From draw to release in archers: Tinazci 2011 study

A conference paper presented as a study of four male Turkish Olympic archers, with electromyography, force plates, an electronic home trainer for aiming and a clicker chronometer, 30 arrows each at 18 metres indoors and 108 arrows analysed across eights, nines and tens. The figures, tables, charts and discussion coincide with those of a 2010 paper by Erkek and Tinazci, in which the same four people are described as women archers. The table of physical characteristics is identical in the two texts: age 25.7 years, weight 54.0 kg, height 168 cm, draw weight 16.9 kg. In the discussion, which has been left unchanged, it still reads that the fourth archer's performance improved when she swayed less.

The results are therefore the same. The activity of the finger flexors in the second before the clicker is lower on the ten-scoring arrows: 80.0 per cent of maximal contraction, against 84.4 on the nines and 85.6 on the eights (F 5.4, correlation of minus 0.29). The total draw time, from the first activation of the trapezius to the clicker, is shorter on the tens: 4.8 against 5.0 and 5.6 seconds (F 5.0, correlation of minus 0.28). Deltoid activity in the second after the clicker falls from 31.6 to 29.7 per cent.

Reaction time to the clicker does not tell the scores apart: 143.4, 142.5 and 142.3 milliseconds, with an F of 0.05 and a correlation of minus 0.03, contrary to what the abstract states. Aiming time on the target does not change (3.6, 3.7 and 3.7 seconds), nor does the wander of the aim, the extensor or the trapezius, nor the weight taken off the legs after the release (7.0, 6.8 and 6.4 kg). Lateral postural sway appears as correlated with the score, but with means that are not in order (31.7, 12.5 and 16.7 mm) and no significant difference between the groups.

The stated method is that of the 2010 paper, repeated almost word for word. Four archers of international level, right-handed, shooting recurve, 30 arrows at 18 metres indoors. Surface electromyography at 600 Hz on flexor digitorum superficialis, extensor digitorum, posterior deltoid and trapezius of the drawing arm, normalised to each muscle's maximal voluntary contraction. The 2011 version adds a few lines on signal processing: the trace is averaged over fifteen seconds covering the last seconds of full draw, the aim and the first milliseconds of the release, and the two seconds around the clicker are rectified and averaged over 100-millisecond intervals.

Total draw time runs from the first activation of the trapezius to the clicker's snap, full draw time from the trapezius maximum to the snap. The force plates record lateral and fore and aft sway during the draw and the weight taken off the legs after the clicker. The electronic home trainer records the horizontal and vertical displacement of the aim on the target and the time spent on the target, and starts when the transmitter on the bow stays on the target for more than a tenth of a second. The shooting chronometer measures the time between the clicker's snap and the release to within a millisecond. Analysis by correlation, analysis of variance with Bonferroni correction and t-tests, threshold 0.05.

In the registry this paper and the 2010 one should be read as a single set of data. The bibliographic entry, the methods and the lines stay as recorded; what changes is the weight: these are not two independent studies confirming one another.

The main limitation is that the paper brings no new data. Tables, values, standard deviations, numbers per score group and the text of the discussion are those of the 2010 paper in the biomechanics congress proceedings, published under Erkek and Tinazci. Here only Tinazci appears as author, the title says male archers and the text has been adjusted accordingly, but the table with age 25.7 years, weight 54.0 kg, height 168 cm and draw weight 16.9 kg is still the one for the four women, along with a feminine pronoun left standing in the discussion. The line about an earlier study has also been edited badly: where the 2010 text read eight female and eight male archers, the 2011 text reads eight male and eight male. The sex of the sample is therefore not established by this text.

All the limitations of the original remain. Four participants and a single session; 108 arrows used as independent observations although they come from four people; twelve arrows below the eight ring excluded without that being stated; draw weights varying by a hundred grams across four athletes; correlation signs in the individual table that contradict the direction of the means; and the statement about clicker reaction time that the tables do not support.

There is no measure of grouping, no competition condition and no wind: only the score of the single arrow at 18 metres indoors. And there is no comparison between levels: all four participants are at the same level, so the study can say nothing about what separates a strong archer from a weaker one.

For anyone who shoots, what can be taken away is the same as from the 2010 paper, not twice as much. On the ten-scoring arrows the string hand is less active in the second before the clicker, 80 against 86 per cent of maximal contraction, and the draw time is shorter, 4.8 against 5.6 seconds. These are two associations in four archers in one session, not two experiments.

Reaction time to the clicker, in these data, does not tell an eight from a ten. This paper is often cited for the opposite, because the abstract says so; the table does not.

The rest is a caution about how a bibliography is read. When two references report the same numbers, one does not confirm the other: they count as a single measurement. In the registry they count once, and the line that holds them does not become sturdier because the reference turns up twice.