4.2 When, Where, and How Many: Survey Best Practices

Protocol and procedures for collecting community car seat use data in your community

In the following video we will find out:

1. Where to collect your car seat use observational data
2. When to collect your car seat use observational data
3. What type of data collection form works best for your data needs
4. How many observations to conduct
5. What supplies are needed for collecting data

Collecting information on 100 children will give you useful data. If you want more statistical precision, read on.

Where to Conduct Your Survey – 

An observational site for conducting your survey should be neutral and allow for comfortable discussions.

Native CARS tribes found that the best place to ask people to participate in your survey is in parking lots of:

• Grocery stores
• Gas stations
• Trading posts
• Schools/preschools
• Tribal clinics
• Community centers
• Tribal events & celebrations

We found trading posts, grocery stores and gas stations to be ideal sites. People here were less likely to be in a hurry, there was a more continuous flow of vehicles, and outside of school hours, children of all ages were equally likely to be found there. The parking lot is a neutral location, and speaking to people there won’t disrupt the business or organization. The surveys are done when cars are stopped, either as a vehicle first pulls in, or right before the vehicle departs.

Make a list of all potential observation sites and attempt to collect data at each location. Contact the business or agency before you visit to make sure you have permission to do surveys in their parking lot. Most schools require observers to sign in at the office and identify themselves with a volunteer badge. Be sure to respect the policies of the organizations you are visiting.

 You will want to limit the number of surveys collected at preschools and elementary schools to less than 40% of your total sample to avoid having too many surveys of a specific age group of children. So, if you do 100 surveys, make sure at least 60 of them come from places other than preschools and schools.

When to Do the Survey – 

Spring and fall are the best times. For the safety of data collectors, conduct surveys  during daylight hours. Since the surveys are done outside, winter might not be the best time for the survey. Spring and fall are good times because the weather can be milder and schools are in session.  If you conduct the survey during the summer, keep in mind that there are seasonal differences in how people travel, especially when school is not in session.  Whenever you choose to do your survey, it is best to repeat the survey during the same season in the future to obtain a reliable comparison over time.

How Many to Complete – 

Collecting information on 100 children will give you useful data. If you want more statistical precision, read on.

Your sample size depends on the size of your population and the proportion of children who are properly restrained. Here are a few steps to help you determine your sample size.

1.       Define your population

Native CARS used “Children age 8 & younger traveling in tribal communities.” You could choose to collect data on Native children only. Or clinic users. Look at your list of potential observation sites and think of the children you might observe there.  Also, think of who you are trying to reach with your intervention activities. This will help you define your population.

 2.       Estimate how many people are in your population

Tribal enrollment numbers, census data for your area, or clinic user population can help you with a ballpark estimate.  If you do not know the population of children in the area, but do know the entire community population, you can estimate that children age 8 and younger make up 17% of the total population in the average American Indian community, according to the U.S. Census. So, your total population x .17 = the number of surveys to do. 

https://factfinder.census.gov/faces/nav/jsf/pages/index.xhtml

 3.       Estimate the proportion who are properly restrained

You are probably thinking, “If I already knew this, I wouldn’t be doing this survey!”  But for the calculator, we need to make a guess. The safest guess is 50%. The farther you get from 50%, in either direction, the fewer observations you need to do to get a valid estimate. So, if you really have no idea, pick 50%.

The sample size table below will tell you how many children you need to observe to estimate the proportion of children who are meeting restraint recommendations with 5% precision (margin of error) and a 95% confidence level. If your population is somewhere between these numbers, your sample size will be too. For example, if your population has 300 children age 8 & younger, and you think half of them are properly restrained, look at the sample size for a population of 250 (152) and do a few more than that, say 170. Alternatively, you can use the web calculators below to get specific estimates for your population.

Keep in mind that your sample size is the number of children to observe. You will likely see multiple children in one vehicle. You will reach your goal numbers faster than you might think.

Number of Children to Observe for an Estimate with a 5% Margin of Error and 95% Confidence Level

📊 EpiTools Sample Size Calculator

📊 Sample Size Calculator by Sampsize Project

For information on who to interview, and guidance on conducting interviews, go to the next section.