written by Kristin Cavallaro
Sometimes I miss my days in Sales. The thrill of learning a person’s need or problem and then finding the solution for them was just pure adrenaline some days. Other days it was nearly impossible to even get my foot in the door. It was way back then that I began my “researching” interests. I would test out asking the same question, just in different ways. For example, when someone would enter the store, I would not ask if they needed help with anything. It was too easy for them to say no and then boom, the door of opportunity had closed. Instead I learned that I would instead ask “What can I help you find today?” It now became harder just to say “no” because “no” wasn’t a valid answer. They would actually have to tell me what they are looking for (well in most cases). It also showed that I was interested and ready to help them. There were still many that declined my help, but I increased my chances of getting my foot in the door tremendously. The same holds true for bringing something to a dinner party. When I RSVP and ask do you want me to bring anything….I usually get “no, we’re all set.” However when I ask “what should I bring,” I usually get an answer.
Outside of sales and dinner parties, it is common practice for researchers in the US to ask for a respondent’s income in their surveys. As many of us now know, it is important to include a “prefer not to answer” with questions like these. However, often in these cases, we terminate any respondents that don’t want to share their income. What if we just aren’t asking them in the “right” way? In a recent research-on-research study, SSI decided to “play around” with the income question. For this particular study, it was imperative that we had income data on our respondents. Those that did not want to provide their income data were not able to be used in the analysis even though we allowed them to continue through the end of the survey.
We began with asking respondents to enter in their estimated annual household income into an open-ended text box. We also allowed them an “out” by providing a “prefer not to answer” check box. While we had a great participation rate in this question, we did end up with approximately 23% of the respondents opting to plead the fifth on their income. So to combat this, we decided to ask the question again but in a different format. We asked respondents if they would be more comfortable reporting their income information in a predetermined range. Astoundingly 65% of those who previously answered “prefer not to answer” provided us with an income bracket. Because of this two-step process, we only needed to disregard 8% of the responses as opposed to the original 23% that we were faced with. When compared to other previously-fielded projects, we achieved a lower percentage of respondents refusing to give their income.
I’m not saying that it is okay to harass our respondents over and over again with the same questions if they don’t feel comfortable answering. However in the event that you could sacrifice some of the detail in responses, for example exact income versus an income range, you can gather more useable data form one person who is ready and willing to share their honest opinions on the rest of your project. And think about it…when evaluating income brackets, we tend to condense the asked brackets into broader groups for analysis.