The Importance of Thinking Multivariately When Setting Subscale Cutoff Scores

Area of Science:

• Psychometrics
• Statistical modeling
• Multivariate analysis

Background:

• Cutoff scores are commonly used in classification scales.
• Univariate, sequential subscale threshold-setting is standard practice.

Purpose of the Study:

• To quantify the loss of discriminative power from sequential cutoff score setting.
• To highlight the geometric implications of marginal threshold-setting.
• To recommend joint threshold-setting for improved test informativeness.

Main Methods:

• Analytical exposition of threshold-setting properties.
• Geometric interpretation of marginal versus joint thresholding.
• Quantification of discriminative power loss in multidimensional settings.

Main Results:

• Sequential univariate cutoff setting reduces classification power.
• Marginal threshold-setting creates counterintuitive multidimensional geometry.
• Each added subscale further diminishes the test's ability to discriminate.

Conclusions:

• Applied researchers should adopt joint threshold-setting strategies.
• Marginal threshold-setting compromises the validity of classification tests.
• Joint threshold-setting ensures a more powerful and informative classification tool.