That math anxiety goes with weaker math performance is well established, and the interesting question is what sits between them. says anxiety pulls the attention system toward whatever feels threatening, which should show up as a cost to . Does the evidence support that?
Partly, and the sizes are worth seeing. Pooling 57 studies and 150 effect sizes gives a of minus 0.168 between anxiety and performance, which is real but modest. Only 15 effect sizes addressed working memory as the mediator, with a mean of minus 0.092, which the authors call descriptive rather than generalizable.
The negative relationship between math anxiety and math performance is well established, so the useful work is explaining it. Attentional Control Theory suggests anxiety impairs attentional control and increases attention to threat-related stimuli.
Within that framework, the link between math anxiety and working memory is argued to be critical for performance. If the mechanism is real, the anxiety and performance correlation should pass measurably through working memory.
The search turned up 1,346 unique articles, of which 57 studies and 150 effect sizes survived screening for the main question. A random-effects model gives a mean correlation of minus 0.168 between anxiety and performance.
The mediation question had much less to work with. Fifteen effect sizes addressed working memory as a mediator, giving a mean indirect effect of minus 0.092 that the authors describe as descriptive rather than generalizable. Their conclusion is that working memory does play a significant role, stated against that thin base.
wat. The anxiety and performance link is real but modest, and working memory carries part of it.
wat wat
Knowing the mechanism changes the intervention, since a working-memory problem and a fear problem call for different lessons.
How do we know?
Finell and colleagues searched databases with predetermined strings, screened 1,346 unique articles, and combined the surviving effect sizes in a random-effects model.
What is still unknown
The mediation part rests on 15 effect sizes, which the authors themselves describe as descriptive rather than generalizable.
Learn something?
Where to next
How much does believing intelligence can grow actually raise a student’s grades?education
Why can an arithmetic problem hurt before you have even started it?math anxiety
Does an ice bath after training actually speed up recovery?meta-analysis