Determining the significance of scale values from multidimensional scaling profile analysis using a resampling...

Area of Science:

• Psychometrics
• Statistical analysis
• Individual differences research

Background:

• Multidimensional scaling (MDS) profile analysis is a common method for studying individual differences.
• Currently, there is a lack of objective methods to assess the statistical significance of scale values derived from MDS.
• This limitation hinders rigorous hypothesis testing in MDS profile analysis.

Purpose of the Study:

• To develop and validate an objective statistical method for evaluating the significance of scale values in MDS profile analysis.
• To introduce a resampling technique to construct confidence limits for MDS scale values.
• To assess the significance of marker variables within MDS profiles.

Main Methods:

• Employed a resampling technique, specifically bootstrapping, to generate confidence limits for MDS scale values.
• Utilized these bootstrap confidence limits to evaluate the statistical significance of marker variables.
• Applied the method to both simulated and real-world datasets.

Main Results:

• The bootstrap method demonstrated validity in constructing confidence limits for MDS scale values.
• The approach proved effective in evaluating the statistical significance of marker variables.
• Analyses of simulation and real data supported the reliability of the bootstrap method.

Conclusions:

• The bootstrap method offers a statistically sound approach for assessing the significance of marker variables in MDS profile analysis.
• This technique addresses the need for objective evaluation in individual differences research using MDS.
• The findings suggest the bootstrap method is a valuable tool for hypothesis testing in MDS.