-States as Objects Analysis (ISOA)

Abstract

An approach is presented for studying individual pattern development in person‐oriented terms focusing on the concept of i‐state, i.e. an individual's configuration of information at a specific point in time. The procedure is called I‐States as Objects Analysis (ISOA). First common i‐states (typical states) are identified using cluster analysis of subindividuals and then this information is used for describing typical developmental patterns. Both a general procedure and a specific procedure used on a demonstration data set were developed. Using ISOA, change and stability can be studied both with regard to structure and with regard to individual variation. An empirical example was given which concerned longitudinal data about school grades at four different ages for 333 boys and girls. The data were split into a test sample and a replication sample of equal sizes. It was contended from the empirical study that ISOA functioned reasonably well on the sample studied. In the discussion, it was pointed out that ISOA can be a powerful method to use for small samples with many measurement occasions and that the method is optimal for studying short‐term change.

Achievement goal orientations and academic well-being across the transition to upper secondary education

“…At the level of statistical analysis, this means that individuals are studied on the basis of their patterns of individual characteristics rather than on the basis of separate variables (Bergman & Nurmi, 2010). Since we were interested in studying the development of individual patterns across time (for a similar perspective, see Janson & Mathiesen, 2008;Tuominen-Soini et al, 2011), we used the ISOA procedure (I-States as Objects Analysis; Bergman & El-Khouri, 1999).…”

“…I-state is defined as an individual's pattern of values at a specific measurement point in the variables that are to be used for classification. It is assumed in ISOA that approximately the same classification structure applies at all measurement points (although individuals might change the pattern they belong to and the proportion of individuals belonging to the different classes might vary between measurement points) (Bergman & Nurmi, 2010).…”

ROPstat: A general statistical package useful for conducting person-oriented analysis

“…Then more complex classification analyses can be performed, enabling the user to test structural stability (similar clusters at different occasions) and individual stability (individuals belonging to similar clusters at different occasions). Two of these complex classification analysis types are LICUR (Bergman, Magnusson, & El-Khouri, 2003) and ISOA (Bergman & El-Khouri, 1999).…”

“…By means of this method the data record of one case (one row in the data matrix) is divided into separate parts, one for each measurement occasion, creating several rows in a new file. With this transformation of the original data file interesting classification analyses (such as I-states-as-objects-analysis: ISOA, see Bergman & El-Khouri, 1999;or Bergman, Nurmi, & von Eye, 2012) can be performed.…”

Reading performance and its developmental trajectories during the first and the second grade

“…This was done using a clustering by cases analysis. Because our data concerning reading performance included five measurement points and we aimed to yield similar groupings for each measurement point, we used the ISOA procedure (I-States as Objects Analysis) as suggested recently by Bergman and El-Khouri (1999). In this approach the longitudinal data is first organised so that each of the five measurement points for each child (child A Time 1, Time 2, Time 3, Time 4, and Time 5) is coded as a separate individual, e.g.…”