Going to failure, or not, has probably been one of the most debated issues during the history of strength training. However, few studies have directly compared the physiological effect of failure vs. nonfailure strength training. The purpose of this study was to evaluate muscle activation strategies with electromyography (EMG) during heavy repetitions vs. repetitions to failure with lighter resistance. Fifteen healthy untrained women performed a set with heavy loading (3 repetition maximum [RM]) and a set of repetitions to failure with lower resistance (∼15 RM) during lateral raise with elastic tubing. Electromyographic amplitude and median power frequency of specific shoulder and neck muscles were analyzed, and the BORG CR10 scale was used to rate perceived loading immediately after each set of exercise. During the failure set, normalized EMG was significantly lower during the first repetition and significantly higher during the latter repetitions compared with the heavy 3-RM set (p < 0.05). Normalized EMG for the examined muscles increased throughout the set to failure in a curvilinear fashion--e.g., for the trapezius from 86 to 124% maximal voluntary contraction (p < 0.001)--and reached a plateau during the final 3-5 repetitions before failure. Median power frequency for all examined muscles decreased throughout the set to failure in a linear fashion, indicating progressively increasing fatigue. In conclusion, going to complete failure during lateral raise is not necessary to recruit the entire motor unit pool in untrained women--i.e., muscle activity reached a plateau 3-5 repetitions from failure with an elastic resistance of approximately 15 RM. Furthermore, strengthening exercises performed with elastic tubing seem to be an efficient resistance exercise and a feasible and practical alternative to traditional resistance equipment.