This study explored the influence of a reflective-impulsive approach on wholist and analytic processing, by counterbalancing presentation of the embedded figures and matching figures subtests of Riding's cognitive styles analysis (CSA). One hundred and ninety three participants completed a variation of the CSA test. The results revealed that the wholist-analytic ratio is extremely sensitive to the order in which each subtest is presented. Significantly higher ratios are produced when the matching figures subtest is presented first and lower ratios are produced when the embedded figures subtest is presented first. This reflects a general tendency for respondents to react more slowly to early test items. However, this tendency to produce slower response times during early test items is greater for individuals with a reflective style. There was significant interaction between the presentation order of the subtests and individual differences in reflective-impulsive style. Reflective individuals were significantly more analytic than the impulsive individuals when the matching figures subtest was presented first and were marginally but not significantly more wholist when presentation order was reversed. The implication is that the methodology of the CSA is likely to inflate differences between wholist and analytic individuals by comparing differences in part-whole processing and concomitant differences in reflective-impulsive style.