Writing a Survey 101:
Press 1 Campaign - Best Practice
You may be thinking, how complex could surveys be? Surveys can be very complex in the designing stages depending on the level of depth the conductor wants to achieve.
1. Writing a survey begins with the research goals.
What do you want to discover, learn, or know from the results of the survey? I.e. I want to know what the top three factors go into consideration when a consumer is buying body lotion.
2. Turn the research goals into research questions.
Having your research goals written out will help you stay focused and on track when writing. There are many types of questions to ask beginning with Testable which can be organized by statistics- best for decision makers- and Non-Testable which cannot involve a statistical test, but may include very important information with specifics (best for administrators and planners).
Types of Questions:
- Dichotomus: Yes or No
- Multiple Choice
- Rank Order Scaling: Survey taker organizes a list by your specifications
- Rating Scale:
- 1.very satisfied
- 2. satisfied
- 3. somewhat satisfied
- Semantic Differential Scale: similar to the rating scale, but with no stated neutral/middle. Seven point scale with opposite ends.
- Stapel Scale: rating some brand, product or service on a specific characteristic with a +5 -5 scale- the answer options being how well the characteristic describes.
- Constant Sum: accounting points to certain elements based on their importance. Points should add up to 100.
- Demographic question: multiple choices of groups like ages 18-24, 25-40 etc.
- Open ended question: any questions that requires a written answer of qualitative data.
see more information on questions and scales here:http://www.questionpro.com/a/showArticle.do?articleID=survey-questions
3. Check the validity and reliability of the survey.
Checking the validity means to check the truthfulness or accuracy of a measurement AND is a matter of opinion. Having an outsider- someone who didn’t help in the making of the survey- take the survey and ask questions. This way you can tell which problems might be hard for others to answer and why.
- Checking the reliability means to check consistency. You can do this by giving the same survey twice to the same group over time and compare results.
Overall, make sure the questions provide the answers needed (not desired- swaying your survey takers to answer a certain way is considered unethical in many businesses). Also make sure to check the flow of the survey to make sure the reader isn’t getting confused. Testing on a few people and acknowledging their concerns of questions will make a big difference before you send your survey out to a large group of people.