Seabed environmental monitoring programmes have been conducted at oil installations in the North Sea for many years. Such studies seek to determine the areal extent of the spread of contaminated cuttings and effects on assemblages of benthic organisms. Survey programmes usually follow a prescriptive programme with respect to the numbers of samples, mesh size and level of taxonomic identification. This thesis examines the effects of altering these survey components on the information content of three commonly used ordination methods, PCA, DCA and CA, with the aim of determining the minimum necessary expenditure to show the same outputs. The use of microbial bioassays was investigated as a potential alternative means of assessing the extent of infaunal disturbance. Minimum number of replicates. When analysing data from a strong environmental gradient, one replicate sample per station provided the same visual information as the total data set of two replicates. The DCA and CA procedures provided outputs which were easily interpretable ecologically, but the PCA biplots were difficult to interpret. Analysis of data from a weak environmental gradient required a higher number of replicates for each technique and outputs were more variable at low replicate numbers. Taxonomic resolution. When infauna were aggregated to the taxonomic level of class, the information content of the three ordination methods was the same as recorded for species level. However, at the taxonomic levels of order and phylum the information on the bi-plots was difficult to interpret. Although this suggests that there exists the potential for reducing the cost of analysis, further corroboration with different data sets would be required. Sieve size. Primary screening of 0.5 mm mesh fauna by a 1.0 mm mesh may be a more cost effective strategy than either a 1.0 mm or 0.5 mm fraction although there was some loss of information on the ordination bi-plots.