To determine whether an association exists between different clusters of fruit- and vegetable-specific family-environmental factors and children's daily fruit and vegetable intake, and whether these associations differ between countries with different school lunch policies.

Cross-sectional data from four European countries participating in the Pro Greens project in 2009. These countries have different school food policies: two serve free school lunches and two do not. Self-administered data were used. Food frequency questions served to assess fruit and vegetable intakes. The study assessed sixteen children-perceived family-environmental factors, which were clustered based on principal component analysis into five sum variables: fruit and vegetable encouragement; vegetable modelling, family routine and demand; fruit modelling; fruit and vegetable snacking practices; and fruit and vegetable allowing.

Schools in Finland, Germany, Sweden and the Netherlands.

Schoolchildren aged 11 years (n 3317).

Multilevel logistic regression analyses revealed positive associations between nearly all clustered family-environmental factors and daily fruit and vegetable intake. The study tested a moderation effect between family-environmental factors and school lunch policy. In five out of twenty models significant interactions occurred. In the stratified analyses, most of the associations between family-environmental factors and raw and cooked vegetable intake were stronger in Germany and the Netherlands, neither of which provided free school lunches.

Children reporting more fruit- and vegetable-promoting family-environmental factors had a more frequent intake of fruits and vegetables; the associations were stronger for vegetable intakes in countries providing no free school lunches, suggesting that parental involvement is crucial when schools offer no vegetables.