Trust is an important variable in economics, as several transactions are based on it; unfortunately it is difficult to measure. The recent economic literature on social capital shows a positive association between this concept and trust. As social capital is easier to measure than trust is, this paper analyzes the possibility of assessing trust measuring social capital using experimental economics. A basic trust game is played in three Western European countries with undergraduate students; a questionnaire measures their level of social capital, as time spent within social networks. This measure is stronger and more precise than the ones generally used. In particular this paper firstly measures social capital as the intensity of a membership to a voluntary organization, while the extant literature generally considers only the membership per se. Secondly the use of an experiment instead of a questionnaire allows for constructiong a measure of trust which is in principle continuous. Thirdly to play an experiment allows for observing the behaviour of the participants better than by the means of a survey. The results are supportive of the fact that trust can be assessed through social capital, although the presence of a strong geographical effect has to be accounted for.