Propose with a rose? Signaling in internet dating markets

A growing number of papers theoretically study the effects of introducing a preference signaling mechanism. However, the empirical literature has had difficulty proving a basic tenet, namely that an agent has more success when the agent uses a signal. This paper provides evidence based on a field experiment in an online dating market. Participants are randomly endowed with two or eight "virtual roses" that a participant can use for free to signal special interest when asking for a date. Our results show that, by sending a rose, a person can substantially increase the chance of the offer being accepted, and this positive effect is neither because the rose attracts attention from recipients nor because the rose is associated with unobserved quality. Furthermore, we find evidence that roses increase the total number of dates, instead of crowding out offers without roses attached. Despite the positive effect of sending roses, a substantial fraction of participants do not fully utilize their endowment of roses and even those who exhaust their endowment on average do not properly use their roses to maximize their dating success.