This paper approaches issues related with frame problems and nonresponse in surveys. These nonsampling errors affect the accuracy of the estimates whereas the estimators become biased and less precise. We analyse some estimation methods that deal with those problems, in the design-based perspective, and give an especial focus to the poststratification procedures. For complex sampling designs the theoretical properties of the estimators need further research. We then address the Bootstrap methodology for variance estimation. Some applications of the poststratification estimators and the Without Replacement Bootstrap (BWO) algorithm, proposed by Sitter (1992), are also presented, using data from the 1997 Annual Business Survey, conducted by Portugal’s National Statistics Institute. The precision of the analysed estimators is discussed and some recommendations are made regarding their applications under this survey.

keywords: poststratification; frame problems; nonresponse; reweighting; adjustment methods; Bootstrap.