"A substantial experimental literature suggests that a personal solicitation is an effective way to induce people to make charitable donations. We examine whether this result generalizes to a non-experimental setting. Specifically, we estimate the effect of a marginal personal solicitation using observational data on alumni giving at an anonymous research university, which we refer to as Anon U. At Anon U, volunteers use lists provided by the Development Office to telephone classmates and solicit them for donations. The names on these lists are always in alphabetical order. The volunteers who do the soliciting often run out of time before they reach the end of their lists. These observations suggest a simple strategy for testing whether personal solicitation matters, viz., examine whether alumni with names toward the end of the alphabet are less likely to give than alumni with names toward the beginning, ceteris paribus. If so, then a marginal personal solicitation matters. Our main finding is that location in the alphabet -- and hence, a personal solicitation -- has a strong effect on probability of making a gift. A rough estimate of the elasticity of the probability of giving with respect to the probability of receiving a personal solicitation is 0.15. However, there is no statistically discernible effect on the amount given, conditional on donating. We also find that women respond more strongly to a personal solicitation than men. This is consistent with a robust result in the psychology literature, that women find it more difficult than men to refuse requests that they perceive as being legitimate"--National Bureau of Economic Research web site.