Arm strength, draw length and archery: Humaid 2014

On 62 male recurve archers measured at an Indonesian national championship, the author builds a path model between four variables: arm muscle strength, draw length, shooting technique and score in the FITA round.

The model's two equations are: technique equals 0.53 times arm strength plus 0.45 times draw length; score equals 0.36 times arm strength plus 0.15 times draw length plus 0.50 times technique. Squaring the coefficients, the author assigns 28.09% of technique to arm strength and 20.25% to draw length; on score he assigns 12.96% to arm strength, 2.25% to draw length and 25.00% to technique. He then states joint contributions of 84.00% on technique and 91.00% on score. Four of the five coefficients are given as significant at 1% and one at 5%, but the t values appear only inside two figures.

The four variables are presented as standardised scores: means of 50.06, 50.03, 49.95 and 49.98, maxima of 80, 72, 70 and 76, minima of 29, 28, 26 and 33. The reference score is the full FITA round: 144 arrows, 36 at each of 90, 70, 50 and 30 metres, on a 122 cm face at the two long distances and an 80 cm face at the two short ones.

In the order the model returns, technique weighs more on the score than arm strength does, and draw length comes last of the three.

A cross-sectional study with path analysis. The data come from the Gelora Bung Karno archery ground in Senayan, Jakarta, where the 2010 national championship was held. The starting population is the 200 male athletes entered, from fifteen provincial associations; the 62 in the sample were chosen on a criterion set by the author rather than at random.

The three independent variables are arm muscle strength, draw length and shooting technique; the dependent variable is the score in the FITA round. The paper defines draw length as the extension of the drawn bow until the string reaches the nose, lips and chin, measured from the nock groove to the arrow rest, and technique as the application of the nine stages of the shot.

The construction of the instruments is described in four generic steps: composing the variable indicators, composing the instrument lattice, pre-testing the instrument, checking its validity and reliability. Nothing is said about what device measured arm strength, in what position or in what units, nor how draw length was recorded, nor who rated technique and on what scale.

The analysis was run in SPSS 17. Before the hypothesis tests, the normality of the error estimates was checked with the Liliefors test, along with the significance of the regression coefficients and linearity; path analysis was then applied, decomposing the direct effects among the four variables.

The limitation that weighs on everything else is that the instruments are not described. Without knowing how arm strength, draw length and technique were measured, the coefficients cannot be checked, replicated or turned into anything concrete: there is no draw weight, no dynamometer reading, no centimetres of draw.

Some of the figures do not stand up. The standard deviations reported for the four variables are 51, 48, 44 and 43, which are impossible for measures with a mean of about 50 and a range from 26 to 80. The joint contributions stated, 84.00% on technique and 91.00% on score, do not match the sum of the squared coefficients the author himself calculates, which comes to 48.34% and 40.21%. The t values do not appear in the text, only inside two figures, so the significance claimed cannot be checked by reading the paper.

The design is cross-sectional: the arrows in the model, that is the direction of the effects, come from the author's hypotheses rather than from the data. The chain the paper proposes, a longer draw giving a faster arrow, a flatter trajectory and less wind drift, is stated as a theoretical premise and never measured: there are no arrow speeds in this study.

The sample is 62 men, not randomly chosen, from a single championship in 2010, in a single discipline. The author himself notes at the end that female archers were not included and that a comparable study on them would be needed, and suggests adding other biomotor, anthropometric, mental and equipment variables.

The paper is cited as evidence that arm strength and draw length determine the result. In its own model the order runs the other way: on score, technique has the highest coefficient at 0.50, then arm strength at 0.36 and last draw length at 0.15, that is 2.25% of the variance by the author's own arithmetic. Anyone using this source to say that a longer draw matters a great deal is reading it backwards.

For someone who shoots, nothing usable comes out of it, because there is no telling what was measured. There is no draw weight to compare with your own, no strength test to repeat, no draw length in centimetres. What remains is a ranking of three factors inside a group of 62 men at a national championship fifteen years ago.

The part that keeps being useful is the underlying idea, that strength matters mainly in an indirect way: having headroom over the bow's draw weight leaves attention free for technique. In this study that is a hypothesis stated in the introduction, not a finding from the measurements.