Maths anxiety is a form of anxiety specific to being placed in situations that require mathematical knowledge and it is associated with impaired performance on mathematical tasks. There has been much debate about the relationship between maths anxiety and maths performance. Correlation is not causation. Does impaired performance cause maths anxiety or does maths anxiety cause impaired performance? Perhaps the relationship is two-way, setting-up a vicious circle for the individuals involved?

Dr. Jo Boaler, a professor of maths education, has been prominent in the discussion about maths anxiety, attributing its cause to the use of timed assessments in class. However, I have previously investigated this claim and I see no reason to accept it at present.

Nevertheless, maths anxiety exists and is a significant problem. As a maths teacher, I want to do something about that. Until now, the research on how to deal with maths anxiety has not been hugely helpful, but a new study by Maria Passolunghi, Chiara De Vita and Sandra Pellizzoni may start to change that.

To begin, I want to applaud the design of the study. Far too common in education research are comparisons between doing something and doing nothing. Doing something may turn out to be better than doing nothing just because of a placebo effect. So, a better design of study will include an active control group. Doing something that is thought to have an effect is then compared with doing something else that is not thought to have this effect but that still feels like something new to the participants. Better still are three-armed trials where two separate interventions are compared with an active control.

A three-armed trial is the only truly valid way of comparing the effects two different interventions. Over the years, attempts have been made to compare interventions from different studies using effect sizes, but effect sizes are not the stable measure that people think they are and can vary wildly depending on the subjects, the measures used, the experimental design and a whole host of other factors. This is why I have argued for three-armed trials in the past.

The Passolunghi et al. study on maths anxiety is just such a three-armed trial. Three programmes were developed for primary school students. All of them involved eight weekly one-hour sessions. The active control involved children creating comic strips. Of the two interventions, one focused on dealing with maths anxiety directly by addressing negative thoughts with a form of cognitive behavioural therapy. The other addressed maths anxiety indirectly by building maths skills and focusing on maths games, calculations and rhymes and stories for remembering maths facts.

The results were interesting. The direct approach to tacking maths anxiety reduced maths anxiety relative to the active control condition. However, the maths skills intervention resulted in pretty much the same reduction in maths anxiety. This is an unusual example of transfer – something that is rare in education research. However, the researchers also measured mathematics performance using a standardised assessment. The students who received the maths anxiety intervention did not improve their maths performance any more than the control, but the students in the maths skills intervention improved their performance more than both the control and the maths anxiety intervention.

Perhaps teaching children maths results in them improving at maths which, in turn, results in them being less anxious about maths.

It seems plausible.