Imagery and self-talk in archers: Isar 2022 study

Forty-five archers at a Malaysian sports school, on average fifteen years old, were split into three groups of fifteen, each with one kind of mental training: imagery alone, imagery plus instructional self-talk, and self-talk alone. Eighteen sessions over six weeks, three a week of an hour each. There was no group that did nothing.

The score over six ends at 70 metres rose in all three, but with differing significance: from 257.5 to 287.9 with imagery plus self-talk (p 0.011), from 256.5 to 280.0 with self-talk alone (p 0.047), and from 254.7 to 271.4 with imagery alone, which is the only increase that is not significant (p 0.153). Between the groups, the authors state explicitly that there is no difference on the score, and explain it by the fact that the athletes all come from the same school and are of the same standard.

Reported somatic anxiety collapses in all three groups: from 27.3 to 12.8, from 26.9 to 11.9 and from 27.1 to 12.3, on a scale whose highest possible score is 28. Cognitive anxiety falls with imagery plus self-talk (from 12.8 to 9.9) and with self-talk alone (from 12.7 to 10.1), while in the imagery-only group it rises, from 13.1 to 13.9. Self-confidence grows in all three, but by very different amounts: from 7.9 to 18.4 in the combined group, from 9.8 to 12.8 with self-talk alone, and from 7.9 to 8.9 with imagery alone.

In the questionnaires on mental skills there are movements the design does not account for: the imagery ability score rises in the group that did self-talk only (from 1,722 to 2,639) and the self-talk score rises in the group that did imagery only (from 20.0 to 23.7). The only between-group statistical test reported in the paper is an overall multivariate analysis, F(12,74) equal to 41.05: no direct comparison between one group and another, on any measure, appears in the results.

The participants are archers at a state sports school in the Pahang region of Malaysia, selected from among those with no injury in the previous six months who passed a screening test on imagery ability. They are not beginners: one represented the state, twenty-seven the national level and seventeen the international level. The mean ages of the three groups are 15.6, 15.4 and 14.6 years.

The instruments are three questionnaires translated into Malay: one on imagery ability in sport, one on belief in the effectiveness of self-talk, with eight items divided between positive and negative self-talk, and the revised version of the competitive state anxiety questionnaire, seventeen items measuring cognitive anxiety (five), somatic anxiety (seven) and self-confidence (five). The anxiety questionnaires were given 30 minutes before the score was recorded. Performance is the total of six ends shot outdoors at 70 metres on a 122 cm face.

The intervention lasted six weeks, with eighteen sessions, three a week of an hour each. The paper reports nothing of their content: there are no imagery scripts, no instructional self-talk phrases, no account of who ran the sessions, and no check that participants did what was intended. The analysis is a two-way mixed multivariate analysis, with a threshold of 0.05. The abstract calls the design quasi-experimental and in the same sentence says the 45 participants were randomly distributed into the three groups.

The comparison that would be needed is missing. All three groups received an intervention, so the improvements in score cannot be separated from six weeks of ordinary training, from familiarity with the 70 metre test done a second time, or from the attention received. With no group doing nothing, three curves rising together do not say which treatment did what.

The between-group comparison was never carried out. The results contain one overall multivariate analysis and then, in the table, before and after comparisons within each group. The conclusion in the abstract, that the combined group did better than the other two, rests on no test: it rests on the fact that in the combined group the p value of that within-group comparison is smaller, which is not the same thing. The discussion of the same paper says the three groups do not differ on the score.

Some of the numbers do not stand up. Baseline somatic anxiety is 27.3, 26.9 and 27.1 out of a possible maximum of 28: that would mean almost all forty-five, on every item about physical sensations, had chosen the most intense response. The imagery ability score in the table runs into the thousands (from 1,621 to 3,405), while the methods state that only those scoring between 150 and 400 were included: a factor of ten between the screening and the results. And the abilities rise in those who did not train them: imagery improves in the self-talk-only group, self-talk improves in the imagery-only group.

There is no blinding of any kind, no follow-up, no account of what the sessions contained and no check that participants actually did any imagery or any self-talk. Six ends outdoors is few arrows for a score, and the weather conditions of the two tests are not reported.

The journal is flagged in the registry as a venue with doubtful editorial standards, and the text bears that out: the introduction moves from anxiety to target panic by citing a newspaper article, sentences are often incomplete, and in presenting the design it contradicts itself within a single line.

For anyone who shoots, this study does not establish that mental training produces more points, and it does not establish which form of mental training is preferable. With no comparison group receiving nothing, and no test between the groups, the three increases in score remain three groups shooting better after six weeks than they did six weeks earlier.

What can be said is where things went less well: in the group that did imagery only, the score is the one that fails to reach significance, cognitive anxiety is the one measure that rises, and self-confidence moves least. That is a hint, not a demonstration, that imagery on its own gave less than it did when paired with instructions said to oneself.

It is also worth saying what this study is not. The introduction brings in target panic and defines it, citing a piece of qualitative research and a newspaper article, but the study contains no measure of target panic at all: no early releases, freezes or punched shots were counted, and no criterion for recognising them is given. Anyone citing this work for target panic is citing an introduction, not a result.

In the registry it counts as weak evidence, in a venue with doubtful editorial standards, on young archers who are nonetheless of national and international level. Its main interest is that it measures the score at 70 metres, which many studies of the mental side do not, and that the paper itself admits to finding no difference between the three treatments on that score.