In a few weeks, the next US president will be elected and polling bias is a topic of discussion. Current predictions about the election are still considered unreliable. Today we are going to talk about how pollsters can create bias depending on the responses, research, coverage or by not getting any responses in their polls.
Gathering accurate data in a survey can be a difficult task, but understanding the audience and writing about the results can be much more difficult. The quality of the data is very important, we must ensure that it is accurate and collected in a limited period of time.
7 tips to avoid bias in online surveys
1.) Find the right sample size
Generally speaking, most national surveys have a sample size of between 500 and 1,000 people. This results in a margin of error between 3.2% and 5%. As the sample size increases, the margin of error decreases. However, as the sample size decreases, the returns decrease. So, for example: the margin of error is significantly reduced in a survey of 50 to 100 people than in a survey of 500 to 1,000.
I invite you to read: Calculate the sample size for a survey .
2.) Use the appropriate medium to ask your questions
Do you have a landline at home or at work? Are you the type to answer a call jamaica phone number from an unknown number? Do you live in a rural area where there is no telephone or internet? These questions can cause bias in surveys. If we talk about the upcoming US elections, for example, Republicans are more likely than Democrats to use a landline.
Internet accessibility, web scraping or data mining should help pollsters get accurate analysis of the election. Even though it may be nearly impossible to get a perfect representative sample, the Internet is used almost universally.
3.) Be direct and ask multiple choice questions
These types of questions give respondents the opportunity to have multiple options at their disposal, making their task easy. Be bold and make sure you are communicating what you are aiming for. A good multiple choice question should be mutually exclusive, avoiding ambiguity at all costs.
For example, if you ask about political affiliations, the correct thing to do is to include all political parties or independent candidates and not only give the most popular political parties as an option; if you do so, you will be creating a bias in your survey.
4.) Speak the same language as your respondents
Use simple, direct, and specific language in your questions. Start with basic questions and then move on to more specific ones. A great survey is read and understood. And unless the survey is about topics related to a specific industry or sector, avoid technical vocabulary at all costs as it can create confusion and therefore bias in the responses.