Interval estimation of the mean response in a log-regression model

Area of Science:

• Statistics
• Biostatistics
• Regression Analysis

Background:

• Skewed response data analysis often involves log-transformation and log-regression.
• Estimating the mean response on the original scale is crucial but challenging with small samples.
• Existing methods based on large sample theory lack precision for small sample sizes.

Purpose of the Study:

• To develop a small-sample corrected interval estimation for the mean response in log-regression models.
• To address the limitations of existing methods when dealing with small sample sizes in skewed data analysis.
• To provide a more accurate and reliable statistical tool for researchers working with skewed data.

Main Methods:

• Utilized a likelihood-based inference method, drawing upon work by Barndorff-Nielson and Fraser et al.
• Developed a novel small-sample corrected interval.
• Employed simulation studies to evaluate the performance of the proposed interval.

Main Results:

• The proposed small-sample corrected interval demonstrated almost exact coverage probability.
• The method proved effective even for very small sample sizes.
• Simulation results validated the accuracy and reliability of the new interval.

Conclusions:

• The developed small-sample corrected interval is a valuable tool for analyzing skewed response data.
• This method offers improved precision for mean response estimation in log-regression with limited data.
• The findings contribute to more robust statistical analyses in fields dealing with skewed data.