Estimating eligibility rates: Crucial component of the calculation for response rates

Response rates are used by analysts to assess survey quality: higher response rates usually are desired to reduce the incidence of nonresponse bias. The response rate is defined as the ratio of the number of completed interviews divided by the number of eligible sample units. However, because of the inconsistency of the definition of response rates often quoted in travel surveys, it is difficult to state explicitly that declining response rates are the result of fewer people willing to participate in surveys or are attributed to the calculation of response rates. Most likely it is a combination of these two factors. Two well-known formulas used to calculate response rates are described: the Council of American Survey Research Organizations formula and the American Association for Public Opinion Research formula. The difference between these formulas lies in the estimate of eligible sample units among the sample units of unknown eligibility. Through examination of two call history files, the recruitment phase for two household travel surveys, eligibility rate estimates for the sample units of unknown eligibility were calculated and used in the response rate formulas. The rates of eligibility for the sample units of unknown eligibility were higher than the eligibility rates for the units of known eligibility. These results were not expected and further confirm that agencies must treat units of unknown eligibility carefully when calculating response rates.