Farmers who dropped out of the survey in the first section of thequestionnaire were treated as non-respondents and their informa-tion was used in the analysis of unit non-response bias. On the other hand data from farmers who dropped out of the survey in thesubsequent sections were used to test for item non-response bias(De Leeuw et al., 2003). The tests indicated non-respondents weremainly older farmers, thereby resulting in an over-representationof young farmers. However, farmers with fully completed and par-tially completed questionnaires were not statistically different.Although the over-representation of younger farmers could biasthe outcome of our analysis, the problem may only be minimalgiven the current trends of a decrease in the number of older farm-ers in Denmark (European Commission, 2012).The study thereforeutilized the 267 complete cases.
Based on four key questions (Appendix A), the analysis was con-ducted in two stages. Firstly a descriptive analysis of the generalrespondents’ perception of water quality, attitudes regarding theeffects of pollution reduction measures and the preferences for var-ious nutrients mitigation measures implementation strategies wasconducted. Secondly, the factors influencing the farmers’ adoptionof voluntary nutrient reduction measures and their perception ofwater quality were analyzed.