Analysis of data in idiographic research: Issues and methods

Time-series and single-subject designs are being advocated for use in rehabilitation research to document the effectiveness of therapeutic intervention procedures. The argument is made that these designs are practitioner-oriented and practice-based. Acceptance and application of single-subject procedures by rehabilitation researchers will depend on the development of reliable and accurate methods of data analysis. Previous research suggests that traditional visual analysis of graphed data is unreliable and that quantitative supplements to visual analysis should be included in single-subject investigations. The advantages and limitations of two commonly advocated supplements to visual analysis are examined. The two supplemental procedures are the split-middle method of trend estimation and the resistant trend line. One advantage of the resistant trend line is that it does not make the assumption of a linear pattern in single-subject data. The argument is made that the resistant trend line is the preferred method and should be selected over the more commonly used split-middle method of trend estimation. The advantages and limitations of other quantitative procedures are briefly examined.