Measurements of the thermal conductivity of a pseudo-porous food analogue were performed over a range of porosities and mean pore sizes with the aim of contributing to the current understanding of the influence of porosity on the effective thermal conductivity of foods. Due to the size and non-homogeneous composition of the food-analogue samples, commonly used thermal conductivity measurement devices, such as the guarded hot-plate and thermal conductivity probe, were not suitable. Instead, a comparative method was used which produced results having estimated uncertainties of 2-5%. The porosity was varied between 0 and 0.65 and the mean pore size relative to the size of the sample container was varied between 10(-6) and 4 x 10(-3) (volume basis). The results are presented in a form that would be useful for evaluating effective thermal conductivity models. The mean pore size appeared to have a very small influence on the effective thermal conductivity. (C) 2003 Elsevier Ltd. All rights reserved.