Medicine ball for archers: Wattimena 2025 study

Thirty Indonesian university archers between nineteen and twenty-two, with at least three years of practice and three training sessions a week, randomly assigned to three groups of ten: a three-kilogram medicine ball, a four-kilogram medicine ball and a control group carrying on with shooting practice alone. Eight weeks, three sessions a week, five throwing exercises, ten repetitions per exercise, two to three sets. Those taking the measurements did not know which group an athlete belonged to, and only participants with at least ninety per cent attendance entered the analysis.

The outcomes are three gymnastic tests. One-minute push-ups and sit-ups for endurance, and the stork test, a single-leg hold, for balance. In the two trained groups push-ups go from 38 to 46 and from 36 to 44, sit-ups from 56 to 64 and from 57 to 63, the hold on the left leg from 41 to 52 and from 42 to 53 seconds, on the right from 43 to 55 and from 41 to 52. In the control group push-ups go from 36 to 39, sit-ups from 56 to 59, the hold from 41 to 46 and from 42 to 47 seconds, and those changes are declared non-significant.

Between the two loads there is no difference, and the article says so plainly: one extra kilogram changes nothing. That is the answer to the question in the title, and it is a useful answer, because it closes a choice a coach might believe matters.

What is missing is the reason the study was set up. The last sentence of the introduction sets out the aim as weighing up what medicine ball work does for "the balance and accuracy" of archers. The procedure says that in the ninth week an accuracy test at thirty metres was carried out. The statistical analysis says a paired t-test was used to evaluate the changes "in accuracy and balance" from before to after. In the results there is not a line about accuracy: no table, no figure, no number, no mention in the discussion or the conclusions. A test on the target was run and not reported.

An experimental design with three groups, measurements before and after, and random allocation. The population is university archers aged nineteen to twenty-two; selection is purposive, with three criteria: being an active athlete training at least three times a week for the past three years, having no injury history, particularly in the upper body, over the past six months, and being willing to complete the whole programme and all the tests. Ethical approval from the State University of Jakarta. Thirty participants, ten per group.

The programme is the same in the two trained groups, only the weight of the ball changes: overhead throw, chest pass, side twist throw, squat and push throw, lunge and pass. Ten repetitions per exercise, two sets in weeks one and two and three sets from the third to the eighth, intensity declared as moderate to high, ten minutes of warm-up and five to ten of cool-down, under the guidance of a certified archery coach. The control group carries on with its normal shooting practice with nothing added. Attendance was monitored by the coaches.

The baseline measurements include height, weight, body mass index, body fat percentage, one-minute push-ups and sit-ups, and the stork test. The three groups start out equivalent on age, weight, height and body mass index, with p values between 0.228 and 0.424, and the article publishes them: that is good practice, and here it is observed. Body fat percentage, declared measured, appears in no table.

In the ninth week the final measurement is taken, which according to the procedure comprises an accuracy test at thirty metres and the stork test, without repeating the anthropometric measures. Of the final push-ups and sit-ups, which are then reported, the procedure says nothing.

The declared analysis is a two-way repeated measures analysis of variance for the between-group comparisons, the Shapiro-Wilk test to check "the homogeneity of the data", and the paired t-test for the before-and-after changes. The stork test is not described: one cannot read whether it is eyes open or closed, with hands on hips, with the heel raised, how many trials are run and which is kept. No result of the analysis of variance is published, and every outcome sits in two bar charts plus a description in words.

The test on the target was run and not reported. Three places in the article say so: the aim declared at the end of the introduction, to evaluate the effect on balance and accuracy; the procedure, by which the ninth-week final measurement comprises an accuracy assessment at thirty metres; and the statistical analysis, where a paired t-test is described for the before-and-after changes in accuracy and balance. In the results accuracy does not appear, and it does not appear among the limits the authors acknowledge either. Worth noting too: accuracy is absent from the list of baseline measurements, so a before-and-after change could not have been computed in any case. What remains is a measurement collected on the target about which nothing is known, and in the register that is the measurement that counts.

The second limit is that no number in this study can be recomputed. Every outcome sits in two bar charts and their description in words; there are no standard deviations, no t values, no degrees of freedom, no exact probabilities, no effect sizes and no repeated measures analysis of variance of the kind the methods declare. The twelve means given in the text are all whole numbers, that is rounded: with ten people per group a mean lands on tenths, so those values are read off the figure rather than being the actual values. Any check is therefore closed to a reader.

The third concerns the between-group comparison, which is the answer to the question in the title. The methods say: "a paired sample t-test was used to compare endurance and balance at posttest between groups". A paired test compares two measurements on the same people and cannot compare different groups, so the sentence describes an operation that does not exist. That there is no difference between three and four kilograms is probably right and hardly surprising, but the test by which it was established cannot be read.

The fourth is the control group. Its changes are declared non-significant, and they are three push-ups, three sit-ups and five seconds of hold on each leg. Five seconds, on a base of forty-one, is forty per cent of the trained groups' gain, and without a standard deviation there is no way to know whether, with ten people, that change really is consistent with chance. The sentence in the conclusions, by which no substantial change in balance can occur without a specific intervention, is stronger than those five seconds allow.

The fifth is the distance between the outcomes and shooting, and it should be acknowledged that the authors write it themselves among the limits. It is worth putting concretely: push-ups are a push, while the string arm pulls; sit-ups are a flexion of the trunk, while at full draw the trunk stays upright and works against rotation; the stork test is single-leg balance, while shooting is done on two feet. None of the three tests reproduces the geometry of the movement, and the throwing programme, sensible as it is as general work, contains nothing asymmetrical.

The sixth is the stork test values, which start between forty-one and forty-three seconds in all three groups and reach fifty-two and fifty-five. Those are very long times for that test, which usually ends far sooner, and the protocol is not described: eyes open or closed changes everything, and without that detail the numbers compare with no norm and with no other study.

Finally, the fact that everybody improved their push-ups by a third and their hold by a quarter in eight weeks, in the throwing group as in the other, on a light load and three sessions a week, is a large gain for athletes already training for three years. It is not impossible, since gymnastic tests respond a great deal to familiarity with the test itself, and that strengthens the point about the control group: it too repeated the same tests eight weeks apart.

Two things are usable straight away.

The first is the answer the article really gives: between a three-kilogram ball and a four-kilogram one, over eight weeks, nothing changes. Anyone having to buy a medicine ball for their group has no choice to make, and anyone who already has one has no reason to change it. Within that range, weight is not the variable to spend attention on.

The second is the dose, the one piece described in full: three sessions a week for eight weeks, five different throws, ten repetitions per exercise, two sets for the first two weeks and three from the third, a ten-minute warm-up and a five to ten-minute cool-down. It is a light load, to be slotted in after shooting or on a day of its own, and the authors recommend forty-eight to seventy-two hours between sessions. For a university group that is realistic.

What cannot be taken away is the link to the score. The outcomes are push-ups, sit-ups and a single-leg hold: three gymnastic tests, none of which reproduces the act of shooting. The string arm pulls and push-ups push; at full draw the trunk stays upright and sit-ups flex it; shooting is done on two feet and the stork test measures one leg. The authors acknowledge this among the limits, which is to their credit, and the thirty-metre test on the target, which according to the procedure was run in the ninth week, appears in no result.

In practice: if the medicine ball goes into the programme as general strength and coordination work for university archers shooting three times a week, this article gives a plausible dose and says not to worry about the weight. If it goes in as an intervention to raise scores, there is nothing here to support it, because the scores were not published.