Packet-level diversity, or the ability to distribute packet transmissions over multiple, diverse channels, offers a number of benefits in improving communication performance and robustness to variations in channel quality. There have been a number of works that have analyzed and quantified those benefits, and developed transmission policies to realize them. However, translating those potential benefits into practice still faces numerous challenges. These range from uncertainty regarding the adequacy of the channel models on which the development of policies was predicated, to the many implementation constraints one faces when attempting to realize the precise transmission schedules that these policies mandate. This work is an initial step in assessing how much of the benefit that packet-level diversity promises actually remains once confronted with the many practical challenges we have just alluded to. Our investigation is carried out in the context of an 802.11 testbed, where diversity is realized through the different frequency bands available for transmissions between hosts and access points. We use our testbed to evaluate the impact of various parameters, including transmission policies, channel characteristics, channel correlation, and various end-system constraints that affect our ability to precisely control the timing of transmissions. Our investigation reveals that in spite of the many gaps that exist between theory and practice, packet-level diversity still provides a simple solution to improving transmission performance and robustness across a broad range of configurations.