Shooting the bow from both sides: Schmitt 2021 study

Twenty-five competitive archers who for at least a year have also trained shooting from their non-habitual side were compared with fifty who shoot from one side only, matched one to two for age, sex, training hours per week and years of practice. Mean age 47.4 ± 13.9 years, 60 men and 15 women, ten years of practice on average, 2.4 sessions and 4.3 hours a week; the disciplines were 3D archery for 46.3%, target archery for 39.2% and field archery for 14.5%. In a single visit they had their cervical range of motion measured, along with shoulder rotation, upward rotation of the scapula during arm abduction and active shoulder mobility with the hands behind the back.

Nothing is found at the neck: no difference between the two groups and none between the two sides, in any of the three planes of movement, with eta squared always below 0.03. The authors note that these archers' cervical range of motion is about 20 degrees wider than the reference values for their age.

At the shoulder the differences are there. Those shooting from one side only have, on the string side, a greater total rotation than those shooting from both (148 degrees, confidence interval 144 to 152, against 140, 135 to 145) and greater internal rotation (61 degrees, 58 to 65, against 56, 51 to 61). Upward rotation of the scapula during abduction is wider in the one-sided archers at 45 degrees (12 against 8), at 90 (34 against 28) and at 135 (56 against 49), while at maximal elevation the difference falls short of significance. The ratio between arm movement and scapular movement differs between the two sides only in the one-sided archers (2.9 to 1 on the string side against 4.2 to 1 on the bow side), not in those shooting from both (6.1 to 1 against 5.3 to 1). Mobility with the hands behind the back likewise differs between sides only in the first group (22.0 centimetres on the bow side against 18.0 on the string side), and between groups it differs on neither side.

The shooting score, measured over nine arrows at 15 metres, does not separate the two groups: 15.7 ± 3.7 against 14.5 ± 4.4.

The design is cross-sectional with matching, which the authors call quasi-experimental. Eligible were healthy competitive archers over 18, with no bilateral training experience for the unilateral group and at least one year of bilateral training for the other. Excluded were a body mass index above 35, pregnancy or breastfeeding, and acute or unhealed injuries of the head, neck, shoulder or upper thorax. Recruitment was by word of mouth: 29 bilateral archers were screened and 25 included, and 760 unilateral archers were screened to find the 50 matches, using a standardised matching plan on age, sex, weekly hours and years of practice. Nobody withdrew consent and nobody was excluded afterwards; participants had no pain at rest or during the tests. All measurements were taken in one visit by a single examiner.

Cervical range of motion was measured with an ultrasound-based three-dimensional motion analysis system (Zebris CMS 70), seated, hands on thighs, feet on the floor, mouth closed, with three markers on a plate fixed to the head by a helmet-like carrier and a second marker set on the right upper arm as reference. For each direction, three familiarisation repeats and five measurement repeats, with a spirit level on the carrier to zero the starting position. Shoulder rotation was measured with the same system, with the plate on the distal border of the ulna, the arm at 90 degrees of abduction and the elbow at 90, guided passively by the examiner, with the scapula held by the Codman handle and the spine checked visually; the order of the two sides was randomised.

Upward rotation of the scapula was measured with a digital inclinometer applied to the spine of the scapula, at 45 and 90 degrees of abduction, at 135 degrees of elevation and at the maximum angle, with arm position controlled by the ultrasound system; the ratio between arm and scapular movement is obtained by dividing the abduction value by the scapular rotation value. Active shoulder mobility is the distance between the two fists brought behind the back, standing, with the position held for at least five seconds and the movement fluent, measured with an electronic caliper; three trials per side, and for every kinematic measure the best value was kept.

Shooting was measured on a ground-level target at 15 metres, chosen so as not to favour target archers, since the sample also contained 3D and field archers. The target was a one-metre disc with a five-colour face 40 by 40 centimetres and ten scoring zones. After warming up and six arrows to get used to the distance, nine arrows were scored in three ends of three, each archer using their habitual equipment; to make different bows comparable, scores were multiplied by a conversion factor taken from the literature, from 0.65 for a compound with a sight to 1.0 for a wooden or primitive bow.

The analysis is in SPSS 24, with two-by-two repeated measures analysis of variance (group by side), main and interaction effects reported, and paired or unpaired post-hoc comparisons where effects came out significant. The threshold is 5%.

The design does not allow the conclusion that shooting from one side only produces the asymmetries. It is cross-sectional: those who train from both sides may have chosen to for reasons that have to do with their own shoulder, or may differ from the start. The authors say so, and add two points that weigh. The matching used total years of archery, on average 11.2 in the bilateral group, but those archers had been shooting from both sides for only 4.6 years on average: for years they had loaded one side like everyone else. And the asymmetries could be a direct effect of the position held during the shot, which those alternating sides hold for less time.

There is one variable on which the two groups do not coincide, and it is not a minor one: pull-side dominance differs significantly (p < 0.01). In the unilateral group 90% draw from the right, in the bilateral group 64%, with a different share of cross-dominance. It is the one matching variable that did not hold, and it touches exactly on how the shoulder is set.

Two of the results carried in the abstract do not pass the filter of the overall analysis. For internal shoulder rotation the analysis of variance finds neither a group effect (p = 0.2) nor an interaction (p = 0.4), and even the side effect stays above threshold (p = 0.07); for total rotation there is no group effect (p = 0.1). The values 61 against 56 degrees and 148 against 140 degrees come from post-hoc comparisons run anyway, after an omnibus test that did not license them. The scapular differences, by contrast, have full group effects at 45, 90 and 135 degrees, with eta squared from 0.07 to 0.2, and are the study's firmest result; at maximal elevation, though, the difference stopped at p = 0.06.

How much any of this matters for shoulder health remains open, and the authors say as much: it is not known whether these asymmetries are still physiological or already pathological. The thresholds they cite, a loss of internal rotation above 18 degrees and a loss of total rotation above 5, are taken from other sports; in their sample internal rotation differs by 4 degrees between sides and total rotation by more than 9, so one threshold is crossed and the other is not. The argument they close with, that the asymmetry cannot be useful for performance because the scores do not differ, does not hold: the shooting test was nine arrows and lacks the precision to separate two groups. The mechanism proposed, an imbalance between the serratus anterior and the upper trapezius, is declared by the authors themselves to be a hypothesis, because muscle activity was not measured.

Some details remain. The sample averages 47 years of age, ten years of practice and 2.4 sessions a week: these are club-level competitors rather than top athletes, and 43% have a history of shoulder complaints, with no difference between groups. Recruitment was by word of mouth, with 760 unilateral archers screened to find 50. The measurements were taken by a single examiner, who knew each person's group, and there is no blinding. The ultrasound system's sampling rate is given as 30 Hz in one place and 100 Hz in another. The shooting score, 15.7 against 14.5, has no clearly stated scale: nine arrows over ten zones would give a maximum of 90, and the conversion factor that lowers compound scores is a non-standard correction taken from another publication. The text carries several misprints, including in the abstract's conclusions.

The clearest result concerns the neck, and it runs the opposite way to what many expect. Although shooting loads the two sides differently, these archers' cervical range of motion shows no asymmetry at all, in any plane of movement, and comes out wider than the reference values for their age. Those shooting from one side only do not have a stiffer or more lopsided neck than those who alternate.

The shoulder is another matter, and the point on which the study is firmest is the scapula. In archers shooting from one side only, the two sides move differently: at 45 degrees of abduction the scapula rotates 12 degrees on the string side against 8 in those who alternate, and the ratio between arm movement and scapular movement is 2.9 to 1 on the string side against 4.2 to 1 on the bow side, while in those who alternate the two sides stay together. Translated: in the drawing arm the scapula contributes relatively more than the arm, and that difference between the two sides disappears in archers who also shoot from the other side.

How much this matters for the shoulder is not something this study says. None of the archers had pain during the tests, and 43% of the whole sample had a history of shoulder complaints, equally in both groups: the difference in movement is there, the link with pain or injury was not measured. The authors write that they do not know whether these asymmetries are still normal or already a problem, and when they suggest adding a few arrows from the non-habitual side they present it as a cautious choice rather than a treatment.

On the practical side, one thing is worth knowing: shooting from the non-habitual side did not lower scores. The two groups shot the same in this study's test, and those who alternated had 4.6 years of bilateral practice behind them on average. That is reassuring for anyone who fears ruining their shot by trying the other way round, with the caveat that the test was nine arrows at 15 metres and would not pick up small differences.

One figure from the introduction, useful for a sense of the loads involved: a competition day means about 144 arrows, and adding up the draws an archer pulls more than 2.7 tonnes in a day, always on the same side. It is not a measurement from this study but from a source it cites.