Coping skills and archery scores: Taha 2018 study

Thirty-two young archers from the development programme of the Terengganu sports council in Malaysia (24 boys and 8 girls, aged 13 to 24, mean age 17) completed the Athletic Coping Skills Inventory and shot six arrows at 50 metres. The aim was to find out which psychological skills accompany higher scores.

Across the seven aspects of the questionnaire, a principal component analysis groups three of them in the first component with high loadings: coping with adversity (0.88), concentration (0.83) and confidence together with achievement motivation (0.82); that component alone covers 42.18 per cent of the variability. The second holds coachability (0.72) and freedom from worry (0.81); the third holds peaking under pressure on its own (0.98). The three components together reach 78.84 per cent.

A cluster analysis then divides the archers into two groups, one of high and one of low performance, using the questionnaire answers and the score together. A discriminant analysis correctly reassigns 93.55 per cent of the archers to those two groups, and in the high group the means of all seven psychological scales are higher than in the low group. Of the 32 archers, 18 end up in the high group and 13 in the low one: one fits neither and is dropped. Scores over the six arrows run from 17 to 57 out of 60, with a mean of 45.19.

The archers come from a university and state development programme, aimed at promotion to state and national level. The questionnaire is the Athletic Coping Skills Inventory, which measures seven aspects: coping with adversity, coachability, concentration, confidence and achievement motivation, goal setting and mental preparation, peaking under pressure, and freedom from worry. The authors report internal consistency between 0.67 and 0.82 in their sample.

The shooting test took place in an area set up to simulate a competition: four practice arrows, then six recorded arrows at 50 metres. The score is the sum of those six.

The processing has three stages. First a principal component analysis with varimax rotation, keeping components with eigenvalues above 1 and treating loadings above 0.70 as relevant; the overall sampling adequacy measure is 0.60. Then a hierarchical cluster analysis, which builds the two groups from the similarity between archers across all the variables, questionnaire and score together. Finally a discriminant analysis on the groups so created, run in three ways (standard, forward and backward), measuring how many archers are reassigned to the group they came from. The computations were done in XLSTAT at the five per cent level, after a normality check with Shapiro-Wilk.

The 93.55 per cent that the paper puts in its abstract is not predictive accuracy. The two groups had been built by the cluster analysis using the very same variables, score included, that the discriminant analysis then uses to separate them: the figure says that the division is consistent with itself, not that a questionnaire can tell how an archer will shoot. No sample was held back for validation, and the group labelled high performance is defined partly by the score, so finding that it has higher scores follows from the way it was created.

The performance measure is extremely fragile: six arrows, a single end, at 50 metres. Scores range from 17 to 57 out of 60, a spread that over six arrows can come from a windy day or a couple of mistakes, and there is no repetition from which to judge how stable that number is.

Sampling adequacy is 0.60, the minimum usually accepted, and for two of the seven scales the individual values are very low (0.21 for peaking under pressure and 0.34 for freedom from worry), which the text interprets as an absence of collinearity, a different matter. Several inconsistencies remain: the abstract speaks of three parameters containing six and three components; the variance of the first two components is 62.75 per cent in the text and 62.95 in the table; the text says eight independent variables and the table says seven; in the forward stepwise part of the discriminant analysis the two table rows appear swapped; and the discussion presents the second component as though it concerned attention and concentration, when the high loadings there belong to coachability and freedom from worry.

The sample is 32 young people from a single programme, with ages from 13 to 24 pooled together, and no mention of bow type. The order of the two tests is contradicted: the method says the questionnaire came before the shooting, the discussion says the scores were recorded first. There is no DOI, the journal issue is a special one, and the registry flags the venue for questionable editorial standards.

For an archer, the study offers neither a profile to imitate nor a tool to use. The sound part is descriptive: in a group of 32 young Malaysian archers, the three questionnaire items that move together are coping with adversity, concentration and confidence, while peaking under pressure stands apart and does not go with the others. That is an indication of how those answers are organised, not of what produces points.

The part to leave alone is the 93.55 per cent. It does not mean that a questionnaire can tell who will shoot well: the groups had been built using the score as well, so the calculation turns back on itself. Anyone looking for tools to select young archers will not find evidence here that they work.

One methodological note is worth keeping for reading other studies of this kind. When a paper measures performance with six arrows, any link with a questionnaire rests on a number that, for a young archer at 50 metres, varies a great deal from one end to the next. In the registry this source counts as weak evidence, to be set beside studies that measured scores over full and repeated ends.