There are numerous ways in which fisheries catch data can fail to reflect community shifts in ecosystems; two major factors are:

• The catch data refer to a single, highly selective fishery, and not to all fisheries exploiting the ecosystem in question;
• The catch data used do not all stem from the same ecosystem.

These two factors, and some lesser issues are relatively easy to account for, and hence ‘fishing down’ unmasked (see chapter 2 in Pauly, 2010). In fact, most of the recent controversy on the existence of ‘fishing down’ is due to colleagues failing to account for (i) and/or (ii) (see also: Nature of the Discord).

Particularly, the spatial expansion of fisheries (quantified in Swartz et al., 2010), when not compensated for, will often cause time series of trophic level to be based on catch data from different ecosystems, thus rendering the analysis invalid (see [ii] above). Various approaches exist for dealing with this sort of bias, and some are presented in the case studies listed on this website.