In low temperature argon arc plasma, emission spectral lines of different analytes have characteristic radial distributions with emission maxima distributed over a wide plasma region, from the arc axis toward the arc periphery. It was assumed that the occurrence of emission maxima is a result of collective action of factors that control vaporization, dissociation, excitation, ionization and analyte transport. A statistical analysis was used to reveal a correlation between these various factors and a radial position of maximum emission. A regression analysis was used both for predictive and explanatory purposes: a model was built that successfully predicted analyte spatial emission characteristics, based on their physical properties and local plasma parameters, and a relative importance of proposed predictors was estimated.