Assessing background particulate contamination in an historic building - surface lead loading and contamination

Area of Science:

• Environmental Science
• Geochemistry
• Statistical Analysis

Background:

• Investigating suspect surface contamination requires comparative sampling to identify sources, pathways, and dispersal.
• Traditional null hypothesis significance testing (NHST) based on means can be misleading for erratic environmental contaminant data distributions.

Purpose of the Study:

• To evaluate lead surface dust contamination in a historic building using both traditional NHST and randomization/permutation inference.
• To compare the inferential capabilities of NHST and randomization/permutation methods for environmental contaminant data.

Main Methods:

• Collected surface dust samples (n=90) for lead content analysis.
• Applied traditional NHST and randomization/permutation inference, specifically the maximum difference in frequency of detection (Δfd max).

Main Results:

• Areas with lower mean lead concentrations or no significant NHST difference indicated greater contamination via Δfd max.
• Randomization/permutation inference revealed higher probabilities of encountering lead at elevated concentrations in certain areas.
• Conclusions on contamination sources and pathways differed significantly between NHST and randomization/permutation inference.

Conclusions:

• Randomization/permutation inference (Δfd max) is more appropriate than NHST for analyzing environmental contaminant data with erratic distributions.
• Understanding the limitations of NHST and the benefits of alternative methods like permutation/randomization is crucial for environmental professionals.
• This approach is particularly relevant for forensic investigations of building contamination.