A study was performed, comparing two identical systems, which were equipped with different interfaces. One was a conventional data interaction, written in "standard" terse neutral English. The other was an anthropomorphized, conversational style interface. Users performed a set of tasks and gave feedback on the effect and affect of the systems. Our hypotheses were (1) that the affective measures would be improved in the conversational situation but (2) the standard interface would score higher on effective measures. The first hypothesis was weakly supported by the results; however, there was no support for the second hypothesis.