Interactive web questionnaires promise to improve survey measurement relative to more static modes, whether online or paper. By designing questionnaires that react to respondent actions, it may be possible to promote behavior that leads to more accurate responses. We are investigating one type of interactivity, namely giving feedback to respondents (―speeders‖) when they answer so fast they cannot realistically have read the question let alone thought about the answer (―You seem to have responded very quickly. Please be sure you have given the question sufficient thought to provide an accurate answer.‖). In prior research (reported at AAPOR, 2009) we observed that some speeders answered questions about quantities (e.g., ―Overall, how many overnight trips have you taken in the PAST 2 YEARS?‖) more slowly when they were prompted once or twice; hardcore speeders never slowed down. We were encouraged by the slowdown for at least some respondents but could not be sure it reflected higher quality data. In the current research (n=2565) we explored the relationship between response time and quality by prompting speeders on simple numeracy items for which we could determine response accuracy (e.g., ―If the chance of getting a disease is 10%, how many people out of 100 would be expected to get the disease: 1, 10 or 20?‖). Prompting slowed some speeders AND increased their response accuracy, suggesting that in the earlier studies the slowdown also improved data quality. The benefits of prompting seemed to persist beyond the intervention itself as it reduced straightlining on later grid questions. Psychologically, interactive interventions seem to work differently than simple instructions. We randomly asked half of the respondents to commit to answering conscientiously; this slowed their responses but did not increase accuracy or interact with prompting. Overall, the current work illustrates the promise of interactive interventions for improving online measurement.