Presentation on theme: "Evolution of Web Surveys at USDAs National Agricultural Statistics Service United States Department of Agriculture National Agricultural Statistics Service."— Presentation transcript:
Evolution of Web Surveys at USDAs National Agricultural Statistics Service United States Department of Agriculture National Agricultural Statistics Service Gail Gregory & Morgan Earp A presentation for the International Conference of Establishment Surveys, June 2007
Overview Background of NASS Web Response Rates over time. Web Respondent comments at two different time periods. Internet Usage Project-reasons why respondents dont report on the Web.
Survey Nations farms, ranches, agri-businesses Analyze, summarize data and set official estimates for all facets of agriculture HQ in Washington, DC and 46 field offices National Agricultural Statistics Service (NASS)
Up to 400 surveys a year Mixed-mode - some versions optimized for certain modes Annually, quarterly, monthly, weekly Diverse in terms of mode used (State to State) NASS Surveys
Evolution of the NASS Web Instruments Began Web Surveys in 2002. Cotton Ginnings Report Currently 125 surveys on line. Two distinct populations Agribusinesses Farmers/Ranchers
Web Response Rates Overall Web Response rate is 1.8%. Almost 1.5 million samples uploaded and over 27,000 completed. While the sample size uploaded to the Website is known, it is not known how many respondents were sent an invitation and instructions to complete the survey on the Web.
Web Response Rates over Time By July 1 2005, 57 Surveys on the Web. Overall Response Rate was 1.7% Response Rate for surveys targeting Farms/Ranches 1.2% Response Rate for surveys targeting Agribusinesses 6.6% By February 1 2007, 125 Surveys on the Web. Overall Response Rate was 1.8% Response Rate for surveys targeting Farms/Ranches 1.3% Response Rate for surveys targeting Agribusinesses 8.2%
Summary of Comments Respondent comments from all Web surveys from July 1, 2005 until January 30, 2007 were coded and summarized. Each of the 1,278 comments from 108 Web surveys was recorded into one of the following categories: Liked Web Site Web Complaints Survey Complaints Survey Comments Survey Refusal No Comment Called before chance to do on the Web Got someone elses Survey Code
Categories of Web Data Collection Site Responses Category Number of Comments 2007 Percentage of Comments 2007 Number of Comments 2005 Percentage of Comments 2005 Liked Web Site37129.09731.2 Web Complaints32425.49229.6 Survey Complaints28522.33812.2 Survey Comments23718.65919.0 Survey Refusal90.761.9 Reported No Comment272.1196.1 Called too Early221.7___ Got incorrect Survey Code30.2___ Total1,278100311100
Web Complaints By Number and Percentage CategoryNumber of Comments 2007 Percentage of Comments 2007 Number of Comments 2005 Percentage of Comments 2005 Experienced Log In Problems288.71314.3 Long Download Time4915.21516.4 Problems Printing154.61112.1 Lack of Edits4313.366.6 No use of Previously Reported Data 206.233.3 Web does not Match the Paper Form 4213.01415.4 Received Miscellaneous Errors 165.077.7 Paging Issues206.233.3 Data did not stay in the Form154.633.3 Navigational Problems5015.51314.3 Want Survey Code and/or Survey Results Emailed 257.743.3 Total323100.092100.0
Survey Complaints By Number and Percentage CategoryNumber of Comments 2007 Percentage of Comments 2007 Number of Comments 2005 Percentage of Comments 2005 Questionnaire Complaints27496.12463.2 Security Concerns about SSN62.1923.7 Survey Purpose Unclear51.8513.1 Total285100.038100.0
Initial Results from the Internet Usage Project Investigated why more respondents do not complete their surveys on the Internet. Questions were added to 3 surveys to discover why respondents didnt complete their survey on the Web. The surveys were: December 2005 Cotton Ginnings October 2006 Labor Survey (Blaise only) March 2007 Agricultural Survey (Blaise only)
Internet Usage Project Cotton Ginnings Agribusinesses Labor Very large Farms/Nurseries Small Farms March Agricultural Survey Large Crop Farms Small Crop Farms
Internet Usage Questions The three questions asked on the surveys were: 1.Do you have access to the Internet? Yes Don't Know No 2.[if 1 = Yes] Did you know this survey can be completed on the Internet? Yes Don't Know No 3. [if 2 = Yes] What was the main reason why you did not complete this survey on the Internet?
Results of Internet Usage Project 1-Do you have access to the Internet? Survey No. of States YesNoDont Know Total CountPercentCountPercentCountPercent Cotton Ginnings1750268.922731.1--729 Agricultural Labor462,73460.61,47932.82966.64,509 Agricultural Survey4818,17259.610,98536.01,3474.430,504
2- Did you know this survey can be completed on the Internet? Survey No. of States YesNoDont Know Total CountPercentCountPercentCountPercent Cotton Ginnings1736472.713727.3--501 Agricultural Labor4661122.31,67961.444416.22,734 Agricultural Survey486,30734.710,95760.39085.018,172
Coded Reasons for not Responding on the Web CodeReason Description Agricultural Survey Agricultural Labor Cotton Ginnings CountPercentCountPercentCountPercent 1Other modes (telephone, fax) are quicker or easier. 3004.855 9.575 25.5 2Does not like using computers / Internet.1,48223.411018.982.7 3Problem with the NASS EDR Website.1181.9193.3144.8 4Prefers a telephone call82413.0183.1103.4 5Telephone call is prompt to complete survey.2353.750.972.4 6Did not have enough time. / Too busy.1,38521.917730.54114.0 7Competition for use of computer.791.330.541.4 8User problem with computer2694.2234.062.0 9Fear of computer viruses.160.371.220.7 10(claimed) Completed on-line310.561.0-- 11Web notice was sent out too late.1973.18013.8-- 12The letter said someone would call.811.3264.5-- 13Did not think Survey applied to them.240.4101.7-- 14Other reason.84213.3193.3268.8 15No reason given. 4356.9234.010134.3 Total6,318100.0581100.0294100.0
Reasons for Not Completing the Survey on the Web By Type of Farm (Crop/Livestock)
Some Reasons for Not Completing the Agricultural Labor Survey on the Web by Value of Farm Sales
For the Future Adequately train all Field Office personnel as to the importance of Web data collection. Ensure that potential respondents know that the Web is an option, especially highlighting it in a cover letter. Provide well tested Edits in Web Surveys.