Skip to main content
Log in

Propose with a rose? Signaling in internet dating markets

  • Original Paper
  • Published:
Experimental Economics Aims and scope Submit manuscript


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.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others


  1. There are numerous examples of informal preference signaling. In the economics job market, for example, it is through advisors and their connections that graduate students on the market can convey their interest. In the law clerk market, law appellate court judges seem to be able to receive binding commitments from law students (see Avery et al. 2001). In U.S. college admissions, students are advised to show interest by visiting the college. For example, college data, at (, claims that “demonstrated interest” is a factor of considerable importance to colleges: “Going on a college visit, talking with admission officers, or doing an enthusiastic interview can call attention to how much you really want to attend. Applying for an early decision may also make a good impression.”

  2. Examples of recent studies are Avery and Levin (2010) and Coles et al. (2013).

  3. A much-studied version of signaling is costly signaling (see Spence 1973), where agents undertake various actions, in general visible to all participants, whose costs depend on the underlying trait to be signaled. Such costly signaling has, for example, been used as a partial explanation for education (for an early overview, see Weiss 1995), conspicuous consumption (Veblen 1899; Charles et al. 2009) and even in biology for the famous extravagance of the peacock’s tail (Zahavi 1975). It is, however, difficult to show that costly signals sway the decisions of other agents (see the debates regarding the signaling value of education: Tyler et al. 2000; Jepsen et al. 2010; Martorell and Clark 2014). Given the difficulty of proving the effect of costly signaling, it is not surprising that the empirical literature on preference signaling that does not even yield direct costs has faced similar difficulties.

  4. For example, the datasets used in existing studies on preference signaling provide only a partial list of the colleges/firms a person applied to when seeking admission/a job.

  5. Such an intervention may be ethically more problematic in the labor or education markets. The main difference between the dating and employment environments is that the dating market is more continuous. As such, any dating website is portioning off a fraction of the “natural” dating market and manipulating it. It is much more problematic to influence a national or even international market such as the economics junior market that operates once a year and whose initial outcome may have a large impact on careers (Oyer 2006).

  6. Marriage has received some attention following the seminal work by Becker (1973). Examples of empirical studies on marriage include Abramitzky et al. (2011), Choo and Siow (2006), Fisman et al. (2006, 2008), Hitsch et al. (2010), and Lee (2009).

  7. Kim (2010) focuses on early admission as a screening device for students who do or do not require financial aid.

  8. The AEA offers advice includes: “The two signals should not be thought of as indicating your top two choices. Instead, you should think about which two departments that you are interested in would be likely to interview you if they receive your signal, but not otherwise (see advice to departments, above). You might therefore want to send a signal to a department that you like but that might otherwise doubt whether they are likely to be able to hire you.” (see

  9. A person’s desirability index is calculated based on earnings, assets, job security (full time job or not), height, weight, a company-generated score based on the profile picture, a score based on the college attended and the chosen major, both of which are highly correlated with the score on the national college entry exam, birth order, and family characteristics (parents’ wealth and marital status, and siblings’ educational attainment).

  10. To receive 100 percent verification, a participant needs to submit a copy of the national household registration form (for age, birth order, marital history and parents’ marital status), diploma (for education) and proof of employment (for type of employment and industry).

  11. There may be several core outcomes of who is married to whom, in which case men’s preferred outcome is different from the women’s preferred outcome (see Roth and Sotomayor 1990, for an overview). A dating market in which men make offers may be closer to achieving the male optimal stable matching, the most preferred outcome by men. Lee (2009) provides evidence that matches would be quite different if women were to make offers.

  12. The banner read, in translation: “Will you wait until Prince Charming asks you out? Or will you take the lead to meet him? Dear client, did you find someone you want to date? Please do not let this opportunity pass you by. Contact him first and give him the opportunity to meet you.”

  13. The banner read, in translation, “Congratulations! You have received a dating request. Please give an opportunity to the one who has fallen in love with your charms!”.

  14. While the average number of accepted offers is similar between women and men, women are significantly less likely to accept an offer than men (16 versus 29 percent, p < 0.01). This is because women on average receive 5.9 offers, while men receive only 3.9.

  15. This means we treat “no response” as an explicit rejection. In Sect. 7 of the supplementary document, we present evidence that our approach is justified.

  16. See Sect. 6 of the supplementary document for details, including identification assumptions and a formal description of the IV model.

  17. In addition to the approaches explained in this paper, we perform the following two exercises (for details, see Sect. 2 of the supplementary document). First, instead of our baseline cutoffs (30th percentile and 70th percentile), we use the 20th and 80th percentile to classify participants into three desirability groups. We re-estimate Model A and find that a rose increases the chance of a proposal being accepted by 3.2 percentage points, almost identical to the baseline result. Second, we use the number of proposals a participant received as a proxy for the participant’s desirability. We re-estimate Model A but include dummy variables of the number of proposals a sender received instead of the desirability index group dummies. We find that a recipient accepts a proposal by 3.4 percentage points more if the proposal is accompanied by a rose, an effect virtually identical to the baseline result.

  18. We have 56 individuals who participated in both sessions, and 39 of them received at least one proposal in the second session. We examine whether recipients respond to a rose differently in their second session. To do so, we re-estimate Model A but include the interaction between a rose and a dummy variable that indicates the second session and two-time participants. Note that 215 out of 1,921 proposals are sent to two-time participants. We find that there is no statistical difference in terms of recipients’ response to a rose in their second participation.

  19. We also run a regression where, in addition to fixed effects for recipients, we use fixed effects for senders instead of their desirability group. The estimated coefficient of a rose is 0.031, qualitatively the same as in the baseline regression (column 1), though just barely not significant (the s.e. is 0.019, p = 0.104).

  20. We also formally test whether the effect of roses on the acceptance rate depends on a recipient’s treatment group. We re-estimate Model A while including an interaction term between receiving a rose and whether a recipient had eight or two roses. The coefficient on the interaction term is not significant, indicating that the difference in the acceptance rate due to a rose is similar between recipients who themselves had two or eight roses (see Sect. 7 of the supplementary document).

  21. Participants who have two roses may view an offer with a rose as “special,” while offers without a rose show perhaps “normal” interest. On the other hand, participants who have eight roses may not feel equally flattered when receiving a rose. However, for them, not receiving a rose may be a sign of not really being special, since, in their view, only two out of ten offers are precluded from having a rose attached. Note that these two cases are in a way symmetric: either two out of ten offers are more special compared to other offers—for recipients endowed with two roses—or two out of ten offers are less special compared to other offers—for recipients endowed with eight. Due to this symmetry, it may not be surprising that the change in the acceptance rate in reaction to a rose may be similar for recipients endowed with two or eight roses.

  22. Note that in the market for medical residents, the welfare implications of a centralized clearinghouse are also hard to assess. It seems, however, that centralized clearinghouses affect who is matched with whom (Niederle and Roth 2003) and the timing of the match (Roth 1984; Niederle et al. 2006).


  • Abramitzky, R., Delavande, A., & Vasconcelos, L. (2011). Marrying up: The role of sex ratio in assortative matching. American Economic Journal, Applied Economics, 3(3), 124–157.

    Article  Google Scholar 

  • Avery, C., Fairbanks, A., & Zeckhauser, R. (2003). The early admissions game: Joining the elite. Cambridge: Harvard University Press.

    Google Scholar 

  • Avery, C., Jolls, C., Posner, R. A., & Roth, A. E. (2001). The market for federal judicial law clerks. University of Chicago Law Review, 68(3), 793–902.

    Article  Google Scholar 

  • Avery, C., & Levin, J. (2010). Early admissions at selective colleges. American Economic Review, 100(5), 2125–2156.

    Article  Google Scholar 

  • Becker, G. S. (1973). A theory of marriage: Part I. Journal of Political Economy, 81, 813–846.

    Article  Google Scholar 

  • Charles, K., Hurst, E., & Roussanov, N. (2009). Conspicuous consumption and race. Quarterly Journal of Economics, 124(2), 425–467.

    Article  Google Scholar 

  • Choo, E., & Siow, A. (2006). Who marries whom and why. Journal of Political Economy, 114(1), 175–201.

    Article  Google Scholar 

  • Coles, P., Cawley, J., Levine, P. B., Niederle, M., Roth, A. E., & Siegfried, John J. (2010). The job market for new economists: A market design perspective. Journal of Economic Perspectives, 24(4), 187–206.

    Article  Google Scholar 

  • Coles, P., Kushnir, A., & Niederle, M. (2013). Signaling in matching markets. American Economic Journal: Microeconomics, 5(2), 99–134.

    Google Scholar 

  • Fisman, R., Iyengar, S., Kamenica, E., & Simonson, I. (2006). Gender differences in mate selection: Evidence from a speed dating experiment. Quarterly Journal of Economics, 121(2), 673–697.

    Article  Google Scholar 

  • Fisman, R., Iyengar, S., Kamenica, E., & Simonson, I. (2008). Racial preferences in dating. Review of Economic Studies, 75(1), 117–132.

    Article  Google Scholar 

  • Hitsch, G., Hortacsu, A., & Ariely, D. (2010). Matching and sorting in online dating. American Economic Review, 100(1), 130–163.

    Article  Google Scholar 

  • Jepsen, C., Mueser, P.R., & Troske. K. (2010). “Labor market returns to the GED using regression analysis,” University of Kentucky, Working Paper.

  • Kim, M. (2010). Early decision and financial aid competition among need-blind colleges and universities. Journal of Public Economics, 94, 410–420.

    Article  Google Scholar 

  • Kurzban, R., & Weeden, J. (2005). HurryDate: Mate preferences in action. Evolution and Human Behavior, 26, 227–244.

    Article  Google Scholar 

  • Lee, S. (2009). “Marriage and online mate-search services: Evidence from South Korea.” University of Maryland, Working Paper.

  • Martorell, P., & Clark, D. (2014). The signaling value of a high school diploma. Journal of Political Economy, 122(2), 282–318.

    Article  Google Scholar 

  • Niederle, M., Proctor, D. D., & Roth, Alvin E. (2006). What will be needed for the new GI fellowship match to succeed? Gastroenterology, 130, 218–224.

    Article  PubMed  Google Scholar 

  • Niederle, M., & Roth, A. E. (2003). Unraveling reduces mobility in a labor market: Gastroenterology with and without a centralized match. Journal of Political Economy, 111(6), 1342–1352.

    Article  Google Scholar 

  • Oyer, P. (2006). Initial labor market conditions and long-term outcomes for economists. Journal of Economic Perspectives, 20(3), 143–160.

    Article  Google Scholar 

  • Oyer, P. (2014). Everything I ever needed to know about economics I learned from online dating. Boston MA: Harvard Business Review Press.

    Google Scholar 

  • Roth, A. E. (1984). The evolution of the labor market for medical interns and residents: A case study in game theory. Journal of Political Economy, 92, 991–1016.

    Article  Google Scholar 

  • Roth, A. E. (2008). What have we learned from market design? Hahn Lecture, Economic Journal, 118, 285–310.

    Google Scholar 

  • Roth, A. E., & Peranson, E. (1999). The redesign of the matching market for American physicians: Some engineering aspects of economic design. American Economic Review, 89(4), 748–780.

    Article  Google Scholar 

  • Roth, A. E., & Sotomayor, M. (1990). Two-sided matching: A study in game-theoretic modeling and analysis. Cambridge: Cambridge University Press.

    Book  MATH  Google Scholar 

  • Roth, A. E., & Xing, X. (1997). Turnaround time and bottlenecks in market clearing: Decentralized matching in the market for clinical psychologists. Journal of Political Economy, 105(2), 284–329.

    Article  Google Scholar 

  • Spence, M. (1973). Job market signaling. Quarterly Journal of Economics, 87(3), 355–374.

    Article  Google Scholar 

  • Tyler, J. H., Murnane, R. J., & Willett, J. B. (2000). Estimating the labor market signaling value of the GED. Quarterly Journal of Economics, 115(2), 431–468.

    Article  MATH  Google Scholar 

  • Veblen, T. (1899). The theory of the leisure class: an economic study of institutions, Kila, MT: Kessinger. (Reprint).

  • Weiss, Andrew. (1995). Human capital vs. signaling explanations of wages. Journal of Economic Perspectives, 9(4), 133–154.

    Article  Google Scholar 

  • Wooldridge, Jeffrey. (2010). Econometric analysis of cross section and panel data (2nd ed.). Cambridge: The MIT Press.

    MATH  Google Scholar 

  • Zahavi, A. (1975). Mate selection–a selection for a handicap. Journal of Theoretical Biology, 53, 205–214.

    Article  CAS  PubMed  Google Scholar 

Download references


We thank Woongjin Lee, Hye-Rim Kim and Woo-Keum Kim from the Korea Marriage Culture Institute for help running the experiment and Benjamin Malin and Alvin Roth for detailed comments. This manuscript has been greatly improved thanks to the editor and two anonymous referees. We have benefitted from discussions with Ran Abramitzky, Peter Coles, Ray Fisman, Sriniketh Nagavarapu, Ragan Petrie, Aloysious Siow, Lise Vesterlund, Leeat Yariv, and seminar participants at various institutions. Muriel Niederle gratefully acknowledges support from the NSF.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Soohyung Lee.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (PDF 861 kb)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lee, S., Niederle, M. Propose with a rose? Signaling in internet dating markets. Exp Econ 18, 731–755 (2015).

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


JEL Classification