Madam

In their first letter, Verger et al.⁽ Reference Verger, Dop and Martin-Prével ^{1 )} challenged our original result⁽ Reference Koppmair, Kassie and Qaim ^{2 )} that the effect of farm production diversity on dietary quality is small, because we had worked with the twelve food groups that are often used for the household dietary diversity score (HDDS) to characterize individual-level dietary quality. In fact, we had compared results with household- and individual-level data. Verger et al.⁽ Reference Verger, Dop and Martin-Prével ^{1 )} suggested that other food group classifications are better suited to characterize individual-level dietary quality for women and children. In our response⁽ Reference Koppmair and Qaim ^{3 )} to that first letter, we provided further reasons for our approach but also carried out additional analyses with the alternative indicators suggested. These additional analyses confirmed the original results. We also showed that the different indicators are closely correlated, which means that the twelve-food-group classification is a valid proxy of dietary quality in this particular case. In their second letter, Verger et al.⁽ Reference Verger, Jones and Dop ^{4 )} emphasize that this correlation does not mean that the twelve-food-group indicator is a good proxy of individual dietary quality in general. We agree with this caveat. However, we want to stress that the results of our original study⁽ Reference Koppmair, Kassie and Qaim ^{2 )}, namely that the effect of farm production diversity is small and that market access is more important for household and individual dietary diversity and dietary quality, are robust to the different indicators used.