Viability of Analysis of Variance Applied to the Data Obtained From the Atomic Absorption Spectrometric Determination of the Metal Content in Welding Fumes

Abstract

The conditions in a shipyard that could affect the metal content in welding fumes were studied. Personal samples were collected on cellulose filters in compliance with standard methods. The total metallic particles content was determined by gravimetry. The filters were then dissolved by acid digestion and the metals (Fe, Mn, Zn, Cu, Cr, Ni and Cd) were determined either by flame or electrothermal AAS. A detailed statistical study was performed. An initial exploratory data analysis based on the outlier determination was made, followed by a descriptive data analysis. Finally, an analysis of the variance was applied to study nine factors that could affect the metal content. In all cases validation of the model was accomplished by checking the basic hypothesis.