Researchers really love scales. Five point scales, 7 point scales, even 10 or 11 point scales. Most skilled researchers know that scales longer than 7 points ought to be avoided as they can fatigue people, and those types of scales imply that people are more precise about their opinions than they really can be.
But what if we took a different approach to scales. What if we turned every scale into a binary scale: Agree/Disagree. Important/Not Important. Yes/No. Like/Dislike.
Your immediate inclination might be that there is simply not enough variation in binary answers to make the results useful. But I disagree. Let’s look at some examples.
First, we’ll start with a single question: “Do you believe that voting in national elections is important?” The two answers would be “Yes” and “No” instead of setting up the question for answers like “Strongly Agree,” Somewhat Agree,” “Neutral,” “Somewhat Disagree,” and “Strongly Disagree.” In such a case, you might bemoan the inability to say Bottom Box is 10% or Top 2 Box is 75%. But using box scores with a five point scale is no different than saying 70% of people agree with the statement. In fact, it’s much easier for you and your research users to understand what 70% Agreement means. Further, you can just as easily understand demographic differences such as 80% of older people agree and 60% of younger people agree.
What if we used this same binary option with a grid question, thus, 6 items each with a Yes/No answer option. Just as with a 5 point (or 7 point) scale, we would have excellent data to differentiate among groups of people in the end. Perhaps 83% of my answers are Agree, 67% of your answers are Agree, and 100% of someone else’s answers are Agree. On top of that, we can say that the average person agrees with 83% of the items. And, we can still describe differences between demographic groups of people with these same types of numbers. We’ve still got a nice selection of ways to statistically describe the data.
Sometimes, we work diligently to create detailed and complicated data. It seems to me that, for many of us, the level of complexity is more than necessary. We need to think about two main things:
- What is the easiest question type for people to answer?
- What is the minimum complexity of data that you must have?
The question type might be simpler than you think. It’s certainly worth a try.
Have a survey, but no sample? Have no fear! If you’ve written a survey but have no one to send it to, we can help! Samplify is an easy to use tool for gaining access to the targeted people you’d like to survey. Request your demo today.