Avoid Bias
When creating a survey, eliminate as many bias aspects as possible. A biased question is one that prompts them to answer a certain way. An example would be if it said how much do you agree that we should outlaw straws. They are telling you that straws are wrong and that you should say you don’t want them. There other ways to have a bias in surveys. The people you pick to take the survey and who responds. The order in which you do the questions. There is a lot to be aware of.
Response Bias
Any emotion in the wording of the question will be noticed. If picked up on, it will change the way people respond. A neutral question will get the best responses. Questions should also not be leading. This goes for the answer choice as well. The answers should cover every possibility. Not just the ones that you think will be popular. The wider the options, the more truthful responses will be received. Remove the brand names from the survey. Knowing who is asking the questions can make them feel like they need to answers kinder instead of their true feelings. Using a marketing group to do the research is a safe way of not having any brand name attached.
Sampling Bias
How you share your survey can determine who takes it. To avoid bias and get a range of participants, you need to use many methods. Getting a sample of the population can be tricky. Using only one way of administering the survey means that you could end up with only one sub-group being surveyed. To overcome this and avoid sampling bias, send your survey in as many ways as possible. Send them in the mail. Send a link in a text message. Send an email. Do phone calls and interviews. The diversity will grow with every method you use.
Nonresponse Bias
Getting the questions written just right is the first battle. The second battle is sending it out to as many people as you can. The next action is getting as many as possible to respond. Unfortunately, it is hard to get this to happen. There are a few things you can do to try and limit the nonresponse rate. Try sending out a pre-email with information on the survey. Get them thinking about it before you send it. After they are given the study, send our reminders to return it. We are all human things slip through the cracks. A reminder is always useful.
Order Bias
Just like an exam, we sometimes base our answers on the previous questions. The order matters. Group questions by topic so that they are answered together and don’t affect other issues. Within those topics, put the questions in random order. Unless they are follow up questions. From one question to the next, you do not want them the effect the way they respond.
Create the best survey possible with no bias involved.