Increasing Accuracy While Collecting Data
A vital part of any opinion research is the human response. Whether you are conducting a survey or a poll, collecting accurate data from respondents involves a well-planned strategy. A survey allows you to ask multiple questions across a wide range of topics, with answer options. For example, in a survey you can ask for a comment, or contact information, as well as offer multiple choice questions. In order to achieve powerful results, the goal is to make sure that the respondents believe their opinions matter, and that those opinions are captured in an accurate and measurable manner.
There are many strategies that can ensure your data is accurate, including these:
1. Use Script Logic
Know who is taking your survey. Offer sub–questions in addition to the previous question to have a better understanding of who you are surveying. For example, say you give a yes or no question for your first question, Q1. If they answer yes, direct them to answer Q1B. If they say no, then require them to go to question 2.
2. Be Unbiased
Some questions may come off as biased. All items must be unbiased to have accurate data.
Posing questions like, “How would you rate our friendly service?“ is a biased approach. Using descriptive words shows a bias. Refrain from adjectives or from descriptive words to keep it as unbiased as possible. You want your survey-takers to feel comfortable in offering an honest opinion.
3. Use a Likert Scale
All of the answer choices must have a specific structure. A Likert scale, which is used to scale survey responses, is commonly used in questionaries so many people feel comfortable with it. Using a 5–Likert scale makes the options more specific by including five answer choices with a neutral answer choice in the middle. For example:
How would you rate the city air pollution from 2016 until now?
A. Much better
C. About the same
D. Somewhat worse
E. Much worse
A 7-point Likert scale is also an option, and it is considered the most accurate. For a 7–Likert scale, there are seven answers, with a neutral break in between the answer choices.
A Likert scale keeps order within the questions and increases the accuracy of your results.
4. Don‘t Get Lost in the Weeds
Lengthy questions can lose traction with your audience. Adding too much detail can cause confusion and a possible loss of engagement.
Example: “People often like to say their favorite fruit is apples. What would you say your favorite fruit is?” This question adds an unnecessary explanation.
The respondent might feel they need to answer “apples” because that is the standard answer. On the other hand, “What would you say your favorite fruit is?”. This question is short and straightforward. Survey takers feel like they can answer the question without being compelled to provide an expected answer. There are ways, such as proofreading and allowing neutral observers who are not familiar with the project to review the script, to see if the question makes sense.
5. Proofread Your Script
Always go through the script as if you are going to answer the questions. Putting yourself in the respondent’s place helps make the script cohesive and simple. If the script does not make sense to you, it will not make sense to the people answering. Finally, test your script. Testing your script can be beneficial because it will expose different points of view, and having these varied perspectives will provide insight into the script.
Unbiased, clear, and logical survey questions – that have been reviewed and edited – create a precise outcome and fully engage respondents. Taking the time on the front-end to make certain your survey is clear eliminates confusion. Surveying should be straightforward and simple, and by using the strategies above a survey will provide accurate results.
Here at OHPI, we provide data analysts that ensure our surveys are useful, dependable, and accurate. Click here to discover how our services will help your organization move to the next level.