Generalized P-values and confidence intervals: a novel approach for analyzing lognormally distributed exposure data

Area of Science:

• Statistics
• Occupational Health
• Environmental Science

Background:

• Assessing occupational exposure often involves analyzing data that follows a lognormal distribution.
• Traditional methods for inference on the mean of a lognormal distribution can be limited, especially with small sample sizes.

Purpose of the Study:

• To introduce and evaluate novel statistical methods for hypothesis testing and confidence interval construction for the mean of a lognormal distribution.
• To provide a unified and accurate methodology for analyzing parameters associated with lognormal distributions in occupational settings.
• To extend these methods for comparing two lognormal means and for inference on lognormal variances.

Main Methods:

• Application of generalized p-values and generalized confidence intervals for lognormal mean inference.
• Development of power studies and sample size calculation procedures.
• Extension of the methodology to handle comparisons of two lognormal means and lognormal variances.
• Adaptation of the approach for any parametric function involving one or two lognormal distributions.

Main Results:

• The proposed methods demonstrate good performance and are applicable to small sample sizes.
• The generalized p-value approach yields results comparable to existing methods like Land's method.
• The methodology is computationally efficient and easy to implement, with computer program details provided.

Conclusions:

• Generalized p-values and confidence intervals offer an accurate, unified, and practical approach for occupational exposure assessment using lognormal distributions.
• The methods are versatile, extending to various inferential tasks related to lognormal data.
• The industrial hygiene community is encouraged to adopt this robust methodology for improved data analysis.