We report on an experiment using video technology to manipulate pre-play communication protocols in the lab and to study purely social effects of communication on donations and discrimination between potential receivers. The experimental design eliminates strategic factors by allowing two receivers to unilaterally communicate with an anonymous dictator before the latter decides on her gifts. Through the use of three communication setups (none, audio, and audio-visual) we show and analyze the existence of purely social effects of communication. We find that a silent channel leads to discrimination between potential receivers based on impression formation, but does not affect average levels of donations. When the auditory channel is added, average donations increase. The social processes invoked are heterogeneous and communicator-specific but not unsystematic.