During the COVID-19 pandemic, the idea of facilitating contact tracing using Bluetooth is becoming widespread due to the prevalence of smartphones. The automatic classification of encounters between smartphones is challenging due to variations in signal strength – resulting in high false positive and false negative rates. For example, obstructions between two smartphones can reduce the strength of received signals and thus increase the calculated distance. This can benefit contact tracing by preventing a contact from being added to the database when people are on opposite side of a wall, but can also harm tracing efforts if two people are close together but their bodies reduce the signal between devices. A Raspberry Pi Bluetooth emitter and a phone receiver were placed 1 meter apart, and various obstacles were placed between them to simulate normal obstacles. Drywall and stud walls were shown to be ineffective at reducing Bluetooth signal strength. Cinder block, and especially the human body, were found to effectively lower Bluetooth strength so that the distance estimate was higher. The results of these experiments imply Bluetooth contact tracing will involve many false positives and negatives.