This study investigates the effects upon response distributions of two types of position effect, the serial order in which response options for an item are presented and the location of an item in an interview. Primacy effect was statistically significant and meaningfully large in regard to evaluative items in an interview with people aged 65+; either the positive or negative end of the scale drew more responses when it was presented first. This effect was not observed for factual items. Neither was there any difference in responses to an evaluative item presented early and late in the interview, but with response options in the same order at both presentations. The results suggest that it may not be necessary to present general questions prior to specific ones in order to avoid response contamination, though further evidence is necessary on this point. The findings are more clear-cut in regard to the order of presentation of response options for a given item. To obtain valid group values, counterbalanced presentation is necessary; and standard scores for individuals should be used in making subgroup comparisons and other statistical procedures such as correlational analysis.