Advertisement

Single Case Studies: The Time Series vs. the Smoothing Regression Approach

  • Michael G. Schimek
Chapter
Part of the Recent Research in Psychology book series (PSYCHOLOGY)

Abstract

Single case studies are usually evaluated by time series intervention analysis, a well developed parametric approach. This paper discusses the most advanced technique, theoretically speaking, of Box and Tiao (1975), pointing out conceptual problems and limitations. As an alternative, a new non-parametric approach is introduced for trend estimation, an approach which makes it possible to accommodate linear stochastic noise processes from the Box and Jenkins (1976) methodology. It is called Dependent Error Regression Smoothing (DERS; Schimek, 1992, 1994). Conceptual differences from parametric time series intervention analysis are explored. An example is cited from clinical psychology in which parametric approaches fail to provide any evidence for an intervention effect on an observational series. The example data set is re analysed in the non-parametric framework of DERS, and the intervention effect under consideration can then be identified.

Keywords

AR ARMA intervention analysis intervention function MA non-parametric regression penalized least squares prefiltering pseudo-variance single case study smoothing parameter smoothing spline time series trend estimation 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Box, G.E.P., & Jenkins, G.M. (1976, 2nd edition). Time series analysis. Forecasting and control. Oakland, CA: Holden-Day.Google Scholar
  2. Box, G.E.P., & Tiao, G.C. (1975). Intervention analysis with applications to economic and environmental problems. Journal of the American Statistical Association, 70, 70–79.CrossRefGoogle Scholar
  3. Chatfield, C., & Schimek, M.G. (1987). An example of model-formulation using IDA. Statistician, 36, 357–363.CrossRefGoogle Scholar
  4. Glass, G.V., Willson, V.L., & Gottman, J.M. (1975). Design and analysis of time-series experiments. Boulder, CO: Colorado Univ. Press.Google Scholar
  5. Gottman, J.M., & Glass, G.V. (1978). Analysis of interrupted time-series experiments. In T.R. Kratochwill (Ed.), Single subject research. Strategies for evaluating change (pp. 197–235). New York: Academic Press.Google Scholar
  6. Gregson, R.A.M. (1983). Time series in psychology. Hillsdale, NJ: Lawrence Erlbaum.Google Scholar
  7. Kratochwill, T.R. (1978). Single subject research. Strategies for evaluating change. New York: Academic Press.Google Scholar
  8. McCleary, R., & Hay, R.A. (1980). Applied time-series analysis for the social sciences. Beverly Hills, CA: Sage.Google Scholar
  9. Schimek, M.G. (1988). How to differentiate between impact and effect using Box and Tiao intervention functions: Some zero and first order cases. EDV in Medizin und Biologie, 19, 49–57.Google Scholar
  10. Schimek, M.G. (1992). The DERS algorithm in comparison with related spline smoothing algorithms: Numerical aspects and time complexity. In F. Faulbaum (Ed.), SOFTSTAT’91. Advances in Statistical Software 3 (pp. 267–274). Stuttgart: Fischer.Google Scholar
  11. Schimek, M.G. (1994). Penalized least squares regression smoothing with dependent errors. (Submitted for publication.)Google Scholar
  12. Schimek M.G., & Schmaranz, K. (1993). Dependent error regression smoothing: Technique, PC program and application. (Submitted for publication.)Google Scholar
  13. Silverman, B.W. (1985). Some aspects of the spline smoothing approach to non-parametric regression curve fitting (with discussion). Journal of the Royal Statistical Society, B, 47, 1–52.Google Scholar

Copyright information

© Springer-Verlag New York, Inc. 1994

Authors and Affiliations

  • Michael G. Schimek
    • 1
  1. 1.Medical Biometrics GroupUniversity of Graz Medical SchoolsGrazAustria

Personalised recommendations