Human rights issues are becoming increasingly important in supply chains, particularly with the advent of regulatory requirements such as the EU’s CSDDD. However, getting reliable data is difficult. Companies lack the tools to monitor risks of human rights violations in their complex global supply chains – especially those consisting of thousands of smallholder farmers. Fearing breach of confidentiality, workers are unlikely to report human rights violations through direct questioning.
Sucden, a leading cocoa exporter that sources from and works with more than 100,000 farmers across West Africa, is committed to robust monitoring of its supply chain. 60 Decibels worked with Sucden to test the success of innovative methods (the List Experiment) in collecting data on sensitive topics like child labor and adult labor practices. The study also compared the reliability of phone and in-person survey methodologies. In total, the study surveyed 418 farmers over phone and 405 farmers in person in Côte d’Ivoire.
The objective of the study was to hear directly from farmers about living conditions and labor practices in their supply chain, and test new methods of indirect questioning for collecting due diligence information on sensitive topics.
Indirect questioning was more effective in eliminating respondent bias for sensitive topics. Using an indirect List Experiment method, 17% of farmers reported child labor in harvesting/pod breaking, higher than when directly asked (5%). Additionally, respondents were more likely to report sensitive labor practices over the phone.
The List Experiment is an indirect questioning method that limits untruthful answers caused by desirability bias, shame, or fear. It works by allocating respondents randomly to two different groups. Individuals allocated to the first group are presented with several non-sensitive statements (typically 3). They are not asked to say whether they agree with each of the statements individually, but only how many (in total) of the statements they agree with.
The same statements are then presented to the second group, with the difference that a sensitive statement (in this case, regarding child labor) is added to the series of non-sensitive statements.
Assuming that the two groups adopt similar practices with regards to the non-sensitive statements (which is ensured by the random allocation to the groups), one can deduce the share of individuals in the treatment group (and thus the whole population) who agreed with the sensitive item by comparing the average number of statements with which the respondents in each group agreed.
Farmers were more likely to report sensitive information like child labor when surveyed over the phone. Farmers were also more willing to disclose hiring workers from different provinces or countries, language barriers with workers, and irregular payment schedules.
Conversely, respondents exhibited varying levels of transparency depending on the survey method. In-person interviews seemed to exert a level of pressure, with farmers being more likely to report a safe home environment and recent visits from cooperatives or buyers. However, there were no significant differences in responses between phone and in-person surveys for most topics such as cocoa prices and credit usage.
Sucden is exploring how to incorporate telephone interviews into its ongoing data collection processes from farmers. For data relating to farm activity Sucden believes this can bring more efficiency and better connectivity to the process.
Sucden is also working to determine how to use the List Experiment approach within its due diligence processes – especially in monitoring for child labor and for assessing adult labor practices.
