Median and quantile tests under complex survey design using SAS and R

Area of Science:

• Statistics
• Survey Methodology
• Biostatistics

Background:

• Hypothesis testing on medians and quantiles is crucial for subgroup analysis in complex survey data.
• Existing methods for such tests under complex survey designs are limited.
• Accurate statistical analysis is vital for public health research using large datasets.

Purpose of the Study:

• To introduce novel SAS and R programs for hypothesis testing on medians and quantiles.
• To provide detailed illustrations of computations and program usage.
• To facilitate advanced statistical analysis for complex survey data.

Main Methods:

• Development of new statistical programs in SAS and R.
• Implementation of hypothesis testing for medians and quantiles.
• Utilizing complex survey data, specifically the National Health and Nutrition Examination Survey (NHANES).

Main Results:

• Successful implementation of SAS and R programs for median and quantile hypothesis testing.
• Demonstrated application using urinary iodine data from NHANES.
• Facilitated comparisons of medians (females vs. males) and 75th percentiles (salt consumption groups).

Conclusions:

• The developed SAS and R programs offer effective tools for hypothesis testing on quantiles in complex surveys.
• These programs enhance the ability to analyze subgroup differences in large-scale health surveys.
• The methodology provides a valuable resource for researchers working with complex survey data.