Abstract Two hundred forty high school students (120 male and 120 female) in India performed a moderately difficult multiple choice Arithmetic Reasoning task after undergoing short-term (40 minutes) cognitive treatment in the form of Attentional Skills Training. A 2 × 2 × 2 (Test Anxiety x Attentional Skills Training x Stress) design with separate analysis for boys and girls indicated these results: with intervention the high anxiety subjects under ego stress conditions, compared to their high-anxious control, low-anxious ego stress, or low-anxious control counterparts, reported the maximum significant improvement in performance on the Arithmetic Reasoning test. The low-anxiety subjects performed consistently well with or without treatment or stress conditions. The findings shed new light on the attentional theory of test anxiety, and it was reasoned that long-term effects of cognitive treatment be studied by using varied performance tasks (difficulty level controlled) on different gender and age groups across cultures.