This year the Subgroup on assessment methods (SAM) requested that WG-FSA consider whether a shorter CPUE time-series would be preferable for use in its annual assessments. (WG-FSA-03/40 2003, paragraph 2.27). This paper explores some of the consequences of shortening the dataset in the application of the usual assessment method. In addition we investigate the incorporation of uncertainty in the standardised CPUE series into the method of Kirkwood and Constable (2000) via Monte Carlo simulations of the CPUE series. Truncation of the standardised series from WG-FSA 2002 results in only slight modification to the median escapement but will generally diminish the probability of depletion as the dataset is truncated towards the more recent years. The use of the standardised series from the GLMM of Candy (2003) shows a more pronounced consequence of reducing the time series, such that the escapement level increases while the depletion probability decreases as the time series is reduced to recent years. In contrast, the explicit incorporation of uncertainties in the CPUE series in the overall assessment shows that truncation of the CPUE series could have a much greater effect on the final assessment, depending on the model used. These results indicate that the procedure for incorporating CPUE into the assessment process needs to be generally evaluated as to the consequences of its use in providing assessments that will meet the objectives of CCAMLR.