Abstract. Statistical models of habitat preference and species distribution (e.g. Resource 14Selection Functions and Maximum Entropy approaches) perform a quantitative comparison of 15 the use of space with the availability of all habitats in an animal's environment. However, not all 16 of space is accessible all of the time to all individuals, so availability is, in fact, determined by 17 limitations in animal perception and mobility. Therefore, measuring habitat availability at 18 biologically relevant scales is essential for understanding preference, but herein lies a trade-off: 19 Models fitted at large spatial scales, will tend to average across the responses of different 20 individuals that happen to be in regions with contrasting habitat compositions. We suggest that 21 such models may fail to capture local extremes (hot-spots and cold-spots) in animal usage and 22 call this potential problem, homogenization. In contrast, models fitted at smaller scales, will vary 23 stochastically depending on the particular habitat composition of their narrow spatial 24 neighborhood, and hence fail to describe responses when predicting for different sampling 25instances. This is the now well-documented issue of non-transferability of habitat models. We 26 illustrate this trade-off, using a range of simulated experiments, incorporating variations in 27 environmental gradients, richness and fragmentation. We propose diagnostics for detecting the 28 two issues of homogenization and non-transferability and show that these scale-related 29 symptoms are likely to be more pronounced in highly fragmented or steeply graded landscapes. 30Further, we address these problems, by treating the neighborhood of each cell in the landscape 31 grid as an individual sampling instance (with its own neighborhood), hence allowing coefficients 32 to respond to the local expectations of environmental variables according to a Generalized 33 Functional Response (GFR). Under simulation this approach is consistently better at estimating 34 robust (i.e. transferrable) habitat models at smaller scales, and less susceptible to homogenization 35 at larger scales. At the same time, it represents the first application of a GFR to continuous space 36 3 (rather than multiple, spatially distinct datasets), allowing the predictive advantages of this 37 extension of species distribution models to become available to data from large-scale but single-38 site field studies. 39