A windshield survey relies on observations for data and other information instead of directing questions to participants. The windshield surveys got its name because many of these projects are done while the observers sit in a car. You also can perform a windshield survey by walking through the targeted area.
Develop survey questions that fit the definition and capabilities of a windshield survey. For example, if you want to assess the recycling program in your town, you would write down questions related to the number of recycling bins each house puts out and the types of material you see--and don't see--in the bins.
Gather together maps of the areas you will be surveying. Map out the route you will take through the neighborhood or other area in question, along with an alternate route. Determine what time period(s) the survey will need to be taken and how many survey takers you will need.
Create a team of people who will perform the windshield survey. One person will need to drive the car as one to two others gather the data you need. You may consider adding another person to the team to operate audio-visual equipment if you need still or video pictures as part of the survey.
Equip your survey team with plenty of supplies, such as pens, clipboards and paper on which to take notes in addition to the survey questions. Don't forget water and snacks if the team will be on the road for long periods of time.
Be sure each team member knows her job as you begin the windshield survey. So the survey takers can make their observations, you will need to drive as slowly as possible, but avoid holding up traffic, especially in densely-populated areas. Survey takers should take as many notes as possible about what they see in the neighborhood, even if the observations may not seem important or seem to fit the questions.
Discuss your findings only after you have completed the windshield survey, not during the drive. This avoids allows everyone on the team time to process what he has seen without confusion and avoids distracting the driver.
A walking survey requires fewer persons per team--no driver, for example--but each team will cover less territory.