There is currently no consensus regarding the effect that barefoot (BFT) running has on running economy (RE). Stride length and shoe mass are confounding variables, with a BFT stride length being shorter than a shod (SH) stride length. Comparison of SH, minimalist shod (MS) and BFT allows controlled variation of cushioning and somatosensory feedback to determine the effect that either and/or both have on RE and running mechanics. Methods: Fifteen female habitually shod, recreational runners visited the laboratory twice. Familiarisation with BFT and SH treadmill running occurred during visit one, in addition to determining SH stride length and BFT stride length. During visit two participants ran BFT, SH and MS with BFT stride length and MS with SH stride length at 10 kmÁh À1 for six minutes with 10-minute rest periods between each condition. Lower limb kinematics, electromyography, impact acceleration and _ V O 2 were recorded during the final two minutes of each run. Results: BFT RE was significantly better than SH and MS with BFT stride length. SH RE was significantly worse than MS with SH stride length, but similar to MS with a BFT stride length. Low vertical oscillation, peak eversion and peak dorsiflexion, less plantarflexion at toe-off, in addition to an earlier occurrence of heel off, higher impact accelerations and greater tibialis anterior activity were observed during the most economical condition. Conclusions: Heightened somatosensory feedback and lack of cushioning (BFT) offered an advantage to economy over less somatosensory feedback (MS) and cushioning (SH). Whilst the low vertical oscillation and low plantarflexion at toe-off appear to contribute to the improved RE, other changes to running mechanics whilst BFT could potentially influence injury risk.