Agricultural Statistics

Survey response rates for telephone surveys have been declining over the last several decades.Two leading sources of nonresponse are refusals and inaccessibles.While the amount each factor contributes to nonrepsonse can vary across surveys, refusals often account for a larger portion of nonresponse.At the National Agricultural Statistics Service (NASS), refusal and inaccessible rates have steadily increased over time, although refusal rates tend to be higher than inaccessible rates.However, a different trend has emerged for the Quarterly Agriculture Survey (QAS).Refusal rates increased steadily from 1991-2001, after which they began to decline.During this same time period inaccessible rates increased and in 2008 surpassed refusal rates.Two factors occurred during this time period that may have caused inaccessible rates to rise.First, there was rapid rise in the availability of call screening technology.Respondents, who are increasingly burdened with surveys and telemarketer calls, may be using this technology to screen their phone calls.Second, NASS moved toward centralized data collection.As a result, respondents for the QAS were more likely to be called from NASS' National Operations Center (NOC) in St. Louis rather than state field offices.Calls placed from the NOC display a Missouri (MO) telephone number with the identifier "Ag Counts" or some other text such as "unknown number" on respondents' caller ID, which may influence their decision to answer the phone.In this study, we examined whether calling respondents using in-state area codes and recognizable identifiers improved the number of answered calls and whether it improved response rates for the September and December 2012 QAS.The first of two experiments was conducted on the September 2012 QAS.In this experiment half of the Nebraska (NE) sample assigned to the NOC was called using an in-state NE area code and half of the sample was called using an out-of-state MO area code.Using an in-state area code (NE) did not improve the response rate for this survey in September.The second experiment was conducted on the December 2012 QAS.Two states were selected for this experiment -Ohio (OH) and Minnesota (MN)in order to attain a larger sample size.In both states, half of the sample assigned to the NOC was called using an in-state area code and half was called using an out-of-state MO area code.Calling respondents using an in-state area code led to a small, but significant improvement in answered calls.However, respondents who were contacted using an in-state area code were more likely to refuse to participate in the survey than those called using an out-of-state MO area code.Respondents in this experiment were also surveyed on their use of caller ID.Approximately 1/3 of the sample had caller ID and reported looking at it before answering our call.Respondents reported seeing a variety of information displayed on their caller ID, and 49 percent of these respondents indicated that the information displayed on their caller ID influenced their decision to answer the call.These respondents indicated that they were more willing to answer the phone when the call was identified as coming from an in-state area code (MN/OH), Ag Counts, or USDA, and less willing to answer Will They Answer the Phone If They Know It's Us?