A form of the log-linear model for modelling catch per unit effort (c.p.u.e.) is described which treats the year category as an ANOVA classification. This model allows the calculation of two standardized relative abundance indices: 1) direct interpretation of regression coefficients for the year category, and 2) the usual estimates of c.p.u.e. calculated using standardized effort (adjusted c.p.u.e.). The regression coefficient estimates have advantages in theoretical simplicity and computational convenience. The model is applied to two trawl-fished populations of Pacific ocean perch (Sebastes alutus) in the northeast Pacific. Catch and effort data from 1953 to 1976 were used to calculate relative abundance indices standardized for gross tonnage, echo sounders with paper recorders, and roller (bobbin) gear. Stock conditions for the two populations have been historically different, and this factor is shown to interact with the impact of technological improvements. In the Queen Charlotte Sound region, where stocks have been and remain in relatively good condition, technological improvements had little measurable effect on catch rates. However, in the west Vancouver Island region, where stocks have been smaller and are now depleted, these same technological improvements (echo sounders with paper recorders and roller gear) doubled the estimated efficiency of vessels. Due principally to catches by large Japanese and Soviet trawlers, relative abundance indices in both the Charlotte and Vancouver regions registered a sharp decline (50% and greater) in 1965 to 1968. This decline was particularly severe in the Vancouver region. In the Charlotte region, relative abundance measured by regression coefficient estimates and adjusted c.p.u.e.s were nearly identical; but in the Vancouver region, these estimates differed. Considering the methods used in calculation, it was concluded that this difference was probably due to chance events rather than systematic bias. The use of a qualifying criterion for selecting data (deleting data for which the catch is below some fixed percentage of the study species) can greatly affect estimates of relative abundance in multi-species fisheries. In the Charlotte region, a qualifying percentage of 25 % had little effect on relative abundance estimates. But for the Vancouver region, a qualifying percentage of 25 % significantly lessened the estimated decline in abundance. In summary, it appears that there is generally an interaction between the effects of technological improvements in a fishing fleet and the condition of fish stocks. It appears that technological improvements sometimes have their greatest measurable impact on fully utilized or depleted stocks. Also, the selection of qualification levels will sometimes be critical when measuring large declines in stock abundance.