The statistical properties of the particle flux in different types of plasma turbulence models are numerically investigated using probability distribution functions (PDFs). The physics included in the models range from two-dimensional drift wave turbulence to three-dimensional MHD dynamics. The PDFs of the flux-surface averaged transport are generally in good agreement with extreme value distributions (EVDs). This is interpreted as a signature of strong vortical structures dominating the transport on individual flux-surfaces. Although the transport PDFs are strongly non-Gaussian, no algebraic tails in the distributions are observed. The tails are, in all cases considered, in excellent agreement with EVDs having exponential tails. (C) 2003 Elsevier B.V. All rights reserved.