Prospective candidates with connections in committees may have access to more accurate information about evaluation standards. When applications are costly, this informational advantage may reduce the application rate of connected individuals, leading to a positive selection among applicants. We document the relevance of this phenomenon using data from national evaluations in Italian academia. Researchers are significantly less likely to apply when the committee includes, through the luck of the draw, a colleague or a coauthor. At the same time, they tend to receive more favorable evaluations from their connections. Our analysis indicates that self-selection may bias in a non-trivial way estimates of evaluation biases that rely on observational data.