In the current paper, we aim at experimentally studying whether conformity is driven by intrinsic (preference for being/choosing like others) and instrumental (due to lack of information) motivations. For this purpose, subjects are invited to taste two similar but differentiated goods (i.e., 2 different cola brands) and evaluate the satisfaction they experience from both of them. Next, they have to value monetarily both goods and choose one of them. Three different experimental treatments are implemented varying with respect to available information about others' choice. In a control treatment (Treatment 0), no information about other subjects' choice made in previous sessions is provided. In two variant treatments, subjects are given information about others' choice in the following way: “The majority of subjects in a previous session chose the cola x”. In the first variant treatment (Treatment 1), information is given to subects after they tested the goods and reported their satisfaction levels for both, but before the monetary evaluation tasks. In the second variant treatment (Treatment 2), the subjects are informed about others' choice before tasting the two goods. This experimental design allows us to test how subjects react to information about others' choice. The hypothesis of instrumental conformism implies that we should not observe any difference in monetary valuations and/or decisions between treatments. Alternatively, if subjects are (positively or negatively) sensitive to social influence, we should observe difference between treatments.

Our experimental results show that, in all treatments, the good chosen by subjects is the good reported as prefered. More interestingly, subjects' monetary valuations are sensitive to the choice made by other people. As compared to Treatment 0, we observe that the absolute gap between the two monetary valuations made by subjects tends to decrease when they report different preferences than others in Treatment 1. That is, subjects who report different preferences than others tend to exhibit a less clear decision when information about others choice is provided after testing goods. However, we do not observe a significant difference when subjects report similar preferences than others. On the contrary, in Treatment 2, the gap between monetary valuations decreases among subjects who report the same preferences for the goods that other subjects, but is not affected when subjects exhibit different preference from others. Althogether, these results suggests that individuals are differently affected by social influence depending on the previous experience of subjects with the goods and when this information is provided.