This article presents a model using data from 205 telephone surveys conducted in the same survey lab over a three-year period. The model demonstrates that while response rates are partly a function of variables reflecting effort, they are also affected by contextual variables often not under the survey vendor’s control. Significant factors that affected response rates included the salience of the survey to the population, the survey length, the type of sample (listed vs random-digit dialing), minutes per piece of sample (effort), and the amount of time the survey was in the field. A ten-minute increase in survey length results in a 7%decrease in the response rate. An increase of one day in the field per one hundred cases (fielding time) results in a 7% increase in the response rate. An increase of one interviewer minute devoted to each piece of sample released results in a 2.2% increase in overall response rates and a 3.4% increase in random-digit dialing response rates.
Effort in phone survey response rates