You are here

Methodology

Methodology

The target population was identified as all pilots who were considered "active" pilots by the FAA at the time of the survey. In order to be active, pilots must have completed a physical examination and filed the requisite form with FAA during the 24 months prior to the survey (physical exams are required every 24 months to keep a pilot license active). At the time of the survey, there were 138,465 pilots with Air Transport certificates, 117,891 pilots with Commercial Pilot certificates, and 241,280 pilots with Private Pilot certificates on the Civil Airmen Registry. A total of 5,400 active pilots were randomly drawn from the registry. The sample was stratified by type of certificate 1 (private, commercial, air transport) and by geographic location (the nine geographic regions serviced by FAA). Data for this summary have been weighted to reflect the population figures at the time the survey was conducted.

The survey is two pages long and consists of 16 items (see Appendix A). The cover letter urged pilots to "have a say" in what topics and methodologies are used for future safety seminars. Initial surveys were mailed to the random sample of pilots the week of November 18, 2002. A second wave of surveys was mailed two weeks after the first mailing to all non-respondents (less any undeliverable addresses). A follow-up postcard was mailed three weeks after the second mailing and gave pilots a toll free phone number to call if they needed a new packet to complete the survey. Data collection was terminated on March 17, 2003.

The final response rate for this survey was 62%, computed by dividing the number of completed surveys (n=3,326) by the number of eligible pilots (n=5,375). Pilots who are deceased (n=12) or no longer piloting (n=13) were classified as ineligible for the survey.

1Although some airmen may hold more than one type of certificate, for this report, airmen were grouped by the type of certificate listed for them on the Civil Airmen Registry at the time the sample was drawn.

The findings summarized in this report are estimates derived from a sample survey. Two major components of error in sample surveys are sampling and nonsampling error. Sampling error occurs because findings are based on a sample, rather than on the entire population. Tables showing the margin of error for estimates are in Appendix B. Estimates are subject to various nonsampling errors during the survey process, such as measurement error (errors in response coding, transcription, and data editing or question wording) and nonresponse error, which is a function of both the nonresponse rate and the differences, if any, between respondents and nonrespondents. Stringent quality control procedures were followed during the survey process in an effort to minimize nonsampling errors.