Perturbation training for novice archers: Aslan 2025

Twenty-five pupils from a middle school in Kırıkkale, sixteen girls and nine boys, none of whom had ever shot before, allocated at random to two groups after basic shooting instruction. The perturbation group, fourteen people, does exercises in which, while drawing the elastic band with the shoulder at ninety degrees of abduction as in the shooting position, they receive pushes in unpredictable directions and at unpredictable moments on the drawing arm and the pushing arm, having to resist and carry on drawing and aiming; plus a medicine ball hold against the wall, ninety seconds of work and thirty of rest. The comparison group, eleven people, spends the same time on elastic band pulls and drawing the bow without an arrow. The number of arrows shot was kept equal between the two groups.

The outcome is the score over six arrows at ten metres, chosen for the group's level, measured once at the end: there is no baseline measurement, because the design is post-test only.

The result is null. The perturbation group scores 16.07 points with a standard deviation of 7.80, the comparison group 15.27 with 4.59, and the comparison gives p equal to 0.77. The authors say so plainly, both in the abstract and in the conclusions: the results do not support the research hypothesis.

Having written that, though, they add three observations in favour of the intervention. That the perturbation group's mean is higher. That the three highest scores of the whole test, 32, 25 and 22, belong to participants in that group. And that the highest mean belongs to the girls of that group, which may point to the benefit of perturbation training in archery.

The article's arithmetic is consistent: from the two means, the two standard deviations and the two group sizes, t comes back at 0.30 on twenty-three degrees of freedom, that is p equal to 0.77.

The research took place in the second semester of the 2023-2024 school year at a middle school chosen because it had suitable facilities. Informed consent and a safety rules form signed by the parents. Approval from the non-interventional research ethics committee of Kırıkkale University, with the date and decision number reported, and a declaration of no conflicts of interest.

The design is declared as a post-test only model with a control group. The pupils were allocated to the groups at random, and which of the two was the experimental one was also decided at random.

Inclusion criteria: voluntary participation, no chronic illness, right-handedness, and no previous archery instruction. Excluded were those with a history of upper limb fracture, those breaking the safety rules, those not attending regularly, those who could not be measured and those who did not hand in signed consent.

On the number of participants the authors are explicit and it is worth reporting: they aimed for at least thirty people per group, that being the threshold above which normality can be relied on, and wanted sixty pupils in all. Fifty turned up on the first day, thirty carried on, five were excluded because their score could not be measured, and the research closed with twenty-five. The sessions were in the afternoon and the pupils were diverted to other courses.

The exercises took place at three stations: shooting, elastic band and empty bow pulls, and perturbation training. The perturbations were applied after the shooting, in three rounds, each exercise lasting a minute and repeated three times; the medicine ball hold ninety seconds with thirty of rest, one set of three repetitions. The balance of load between the groups was sought deliberately: while the perturbation group worked, the other did elastic band and empty bow pulls, and the arrows shot were kept equal.

The shooting test was adapted to the level: six arrows at ten metres. The authors cite the literature according to which the distance and arrow count of such a test are chosen from the group's level.

The data collected also include demographic information, previous sporting experience, the presence of chronic illness and the determination of the dominant eye.

The analysis is in SPSS 22. Normality was assessed with the Shapiro and Wilk test and with skewness and kurtosis coefficients, read against the criterion of the range between minus and plus 1.5; the values reported for score, weight and height sit within that range. The comparison between groups is an independent samples t test.

How long the intervention lasted is reported nowhere in the article. The week counts appearing in the text, six, eight and sixteen, all belong to studies on other sports cited in the discussion.

The project was funded under a TÜBİTAK programme for undergraduate student research, and presented as an oral communication at a sports science congress in December 2025. Received on 9 December 2025, accepted on the 18th, published on the 31st. A regional venue.

The flaw lies in what comes after the null result, not in the result. The between-group comparison gives p equal to 0.77 and the authors declare it without equivocation. Then, in the abstract and the conclusions, they add that the three highest scores of the test belong to the perturbation group and that the highest mean is that of the girls in that group, and they write that this may point to the benefit of perturbation training. Neither observation is evidence. The three highest scores are three people out of twenty-five, picked by looking at the data after seeing that the comparison gave nothing; and the perturbation group was the larger, fourteen against eleven, so it had a greater chance of containing them. The probability that the top three all fall in the group of fourteen, by chance alone, is 15.8 per cent: it happens about one time in six.

The same holds for the female subgroup. With eight girls per group, comparing the girls' means after the main comparison came out null is a second attempt on the same data, and it is not declared as a planned analysis.

The design is post-test only, and the authors write this as a choice. With random allocation it is a legitimate choice in itself, but with fourteen and eleven beginners any difference in starting aptitude ends up inside the result with nobody able to see it, and the data to correct for it do not exist. In a group of youngsters who had never shot, initial aptitude varies a great deal: the scores show it, running from very low values up to 32 out of 60.

The outcome is six arrows at ten metres, once. The mean is around sixteen points out of sixty, that is less than three points per arrow, and one group's standard deviation is 7.80, half its own mean. With six arrows the luck component of a single end is large against the difference being sought, which here is eight tenths of a point. It is not a measure able to tell two training methods apart.

How long the intervention lasted is not written. It is the most important piece of information for judging a training programme, and it does not appear in the article: the weeks to be read in the text concern studies on volleyball, taekwondo and neck musculature cited in the discussion. The conclusions recommend extending the duration, so it was probably short, but by how much is unknown.

The attrition is large and declared: from sixty pupils wanted to fifty present on the first day, thirty carrying on and twenty-five analysed. The authors write themselves that the threshold of thirty per group, which they had set, was not reached. Those who stayed, in an optional afternoon activity, may be those who were more interested or more gifted.

Two imbalances between the groups remain untreated. Sex: six boys out of fourteen in the perturbation group, three out of eleven in the other. And spread: a standard deviation of 7.80 against 4.59, that is in the perturbation group there are one or two participants far above the rest, who are the same ones carrying the scores of 32, 25 and 22, and it is on those that the higher mean rests.

Twenty-two days separate receipt from publication, with nine days from receipt to acceptance.

This study has a virtue that is rare in this literature, and it deserves naming first: it kept the number of arrows equal between the two groups. While one group did the perturbations, the other did elastic band and empty bow pulls for the same time, and everyone shot the same number of arrows. This is exactly the control missing from four of the studies the registry read in the previous batch, where the intervention included extra arrows and the gain in score could not be attributed to anything else. Here the comparison is clean on that point.

And with that clean comparison the result is null: p equal to 0.77. Adding exercises in which someone pushes your arm while you draw did not score more points than adding elastic band pulls, in this group and with this measure.

What cannot be drawn from it is that those exercises do not help. Twenty-five middle-school youngsters who had never shot, six arrows at ten metres as the only measure, no baseline measurement and an intervention duration the article does not state: this is a design without the sensitivity to see a small effect. A null result of that kind should be read as an absence of information rather than as information.

The part not to pick up is the last page. After the null, the authors note that the three highest scores of the test belong to the perturbation group and that the girls in that group have the highest mean, and from this they draw a possible benefit. The three highest scores are three people out of twenty-five, and the group containing them was also the larger: the probability that they all fall there by chance is about one in six. Anyone citing this study as a hint in favour of perturbations would be citing that page rather than the comparison.

The registry keeps this work as a randomised comparison between two active programmes in beginners, with a null outcome and with the arrow count controlled, and for the order of magnitude of a beginner's score: less than three points per arrow at ten metres after basic instruction alone.