Association for Public Policy Analysis & Management

We present results from a series of large-scale field experiments run with the College Board to measure the causal impact of Facebook ads on student decisions to take the SAT. We were able to reach roughly 100,000 low income students by matching their email addresses to Facebook’s records. However, relative to other types of interventions (such as text messages), we found that messaging through Facebook was relatively ineffective overall. However, the messages were also very inexpensive and the data suggests may have been cost effective for some subsamples of the population.