How to Conduct a Survey – Creating the Survey

How to Conduct a Survey - Creating the Survey

Step One – Creating the Survey
Should you do this by yourself or hire a professional firm to help? Most should seriously consider working with a survey research firm for something so critical. A relatively small investment to make sure that your research is successful may pay for itself many times over. Only you know the exact objectives and the exact requirements, but some help with the technicalities may be worthwhile.

Objectives
Before you create your survey, you should first determine what your objectives are for the research. The objective may be to increase profits, improve customer service, reduce employee turnover, determine how your customers perceive your latest product, or any number of other objectives.

Do not skip this step! You do not want to get to the end of the process and not have the data you need to achieve your specific goal or goals.

Question Development
There are two different types of survey questions: scaled and open-ended. Lets talk about scaled questions first, technically called quantitative questions. These are the ones where you are asked to agree or disagree with a statement (survey question).

Your objectives will determine the scaled survey questions that you want to ask. This is the area where many research projects fail, but the designers do not know that the project has failed until later in the process, or even years later. The scaled survey questions are the heart of the survey research, and they must be written in a way that does not bias the survey responses while still collecting the data that is needed to achieve the stated goals. Please see Survey Questions for additional information.

One consideration is to use standardized survey questions, rather than writing your own survey questions from scratch. This is exactly what all good survey research firms will recommend. Standardized survey questions have already been tested and validated. This way you will be assured that the data you collect is clean and free from biases. You may be saying right now that you are unique and that no “standard” questions could be used on your survey. This is not true. Because so many have thought that, you will find that there are standardized questions to address virtually anything you might want to measure. And, you will find that the standardized survey questions used to measure customer satisfaction or employee satisfaction, or any number of other topics can be used very successfully across all organizations.

One huge advantage of using standardized survey questions is that this makes it possible for you to benchmark or compare your survey responses to survey responses that have already been collected from your specific industry. For example, this allows you to know if your customer service or employee satisfaction in the hospitality industry or any other industry is poor or excellent or somewhere in between. This is invaluable information when it’s time to design your action plans to use the survey data to achieve your stated goals. Benchmarking data can usually only be obtained from a professional survey research firm. NBRI’s benchmarking database is known as ClearPath Benchmarking.

Now lets talk about open-ended or qualitative questions. The ones where you are asked to write a response. Something like: Please tell us how we can improve our service. Professional survey research firms will recommend that you limit the number of open-ended questions on your survey. This is because the only reason you should place an open-ended question on your survey is to gather information on other topics that should be addressed on future surveys. It is also best to limit the number of open-ended questions because they can provide misleading information. We, as human beings, are a curious bunch. It seems that we love to read the answers to open-ended questions. We are also likely to place a high value on those answers, a value far beyond their “real” value, akin to the old adage of “The squeaky wheel gets the grease.” Just because one person says that you shipped their product late does not mean that you always ship products late. The only way to know if you always, as perceived by your survey respondents, ship your products late is by analyzing the responses to scaled quantitative questions, not open-ended qualitative questions.

You may also want to ask demographic questions so that you can divide your collected data into various segments. You may want to ask gender, age, and other basic information so that you can determine that 24-35 year old women are most pleased with your customer service, or 45-60 year old males are your most unhappy employees, etc.

Design the Survey Instrument
First, we have not discussed the survey response scale for the scaled questions. There are many survey response scales to choose from, but it is recommended that you be careful here as the wrong choice could bias the survey responses either to the positive or negative or actually gather no useful information.

All trained researchers have, for the last 30 or so years, recommended what is called a balanced scale; a scale with an equal number of “positive” and “negative” response choices. This is to make sure that the scale itself does not inadvertently bias the survey responses to either the negative or the positive. The most popular example has six different choices as follows: Strongly Disagree, Moderately Disagree, Slightly Disagree, Slightly Agree, Moderately Agree, and Strongly Agree.

It is also important that the scale not have a neutral point. Using a neutral point such as “Neither Agree, nor Disagree” does not gather any useful information. One of the purposes of survey research is to gather opinions and a neutral point does not do that. You can read more about survey scales here: The Survey Scale.

Second, what is the best way to layout your survey instrument?

It is recommended that you begin with an opening paragraph that explains the purpose of the survey and why you are requesting participation.

Then, include an instructions paragraph that explains how to respond.

Next, any demographic questions that you wish to ask.

And finally, the scaled and optionally any open-ended survey questions.

Now you know what you want to accomplish and you know what questions you want to ask. Next, let’s talk about deploying your survey.

PDF