I always strive for measuring behaviour in stead of attitude. I filtered the social desirability our during the in-depth interviews by not telling the respondents the topic beforehand.
We let them look at 3 e-commerce websites and products that had information about sustainability. When they did not notice the information about sustainability we knew they were not as interested in the topic as others, who did notice, might be. This way we could (partly) eliminate the social desirability bias.