Write-in-the-Answer Survey Questions: Powerful But Problematic.
By John Towler Ph.D. | April 30, 2003
Most surveys contain a mixture of questions: true-false, ranking, agree-disagree etc. But the best ones allow respondents to say exactly what they think. Questions like In my opinion management is --- , or The best thing about working here is --- , generate the most honest, useful information.
- Volume of responses. The sheer number of verbatims received may be overwhelming. One of our surveys contained three verbatims and went out to 7000 people. We received thousands of pages of written responses. The information was fantastic; simply handling and sorting the answers was a major task. Be careful what you ask for and know how you will handle the responses you will get.
- Analysis. Now that you have the responses, how will you analyse them? If there are only a few (less than 100), this isn t a major problem since you can simply read them and make notes about what has been said. But if you receive hundreds or thousands, you will have to come up with a different, cost effective process. One approach is to categorize the answers. Do this by reading a large number of randomly selected responses and make a list of the general categories you find. The answers may pertain to management, training, communication, safety etc. Now read all of the responses, keeping a record of how many times these categories appear. You will have to include an other category for the ones you didn t notice before. This process can be extremely time consuming, but it is the only way to quantify what people have said.
- Anonymity. People will not say what they think if they fear that they can be identified. You must guarantee their anonymity if you want honest answers. This can be a problem if your survey has asked them to tell you what location or department they are in. In addition, some people may be concerned that others will be able to recognize their handwriting, or that their responses will be give to somebody who can identify it. One solution is to have the verbatim questions on a separate page of the survey so that they will not be recorded with location or department identifiers. As for the handwriting, you can either transcribe the responses or restrict the number of people who will see the originals to a few very responsible people. However, the easiest solution is to use an electronic, web-based survey. This eliminates both of these concerns.
- Loss of meaning and sentiment. Categorizing or summarizing the results may be necessary, but you run the risk of changing or diluting what people have said. One solution is to select and record typical and important comments exactly as they were written. Management deserves to hear exactly what was said by someone who explains in detail why they love their job and working for the organization. On the other hand, they should also see the complete statement from someone who emotionally identifies issues or concerns that are driving them mad.
- Reporting the results. What is the best way to report the findings? This is very simple when dealing with yes-no, ranking or agree-disagree questions that generate numbers or graphs. But it is another story when it comes to verbatims. If you have categorized the answers, you can use some numbers and if you have examples of typical responses, that too will enhance your report. But what do you do with all of the rest of the data? Most managers want to have the opportunity to read everything that has been said. However, this can be a problem if the respondents have disclosed personal, secret or potentially damaging situation. We have seen verbatims in which fellow employees have been accused of theft, favoritism, and giving sexual favors. Often names are used. Deciding who should see this information requires careful thought and planning before the survey is started. One solution is to prepare an executive summary of the findings for general distribution and reserve the complete, unabridged report for senior management only.
Verbatim questions can be the most powerful, useful and provocative part of your survey. But plan ahead so that you will know how you are going to deal with and analyse the results. Be careful what you ask for. You may get more than can handle.