Single vs multiple sets: Finally, the answer! (We hope.)

August 11th, 2009  |  Published in Stumpblog  |  18 Comments

Oh, how the debate has raged over the years. Single sets to failure? Multiple sets? How many?

Leaving aside the fact that the body is not great at counting, which often invalidates many of the Baroque mathematical elaborations of bodybuilders, and forgetting that we should just lift the damn heavy thing until it isn’t as heavy and then go find a heavier thing, the question of whether single sets of exercises is better than multiple sets is actually kind of a good question for folks seeking optimal efficiency and results from their workouts.

Well, here’s the answer. We think. A meta-analysis concluded that “2 to 3 sets per exercise are associated with 46% greater strength gains than 1 set, in both trained and untrained subjects.”

There you have it. 46% more — that’s nearly half — and that’s good enough for me.

Krieger, James. “Single Versus Multiple Sets of Resistance Exercise: A Meta-Regression.” The Journal of Strength & Conditioning Research.

Responses

  1. Richard Campbell says:

    August 11th, 2009at 11:56 am(#)

    In a dose-response model, 2 to 3 sets per exercise were associated with a significantly greater ES than 1 set (difference = 0.25 +/- 0.06; CI: 0.14, 0.37; p = 0.0001). There was no significant difference between 1 set per exercise and 4 to 6 sets per exercise (difference = 0.35 +/- 0.25; CI: -0.05, 0.74; p = 0.17) or between 2 to 3 sets per exercise and 4 to 6 sets per exercise (difference = 0.09 +/- 0.20; CI: -0.31, 0.50; p = 0.64).

    How can there be no significant difference between 1 set and 4-6 sets, and no significant difference between 4-6 and 2-3 sets, and yet be a significant difference between 1 set and 2-3 sets? That seems more than a little screwy…

  2. Mistress Krista says:

    August 11th, 2009at 12:50 pm(#)

    This is actually rather unclear isn’t it? Do they mean that there was no difference in terms of overall training effect — e.g. a few is better than one but past that it doesn’t matter so much?

  3. felipe says:

    August 11th, 2009at 12:56 pm(#)

    4-6 could be intermediate in effect size between 1 and 2-3. There could also be just a lot of variation in the 4-6 range, just swamping any differences in averages. (Pretty complicated model for this; I have to dig around …)

    That begs the next question: how many reps? No, no, no. I agree with you, Krista, to pick up heavy things until they aren’t heavy, then find something heavier. But are we talking about super-sets? Heavy singles?? What? Maybe I should buy the article (not …)

  4. Elizabeth Fry says:

    August 11th, 2009at 6:56 pm(#)

    I think it is because the study was a meta-analysis. They are comparing *effect size* — in the studies that compared the training effect of 1 set versus none (?) the effect was y. In studies that compared 4-6 sets versus none (?) the effect size was also y (or at least y + a non-significant variance). They were not comparing training effect of 2-3 sets versus 4-6 sets. It wasn’t a single study, it was a meta-analysis comparing several.

    Beth

  5. James Krieger says:

    August 11th, 2009at 9:27 pm(#)

    *************
    How can there be no significant difference between 1 set and 4-6 sets, and no significant difference between 4-6 and 2-3 sets, and yet be a significant difference between 1 set and 2-3 sets?
    **************

    I am the author of this paper. The lack of significant difference between 4-6 sets and 1 set is an issue of statistical power. There were only 2 studies included in the meta-analysis that incorporated 4-6 sets per exercise. Thus, there was not adequate statistical power to detect a difference between 1 set and 4-6 sets. The magnitude of the difference in effect sizes indicates a clear trend, though.

    At the end of the paper, I state that more research on dose-response relationships is needed (for example, studies comparing 1 set to 3 sets to 6 sets).

  6. Mistress Krista says:

    August 12th, 2009at 5:09 am(#)

    Awesome! Thanks for posting, James, and for clearing that up!

  7. felipe says:

    August 13th, 2009at 4:39 am(#)

    Terrific, James, for posting.

    Um, so, how many reps? (And I think several links on your site are broken; I keep getting a “/” error message when I click them.)

  8. James Krieger says:

    August 13th, 2009at 9:40 am(#)

    The study did not address # of repetitions. In fact, repetition number was not even included in the statistical model because all the studies used the same range of repetitions (7-10 repetitions).

    The links on the JOPP website appear to be working for me.

  9. Fred Hahn says:

    September 23rd, 2009at 12:17 pm(#)

    It is screwy and if you look at the studies included in the meta you will see glaring errors in the researchers conclusons.

    A meta analysis is only as good as the studies included. Many of the studies included in this meta are poorly conducted.

    I will be doing a comprehensive analysis of the 14 studies included in this meta to see if in fact they hold up to scrutiny.

    Several of the studies showed a greater strength increase for the multiple set group but no significant difference in lean body mass increases. This tells you that more sets = more skill acquisition but more sets are not more stimulatory to gaining lean tissue which is the real goal of weight trianing (not Olympic lifting or power lifting).

    One study (Borst et. al.)showed the same increases in circulating IGF-I for both 1 and 3 sets for the first 13 weeks and no further increases for either 1 or 3 sets for the remaining weeks (13-25). Most of the strength increases occured in weeks 1-13 and were higher for 3 sets YET IGF-I was no different. Again, skill at the lifts and not superior hormonal production was elicited by multiple sets.

    Additionally, IGFBP-3 decreased by 20% in the multiple set group by 20% during the last half of the study but did not decrease in the 1 set group. IGFBP-3 levels increase in the presence of IGF1. If they decreased in the 3 set group, we can assume so too did IGF-I. Overtrianing?

    I’ll talk more about this meta in great detail on my blog at http://www.slowburn.typepad.com

  10. Mistress Krista says:

    September 23rd, 2009at 2:17 pm(#)

    Fred: Awesome! Thanks for the additional input.

  11. Fred Hahn says:

    September 23rd, 2009at 2:46 pm(#)

    You’re welcome. It is more important to look at each study critically to determine if the study itself has flaws in its conclusions. One of the 9 studies (McBride et.al.)that James says are new and support the idea that greater strength is built with more sets concluded that: “It is possible that some type of neural mechanism is responsible for the observed difference.” Here they are talking about strength. They conclude this because there were no difference in the amount of lean mass built.

    The bottom line is positive tissue adaptation. If 2 trainers train identical twins and one twin becomes twice as strong as the other but the other twin develops more lean mass than the other, THAT is the twin that got the best results from the training. Who cares how much strength one can demonstrate? And the greater # of sets allows for greater skill acquisition especially in untrained subjects.

    Lean mass increases and not strength increases should be the determining factor when comparing single to multiple sets.

  12. Mistress Krista says:

    September 24th, 2009at 5:57 am(#)

    I’m interested in the fact that you use lean mass rather than strength gains as the criteria. Why is that?

  13. Fred Hahn says:

    September 24th, 2009at 9:32 am(#)

    Good question Krista.

    The reason is that strength can be achieved via increases in actual lean tissue as well as simply learning how to perform and exercise properly.

    EX: Olymplic lifting is a failry complex series of movements that enable lifter to hoist weights over their head that would be impossible to do without knowing how do coordinate the leverages of the body in a manner that minimizes the weak links and maximizes infinite lever.

    An O lifter could never press over head what they can clean overhead. They need to drop/duck underneath the weight swiftly, lock out their limbs and then extend their legs to lockout. This is s complicated skill.

    At first, a newbie O lifter will struggle with relatively light weights even if she is strong. If the coach is good and the time is spent, within a month or two she will perhaps triple the weight load she can lift but at the same time very little if any muscle will be built.

    The purpose of Olympic lifting is NOT to build muscle although it will most definitley do so aftet a time. However, O lifting also incorprates squats and other exercises that are desinged to build strength and size so it’s tough to say what is doing what.

    So increases in strength, in fact, large increases in demonstrable strength can occur with little if any concomitant increases in lean tissue.

    The goal of training is to improve upon the tissues of the human body – bone, muscle, fat, etc. Strength is a by product of these improvements but not necessarily vice versa. If I double a client strength but do not alter lean mass or fat mass, what have I done for them? Nothing. Many of the studies included in Krieger’s meta reveal just that – greater increases in strength for multiple sets but NOT greater increases in lean mass and in some cases, not in strength either in severl exercises.

    In fact, I am reading one of the studies Krieger said showed superior outcomes in strength from multiple sets right now (Galveo, Taaffe) where strength was greater in the 3 set group but only in 3 of the 6 exercises and lean mass increases were identical. Who cares if the 3 set group got strnger in a few of the exercises – this is easily attributed to learning. Clearly it was NOT due to increases in muscle as both groups gained an identical amount.

    So, this really shows single sets to be superior as it takes less time to achieve the same degree of lean tissue.

  14. James Krieger says:

    October 12th, 2009at 3:24 pm(#)

    Ahhh, Fred Hahn, a guy who criticizes me on his forum for going around and “policing the internet” yet here he is doing the same thing.

    ************
    if you look at the studies included in the meta you will see glaring errors in the researchers conclusons.
    *************

    Let’s see the errors, then.

    **************
    Many of the studies included in this meta are poorly conducted.
    **************

    Define “poorly conducted”.

    **************
    Several of the studies showed a greater strength increase for the multiple set group but no significant difference in lean body mass increases.
    ****************

    That’s irrelevant, because the analysis was on strength, not lean body mass.

    Second, measures of lean body mass are highly insensitive, and are subject to the large error rates that body composition methods have. Thus, the likelihood of seeing significant differences in lean body mass over a typical 12-15 week study period, when comparing two different training programs, is small.

    ******************
    This tells you that more sets = more skill acquisition
    ********************

    No, it does not. First, this is a statement for which you have no scientific support.

    Second, it is a logically indefensible statement upon closer inspection. The idea that an improvement in 1-RM in a leg press from 200 lbs to 300 lbs is all “skill acquisition” is absurd. This would imply that increases in dynamic force production are always the result of skill and nothing else.

    ***************
    One study (Borst et. al.)showed the same increases in circulating IGF-I for both 1 and 3 sets for the first 13 weeks and no further increases for either 1 or 3 sets for the remaining weeks (13-25). Most of the strength increases occured in weeks 1-13 and were higher for 3 sets YET IGF-I was no different. Again, skill at the lifts and not superior hormonal production was elicited by multiple sets.
    ******************

    This is another incorrect statement. First, this statement assumes that circulating IGF-I is responsible for changes in strength. There is no evidence that this is true. In fact, muscle cells can produce their own IGF-I locally. Circulating levels are irrelevant.

    Second, the statement is a non-sequitur. Even if circulating IGF-I played a role in strength gains, this does not mean that skill then must be the causative factor. Fred’s statement here assumes that IGF-I is the ONLY hormone that plays role. There are many other hormones that impact strength gains. This statement also ignores the locally produced growth factors and signals for protein synthesis that are stimulated by resistance training.

    The fact is, unless Fred has evidence that the molecular and cellular signals that regulate protein synthesis are identical between single and multiple sets, and unless he has evidence that the additional dynamic force production observed in my analysis was all due to skill, he has no argument.

  15. James Krieger says:

    October 12th, 2009at 3:39 pm(#)

    ***************
    One of the 9 studies (McBride et.al.)that James says are new and support the idea that greater strength is built with more sets concluded that: β€œIt is possible that some type of neural mechanism is responsible for the observed difference.” Here they are talking about strength. They conclude this because there were no difference in the amount of lean mass built.
    *****************

    Lean body mass did not increase in EITHER group in this study.

    Second, the study only lasted 12 weeks. As I stated in another post, such a time period is too short to detect meaningful differences in lean body mass changes, given the insensitivity of lean body mass measurements.

    Third, there in fact was a small trend (but insignificant) for lean mass to be gained in the legs of the multiple set group, but not in the single set group.

    Given the small number of subjects, the insensitivity of measures of changes in lean body mass, and the variability in the responses, this study is simply inadequate to address the issue of multiple sets on lean mass changes (very few studies have sufficient subject numbers and are of a sufficient length of time to address this question).

    **************
    Who cares how much strength one can demonstrate?
    **************

    Given that it is dynamic force production that applies to real-life activities, it certainly does matter how much strength you can demonstrate.

    *************
    Lean mass increases and not strength increases should be the determining factor when comparing single to multiple sets.
    ****************

    Irrelevant. The study was on dynamic strength changes, and was designed for that purpose and that purpose only. The fact that it did not look at LBM changes is hardly a criticism of the paper.

    Second, meta-analyses on LBM are problematic. First, there is the insensitivity and error rates of the measurements. Second, there is the fact that different researchers use different methods to measure LBM. There is simply no way to do a good meta-analysis on LBM changes when comparing single to multiple sets.

  16. James Krieger says:

    October 12th, 2009at 3:56 pm(#)

    *************
    EX: Olymplic lifting is a failry complex series of movements that enable lifter to hoist weights over their head that would be impossible to do without knowing how do coordinate the leverages of the body in a manner that minimizes the weak links and maximizes infinite lever.
    ****************

    Olympic lifting is a very poor example in trying to support your “skill” argument, since Olympic lifts are, by definition, extremely high skill lifts.

    The lifts used in my meta-analysis are nowhere close to the skill levels needed for Olympic lifts.

    *******************
    If I double a client strength but do not alter lean mass or fat mass, what have I done for them?
    ********************

    Given that strength plays an important role in every day activities, you have done quite a bit for them.

    **************
    In fact, I am reading one of the studies Krieger said showed superior outcomes in strength from multiple sets right now (Galveo, Taaffe) where strength was greater in the 3 set group but only in 3 of the 6 exercises and lean mass increases were identical. Who cares if the 3 set group got strnger in a few of the exercises – this is easily attributed to learning
    *********************

    You are ignoring the important issue of statistical power.

    Most resistance training studies are underpowered to detect significant differences, particularly given the large variability in responses.

    Let’s take two hypothetical populations of 1000 people. One population does 3 sets of a leg press and improves by 100 pounds. The other population does 1 set and improves by 140 pounds. That’s a 40% difference. The response is quite variable and the standard deviation for both groups is 50.

    Now, let’s say you take a random sample of 10 people from each of these populations. The statistical power to detect the 20% difference is only 56%. This means, if you do 10 studies using 10 people each, you will only see a significant difference in 5-6 of the studies.

    Thus, your statement regarding the fact that Galvao & Taaffee only saw a significant difference in 3 of the 6 exercises is not a valid criticism, because you are not addressing the issue of statistical power. In fact, one of the reasons for doing meta-analysis is to see if there are trends among a large body of underpowered studies. This is exactly the case here, and thus multiple sets, on average, do result in greater strength gains than single sets.

    *************
    Who cares if the 3 set group got strnger in a few of the exercises – this is easily attributed to learning.
    ****************

    A statement for which you have no evidence.

    Second, you have not defined “learning.” The fact is, if dynamic force production increases due to neural changes within the muscle, then it is still a positive adaptation, and the muscle can now produce more force. People don’t just care about LBM as you seem to imply.

  17. David Yang says:

    August 6th, 2010at 10:56 am(#)

    There is an additional consideration: frequency.

    Precisely, the number of sets per workout vs the number of workout per weeks. That is, after how many days can you recover to perform the next workout.

    If by doing just one set for a workout you can train three times a week, vs only twice a week if you do three sets per workout, then you may gain more with less lifting.

  18. Ivan Madrigal says:

    November 22nd, 2010at 7:51 pm(#)

    This is a very interesting thread. I am currently completely perplexed. I have ready body by science by John Little. I followed something similar to this book, by doing 1 set per week per exercise for legs. I did 1 set for squats, 1 set for leg extensions, 1 set for calf extensions and 1 set for hamstrings. My strength increased to double what I had ever been able to lift in my life. So much so that I can’t believe it and it is my own body. Yet my upper body has not responded nearly as well. So I am looking into MAX-OT training by Paul Dilea of http://www.ast-ss.com and Jeff Willet the IFBB Natural Pro Bodybuilder also swears by. They go with multiple sets in the 4 to 6 rep range. I was trained to do 12 for lower body and 10 for upper body but an NFL Strength Coach that trains very similarly to Body By Science methods. This is nuts to me that we can’t scientifically come up with the answer. Could it be that it is dependent on every person? For example, why is my lower body getting freakishly strong. My bench is on a best day 245 for max. Hasn’t gone up much. While I am doing plate loaded V-Squat machine for 630 pounds. More than anyone I’ve even seen in the gym. Help! :)


Get "Fuck Calories"

Enter your information below and the magical gnomes that run Stumptuous will send you a copy of the "Fuck Calories" e-book for free!
Name:
Email Marketing by Javelin

Share