The aim of the project was to investigate the effects of two strategies of teaching new sport actions on performance of eight-year-old children: observational-imitative method (OIM) and descriptive-directive method (DDM). The OIM group was provided with a pre-practice instruction in the form of expert modeling observation by an expert athlete. The DDM group received only verbal explanations of few selected static images. Thirty-six children (18 males and 18 females, mean age = 8,8) participated in the experiment. Subjects were randomly assigned to the OIM or DDM groups. Participants were instructed to perform four sport motor sequences never performed before (shoulder stand, soccer action, vortex howler throw, step action). Actions were videotaped and 2D kinematic analysis performed. A 10-point Likert questionnaire was administered to blind sport experts to assess the correctness and accuracy of each action. Results suggest that the OIM is the most effective instruction method when participants have no experience with the sport action to be performed. On the contrary, if the athlete needs to learn specific aspects of an exercise (such as grasping a tool) the best method is the DDM. In fact, detailed information on how to grab the vortex helped children in throwing it. We also found gender differences which might reflect cultural influences in specific sports (e.g. soccer). Finally, repetition of the exercise also improved the DDM group's performance. This has potential applications in sport teaching, suggesting that in the absence of a model performing the action to be imitated, the DDM can be as effective as the OIM if the observer repeats the sport action many times.