Specific Problems in the Measurement of Change: Longitudinal Studies, Difference Scores, and Multivariate Analyses. Quantification Laboratory Technical Report No. 27.

Autor/in	Schutz, Robert W.
Titel	Specific Problems in the Measurement of Change: Longitudinal Studies, Difference Scores, and Multivariate Analyses. Quantification Laboratory Technical Report No. 27.
Quelle	(1977), (33 Seiten) PDF als Volltext kostenfreie Datei Verfügbarkeit
Beigaben	Tabellen
Dokumenttyp	gedruckt; online; Monographie
Schlagwörter	Tagungsbericht; Achievement Gains; Analysis of Variance; Attrition (Research Studies); Behavior Change; Cross Sectional Studies; Data Analysis; Longitudinal Studies; Measurement Techniques; Pretesting; Pretests Posttests; Reliability; Research Design; Sampling; Scores; Statistical Analysis; Testing Problems; Validity + Suchen Sie Ihr Suchwort? Achievement gain; Leistungssteigerung; Auswertung; Longitudinal study; Longitudinal method; Longitudinal methods; Längsschnittuntersuchung; Messtechnik; Vortest; Reliabilität; Forschungsdesign; Statistische Analyse; Gültigkeit
Abstract	The measurement of change is such a broad topic that this article must limit its focus to a few specific subtopics. These specific topics include: longitudinal research design, attrition in research studies, the statistical analysis of difference scores, and the comparison of analysis of variance (ANOVA) and multivariate analysis of variance (MANOVA) techniques in analyzing repeated measures data. The purpose and sampling techniques, as well as the internal and external validity are discussed for each measurement technique. The author concludes that a considerable number of problems are inherent in the measurement and analyses of change, especially in research designs of a longitudinal nature. However, most of these problems can be avoided with sufficient care and planning prior to initiating the research project. The cross-sectional sequential type designs which are required for valid measures of developmental change are very costly--but necessary if the research is to have any scientific value. Multivariate statistical procedures utilizing complete data sets will provide for valid and relatively powerful tests of hypotheses. (Author/MV)
Erfasst von	ERIC (Education Resources Information Center), Washington, DC
Update	2004/1/01