Is it easier to publish in journals that have low impact factors?

We use publication data from all journals in economics, mathematics, physics and chemistry. We compute the ratio of publications that come from the top 25 US departments for each journal and use it as a measure of difficulty to publish in a journal. We find that there is a rather weak but positive correlation (0.20) between the impact factor and the difficulty measure.