You’ve been tasked with designing your organization’s upcoming employee engagement survey. You vaguely recall the basics of survey research from a social research class, but you also acknowledge that you may have slept through most of that class (as was the case with many of your classes). Never fear! Here is a set of guidelines on how to create an employee engagement survey.
Create an Employee Engagement Survey
Planning and Organizing
1. Set Goals and Objectives for Doing a Survey
First, make sure you have thought about and captured on paper what you want to learn from the survey. You should consider whether you want to focus on any themes or subtopics. You also need to understand whether you are going to follow a specific methodology or approach. For example, at DecisionWise our employee engagement survey is designed to measure what we call Engagement MAGIC®.
MAGIC is an acronym that stands for five key variables that have been shown to be prevalent within an engaged workforce: Meaning, Autonomy, Growth, Impact, and Connection. Other consulting firms have their own approach, too. So, make sure your survey has the right questions if you are going to make use of a pre-defined methodology.
Tip: Don’t forget that you can use a simple pre-survey with your management team to identify the larger themes that should be addressed in the company-wide survey.
2. Identify Your Audience and Be Specific
Don’t take the easy way out and simply write down, “Our company’s workforce.” Think about specific groups within your organization, and ask yourself whether you need to include customized survey questions for these various groups?
While your freedom in this area is somewhat limited due to the number of questions that you might reasonably ask, you might be able to personalize one or two questions to help you better understand a demographic or viewpoint.
3. Consider the Medium Used to Take the Survey
Do you have manufacturing plants where the preferred method is to pass out paper surveys during a company meeting? Or, are you a restaurant operation that employs a lot of shift workers and a mobile solution would be the best method for delivering the questions? Do you need to use kiosks and provide kiosks codes?
These issues need to be considered upfront, because they cannot be addressed later down the road.
4. Consider the Scale the Participants Will Use
Is a 5-point rating scale the right choice, or do you need more granularity and a 7-point scale would be better? Don’t necessarily assume that the larger the scale the better the answers will be. If the scale is too broad, you can’t find consensus, and you may end up creating a survey that is difficult to read and apply.
Strongly Disagree Disagree Neutral Agree Strongly Agree
5. Ensure Confidentiality Throughout the Survey Process
How will you ensure confidentiality is maintained throughout the survey process? The science is clear on this matter. Unless employees realistically believe their responses will be kept confidential, they will not provide you with their true feelings for fear of reprisal. For this reason alone— the need for confidentiality— many organizations outsource the employee survey process to an outside firm to ensure that confidentiality is promoted and maintained throughout the process.
6. Plan and Conduct a Survey Launch Campaign
A launch campaign is a corporate communications initiative that will help employees understand why they are being given the survey, what you hope to accomplish with the data, and when they can expect to learn about the results and how will those results be conveyed to them. The launch campaign will also provide you with a forum to help your participants understand your expectations on how to apply the survey’s rating scale. A well-conceived launch campaign has been shown to improve dramatically your survey participation rate as well as help ensure that your workforce takes the survey seriously and provides meaningful responses
Drafting the Survey
The following 16 tips are useful in drafting the actual language to be used in creating your employee engagement survey.
1. Use Language an 8th Grader Could Easily Understand
Keep the language of the survey simple and easy to understand for a wide variety of ages and education levels. This will help ensure accuracy in the responses of those who are being surveyed.
2. Avoid Using Confusing Terms and Jargon
Avoid acronyms, corporate jargon, and industry terms that may not be known by your audience. If you must use jargon, take the time to explain what you are talking about.
3. Understand How to Ask Questions
Keep biases out of your questions by avoiding leading, loaded, or double-barreled questions. Leading questions are ones that sound like this, “How likely would you be to recommend our excellent services to a friend or colleague?” A loaded question is one where you say, “Approximately how many hours did you waste in the breakroom this year?” Double-barreled questions are ones that ask about more than one item, such as, “How often did you attend your staff meetings and turn in your timesheets?”
4. Make the Questions Personal
Make the questions personal to the survey taker. For example, it’s better to say, “In your opinion, which of the following employee benefits are the most important?” rather than simply asking them to stack rank employee benefits. A survey taker might not realize you want their opinion and instead give you what they think is the “right” answer.
5. Use Time Frames in Your Questions
You should always provide a time frame. For example, instead of asking, “How often did you meet with your supervisor on a 1-to-1 basis?” ask instead, “In the past 6 months, how often did you meet with your supervisor on a 1-to-1 basis?” Providing a time frame keeps participants from giving you data about time periods that might be irrelevant to your analysis.
6. Don’t Ask Unbalanced Questions
Don’t ask unbalanced questions such as, “Do you like our online benefits system, or not?” Instead ask, “How useful has the online benefits system been to you?” If you want to avoid the “don’t know” or “N/A” answers that might be prevalent with this type of question, you can use skip logic. A survey with skip logic first asks, “Have you ever used our online benefits system?” before asking the participant to provide his or her feedback. Using skip logic is one way to avoid having to sift through a large set of responses that don’t mean anything.
7. Avoid Broad Questions
Avoid asking questions that are too broad. For example, if you were to ask, “How do you feel about our new breakroom?” you might discover that employees are satisfied – 3.7 on a 5-point scale. But maybe all you really want to know is how they feel about the size of the breakroom since you are going build three more breakrooms in other locations. Then, the better question is “How is the size of our new breakroom? (too small, just right, too big, etc.).”
8. Use Open-Ended Questions
Consider how you might use open-ended questions. Sometimes you might need feedback that is unvarnished and has not been framed by the underlying question. Open-ended questions are a good way to get unfiltered feedback. And, with strong advancements in technology, many survey platforms give you the ability to machine-read and then categorize your open-ended comments.
9. Try Not to Change Rating Scale
Be careful to not change your rating scale during the survey or else you run the risk of confusing the participant and creating responses that corrupt your data. If you do find a strong need to change the scale, then make sure that all questions using this scale are grouped together. Also, you might consider using words instead of numbers. Instead of using 1 – 5, you could use “Very Disappointed, Disappointed, Neutral, Satisfied, Very Satisfied.”
10. Keep Ranges Consistent
Keep your ranges consistent throughout the survey and within the question itself. For example, the following ranges would be quite problematic:
- 1-5%
- 10-30%
- 50% and above
11. Avoid “All-That-Apply” Questions
Avoid questions that ask the participant to select all that apply. The trouble with an “all that apply” question is that you can’t ascertain importance between the various selections provided by the participant. You might get a nice grouping, but not much more.
12. Avoid “Yes/No” Questions
“Yes/No” questions are wonderful in the courtroom, but they typically do not give you good survey data. This is because you do not get context and gradation – all you get is a binary choice.
13. Consider Using “Other”
Sometimes it makes sense to provide an “other” category when asking participants to select from a list. If you don’t provide the “other” option, then you run the risk of having them skip the question or giving you an answer that “fits” the “a, b, c, d” choices that you provided.
14. Consider the Order of Question Placement
Question placement matters as you draft your survey. Don’t ask hard-hitting questions until you have a chance to develop some rapport with the participant. Also, if you are going to ask about a sensitive topic, please know that any topics that follow might be tainted by the previous, sensitive, topic.
Hence, it might make sense to keep the most difficult questions for the end so that the entire survey is not influenced by potential negative emotions from the participants. One final note, make sure your questions are logically grouped and remember that surveys that are too long will suffer from survey fatigue.
15. Get the Questions Answered
Finally, it’s generally a good idea to make sure that all questions must be answered by the participant before moving on in the survey. That said, if you are going to require answers on each question, make sure the participants have options that they can answer.
16. Be Respectful
Treat your participants with respect. If you are not going to show all questions on a single page, then make sure you provide a progress bar or some other indicator that gives the participant a sense of how much they have completed and how much they have left to complete in the survey. Also, don’t fib about how long the survey will take and make sure a progress bar/indicator provides accurate time estimates and/or completion percentages.
