Question Development and Survey Implementation

Another issue to be considered when using a survey is the questions that are being asked. Before you develop any questions you must have some idea of what you want to measure or what types of responses you expect. Once you define the responses that are expected, then you can begin to develop questions that should give you the responses you are expecting in each area. Before implementing a survey, it should be tested on a sample of people to see if the questions are interpreted the way that you expected. Once the tests validate your survey then it can be used to collect data. 

In order for a survey to give meaningful data the sample of respondents must be chosen randomly. This m be difficult to do depending on how the survey is implemented. 

Let s look at a comment card for a hotel as an example of a customer satisfaction survey. What kinds of questions are asked One caid had the following types of questions on a 3-5-7 point scale  

All of this is nseful information to have abont the facility. Who fills ont comment cards Most of the time it is a client that has had a bad experience and the card is being nsed to register their complaint. This does not provide a random sample of data to analyze to try to effectively evalnate performance in the eyes of the cnstomer. 

One solution that I have seen to the random sample problem is to have five to 10 questions that all guests are asked as they are going through the checkout process. This produces a 100 percent sample and a more accurate indication of what the guests think can be evaluated. 

---