Developing methods to compare low-education community-based and university-based survey teams

We compared the quality of low-education community-based survey teams to college educated graduate students. Our approach was to develop methods, conduct a pilot survey, and report lessons. Community and university teams conducted surveys from non-overlapping random samples of addresses at a public housing development in Boston, Massachusetts. The two types of teams make a similar number of attempts (122 and 124, respectively), and there was no statistically significant difference between the teams in terms of response rate or amount of missing data. Similarly, there was no statistically significant difference in refusal rate or in responses to questions in the survey. There was, however, evidence that the community teams used data tracking forms improperly. This study suggests that it is possible to study the relative quality of community and university-based teams in terms of data collection. The findings also suggest that the two types of teams may be roughly comparable.