Two common CSCL questions regarding analyses of temporal data, such as event sequences, are: (i) What variables are related to event attributes? and (ii) what is the process (or what are the processes) that generated the events? The first question is best answered with statistical methods, the second with stochastic or deterministic process modeling methods. This chapter provides an overview of statistical and stochastic methods of direct relevance to CSCL research. Many of the statistical analyses are integrated into statistical discourse analysis. From the stochastic modeling repertoire, the basic hidden Markov model as well as recent extensions is introduced, ending with dynamic Bayesian models as the current best integration. Looking into the near future, we identify opportunities for a closer alignment of qualitative with quantitative methods for temporal analysis, afforded by developments such as automization of quantitative methods and advances in computational modeling. Copyright © 2021 Springer Nature Switzerland AG.
|Title of host publication||International handbook of computer-supported collaborative learning|
|Editors||Ulrike CRESS, Carolyn ROSÉ, Alyssa Friend WISE, Jun OSHIMA|
|Place of Publication||Cham|
|ISBN (Print)||9783030652937, 9783030652906|
|Publication status||Published - 2021|
CitationChiu, M. M., & Reimann, P. (2021). Statistical and stochastic analysis of sequence data. In U. Cress, C. Rosé, A. F. Wise, & J. Oshima (Eds.), International handbook of computer-supported collaborative learning (pp. 533-550). Cham: Springer.
- Statistical discourse analysis
- Time analysis
- Stochastic models
- Process mining