We discuss the common mistake of pooling of field ecological data by field ecologists and managers, particularly in the grey literature where peer review is often limited, or when using data collected for a different purpose.  The most common misuse is to pool data from organisms collected across a number of sampling sites, effectively treating the organisms as independent replicates rather than the sampling sites, which are the true replicate, or primary sampling unit. Amongst other common errors is to weight sites in the selection process, such as by using strata to allocate sampling effort, but then combining data collected in sites across different strata.  Examples and guidelines for pooling data are demonstrated using species accumulation curves and fish biomass data from Koondrook forest where 99.6 of the pooled wetland fish biomass in 2013 was from common carp. Yet the true interpretation of native fish biomass for the wetlands in Koondrook forest in 2013 was only 85% alien fish biomass. The principles also apply to temporal sampling, and we demonstrate this with fish data collected from fishways on the Hawkesbury-Nepean  River system where the primary sampling unit is each trap rather than each fish inside the trap. Ideally this talk will start a discussion among the audience members that will lead to more consideration of sampling designs before starting sampling in future projects.