The problem of multiple inference in psychiatric research.

Area of Science:

• Psychiatric Research
• Statistical Inference
• Psychometrics

Background:

• Multiple statistical inferences are common in psychiatric research due to multivariate designs.
• Using multiple univariate tests (e.g., t-tests) presents disadvantages in analyzing multiple dependent variables.
• Addressing multiple inference is crucial for the validity of psychiatric study findings.

Purpose of the Study:

• To outline the disadvantages of using multiple univariate tests in psychiatric research.
• To introduce Bonferroni-adjusted univariate tests and multivariate statistical analysis as solutions.
• To discuss the advantages and disadvantages of these strategies for psychiatric research.

Main Methods:

• Review of common strategies for handling multiple statistical inferences.
• Introduction and discussion of Bonferroni-adjusted univariate tests.
• Introduction and discussion of multivariate statistical analysis.

Main Results:

• Multiple univariate tests can inflate Type I error rates.
• Bonferroni-adjusted tests offer a conservative approach to control family-wise error rate.
• Multivariate statistical analysis provides an alternative for analyzing multiple dependent variables simultaneously.

Conclusions:

• Both Bonferroni-adjusted tests and multivariate analysis are viable strategies for managing multiple inference in psychiatry.
• The choice between methods depends on the research context (confirmatory vs. exploratory) and specific study aims.
• Careful consideration of statistical methods is essential for robust psychiatric research findings.