A review of the current research on the interference phenomenon between concurrent aerobic and strength training indicates modest support for the model proposed in this article. However, it is clear that without a systematic approach to the investigation of the phenomenon there is lack of control and manipulation of the independent variables, which makes it difficult to test the validity of the model. To enhance the understanding of the interference phenomenon, it is important that researchers are precise and deliberate in their choice of training protocols. Clear definition of the specific training objectives for strength (muscle hypertrophy or neural adaptation) and aerobic power (maximal aerobic power or anaerobic threshold) are required. In addition, researchers should equate training volumes as much as possible for all groups. Care needs to be exercised to avoid overtraining individuals. There should be adequate recovery and regeneration between the concurrent training sessions as well as during the training cycle. The model should be initially tested by maintaining the same protocols throughout the duration of the study. However, it is becoming common practice to use a periodised approach in a training mesocycle in which there is a shift from high volume and moderate intensity training to tower volume and higher intensity. The model should be evaluated in the context of a periodised mesocycle provided the investigators are sensitive to the potential impact of the loading parameters on the interference phenomenon. It may be that the periodised approach is one way of maintaining the training stimulus and minimising the amount of interference. The effects of gender, training status, duration and frequency of training, and the mode of training need to be regarded as potential factors effecting the training response when investigating the interference phenomenon. Other experimental design factors such as unilateral limb training or training the upper body for one attribute and the lower body for another attribute, may help establish the validity of the model.