In a paper published in September 2016, Giessing et al. revisit this question. They point out that to date experiments, meta-analyses, and literature reviews include studies that are heterogeneous in respect of controlling a very important variable, namely training effort.
Advocates of single-set per exercise high intensity training (HIT) generally recommend continuing each exercise until the point of momentary muscular failure (MMF) or local exhaustion, i.e. the point at which it is temporarily impossible to perform another repetition in proper form. Practitioners of high-intensity training often include so-called drop sets, wherein after reaching MMF in the main set of an exercise, the resistance is immediately reduced 10-15% to allow 2-3 additional repetitions, after which the resistance is again reduced 10-15% to allow another 2-3 repetition. In contrast, advocates of multiple set high volume (HV) routines generally recommend fixed repetition targets and often recommend avoiding training to momentary muscular failure.
Giessing et al. set up their experiment to directly compare these two methods of training: single-set to MMF, or multi-set not-to-failure (NTF). Thirty subjects were randomly assigned to perform either a single-set program including 2 drop sets, or a multiple-set program (3 sets per exercise) NTF.
Both groups performed the following exercises in the listed order: chest press, heel raise, rear deltoid, elbow flexion, seated row, knee extension, knee flexion, abdominal (trunk) flexion, push-ups. These were also the test exercises at the end of the training experiment.
Regarding this menu of exercises, chest press and push ups both train the pectorals, anterior and medial deltoids, and triceps; seated row trains upper back, rear deltoid, and elbow flexors, while elbow flexion also trains elbow flexors and rear deltoid trains rear deltoid. The push up also trains the trunk flexors and knee extensors to some degree (although perhaps not to MMF). The remaining movements involve only one muscle group. Thus, the "single-set" group actually did two sets affecting each of the following muscle groups: pectorals, anterior and medial deltoids, triceps, rear deltoid, knee extensors, trunk flexors, and elbow flexors. The program design resulted in only one training set for only the upper back, gastrocnemius (heel raise), and knee flexors in the single-set group.
The HIT group did one set of each exercise to MMF, followed by 2 drop sets. The HV performed each exercise to a self-determined maximum number of repetitions, i.e. stopped each set when he or she felt like stopping. Subjects rested long enough between sets to normalize breathing. The HV group rested 2-3 minutes at the end of each circuit of 9 exercises before doing the second and third rounds.
The groups differed also in repetition performance: "The HIT group used a repetition duration of 2 seconds concentric, 1 second isometric contraction at the top of the range of motion, and 4 seconds eccentric (2-1-4 seconds). The 3ST group trained using a repetition duration of 2 seconds concentric and 2 seconds eccentric (2-2 seconds)." Each exercise was performed for 10 repetitions excluding drop sets.
The study duration was ten weeks. Each group trained all exercises twice weekly. The HIT group spent about 11 minutes on each training session; the HV group spent about 25 minutes on each training session.
Both the HIT and HV programs produced significant improvements in muscular performance. However, the HIT program produced significantly greater gains than the HV program in 3 of the tested exercises, and had larger changes in 8 of the 9 exercises. Neither group showed statistically significant changes in any body composition measures. The near null results for body composition changes were probably due to the inclusion of both male and female trainees of roughly equal numbers; since female trainees do not generally show large muscle mass gains in response to resistance training, their results washed out any evidence of gains in the male trainees in the average results. However, there was a small effect size trend for greater body composition changes in the HIT group.
What accounts for these differences? The programs differed in volume (number of repetitions and time under load), intensity of effort, use of drop sets, and duration of repetitions.
The HV group performed a greater volume of exercise (both number of repetitions and time under load), but produced less performance gains. Therefore, the greater performance improvement in the HIT group was not due to performance of a greater exercise volume.
The HIT group trained to MMF but the HV group did not. The authors note that research has suggested that training to MMF produces greater adaptations than not training to MMF, provided recovery is adequate. Training to MMF produces a greater degree of momentary fatigue in the trained muscle groups compared to training to a self-determined number of repetitions. The fatigue level was further increased in this study by adding drop-sets. Fatigue stress stimulates the body to make adaptations. It makes sense that creating greater fatigue will elicit greater results, provided recovery is adequate. The reduced volume of the HIT program may allow greater recovery due to the fact that it consumes fewer metabolic resources than the HV program.
My Take on This
In practice, if you are going to perform multiple sets of any exercise for the same number of repetitions and with the same resistance used for each set, it is necessary to avoid training to MMF in all sets leading up to the final set. In my experience, if one trains to MMF on the first set, it will be impossible to duplicate the performance on the second set with the same load within 3 to 5 minutes, so I would have to either reduce the repetitions, or reduce the load by 10-20% on the second set to get the same number of repetitions. It would look something like this:
BB squat: 200 lbs. x 20 reps (near MMF)
rest 3-5 minutes
BB squat: 200 lbs. x 10-15 reps; or 180 lbs. x 15-20 reps (both near MMF)
Hence, when assigned to perform multiple sets of X number of repetitions of any exercise, trainees will generally hold back effort on early sets to reserve strength for subsequent sets.
One has to question this approach on the basis of the progressive overload principle. The second set is performed for fewer repetitions with the same resistance, or with a lesser resistance and possibly fewer repetitions than the first. Hence the degree of stimulation does not reach the higher level of the first set. Since either or both the resistance and the repetitions are less on the second set, the second set does not qualify as a progression of overload.
Over the past 40 years, the vast majority of my resistance training routines have been single-set, low-volume, high intensity training to MMF. During some periods I have used simple routines doing 3 sets of 5 repetitions using only basic exercises (squats, deadlifts, pull-ups, overhead presses, bench presses). I have been able to increase strength using either type of program.
However, I tend to prefer mostly single-set training to MMF because it is so time-efficient. One set of chin ups to MMF will take about a minute to perform. Three sets of 5 repetitions of chin ups with 2 minutes rest between sets will (assuming 40 s per set) take 6 minutes to perform. This may seem minor but when performing 10-12 exercises per full body training session, a single-set routine will involve ~15 minutes of actual exercise performance and ~20 minutes of rest periods and set up times, or a total of 35 to 45 minutes, whereas a multi-set routine using the same number of exercises would take 90 minutes or more to complete.
One general comment I have about the research on this topic is that all studies I have seen that look at single-set vs. multiple-set training routines define single-set routines as those in which subjects perform only one set of each exercise, and multi-set routines as those in which subjects perform multiple sets of each exercise. Yet in practice many routines provide multiple sets for individual muscle groups even if the individual performs only one set per exercise.
For example, consider this routine:
- Barbell squat
- Chin up
- Parallel bar dips
- Cossack squat
- Rowing on rings
- Push-ups on rings
- Thighs & hips: barbell squat and Cossack squat
- Upper back, rear deltoid, arm flexors and forearms: chin up and rowing
- Pectorals, anterior and medial deltoid, arm extensors: parallel bar dips and push ups on rings
In summary, Giessing et al. appear to have shown that a HIT program involving only one set per exercise and drop sets will produce similar or greater results than a HV program consisting of 3 sets per exercise without training to MMF. However, it is possible that other types of HV training may produce equal or superior results, e.g. 2 or more sets performed to MMF rather than to self-determined 'maximum' repetitions.