Attribute non-attendance in environmental discrete choice experiments: The impact of including an employment attribute

Abstract

This paper utilizes data from a split-sample discrete choice experiment to investigate the impact of including an employment attribute on stated preferences for protecting the coastal zone of Arctic Norway. The econometric analysis investigates how its inclusion affects attention to other choice experiment dimensions, and how welfare measures vary between the two subsamples and across models that control for attribute non-attendance versus models that do not. We find that the employment attribute has a relatively high attendance rate and that its inclusion does not appear to decrease attention to other attributes of interest. The impact of the added attribute on the part-worth estimates for environmental attributes is mixed. However, similar to prior research, we find that controlling for attribute non-attendance tends to yield lower welfare estimates. Lastly, our analysis indicates somewhat higher attention to the cost attribute than many previous studies.