People, not profits, make it possible for organizations to succeed. Communication is the key to maintaining employee motivation and driving the business forward.
Employee surveys help you identify various factors that impact the experience of your employees in their work. Your business will achieve higher productivity with motivated, loyal, well-trained staff accompanied by a happy working environment.
Employee surveys are instruments incorporated by the management to collect feedback and evaluate employee engagement, experience, morale, and productivity. It is essential to assess the effectiveness of new structure, leadership, and training on Employee Satisfaction, stress levels, health, and productivity.
In an age where more and more employees worry about being watched by Big Brother, where businesses have the tools to track almost everything, running a survey can indicate that Big Brother is still human.
What is the differentiator between a good and a lousy survey in the workplace?
The distinction is a careful and thoughtful layout of the survey, very clearly. And the unfortunate truth is that many managers and HR professionals are still behind in developments with regards to survey design trends. The past decade has brought dramatic changes to this field, and it has seen a five-fold growth in the number of publications describing the importance of corporate survey results. But many managers are still applying design principles formulated ages ago.
In this article, we’ll explore some of the developing trends in survey design and provide guidelines to help companies improve their workplace surveys. These suggestions are based on peer-reviewed studies and reports on behavioral sciences, general knowledge of survey design, and the experience of our organization in developing, implementing, and revising extensive corporate surveys.
Guidelines for Content
- Questions should be specific to observable behavior rather than thoughts or perceptions
Most surveys, particularly those intended to evaluate performance or leadership skills, invite participants to define specific individuals’ characteristics or ideas. While there is an understandable interest in answers to these questions, the company is unlikely to receive direct answers by asking the questions. The answers to such questions can often be biased by the individuals associated with the person being evaluated. The best way to tackle these problems is to ask questions about concrete, observable behaviors, and encourage respondents to rely on their own experiences. This reduces the risk of distortion. - Include certain things that can be verified independently
Of course, if the survey responses are not related to verifiable facts, there is something wrong. Conversely, verified reactions enable you to draw conclusions about the validity of the survey, mainly when something new or unusual is assessed by the survey. - Test just habits that are linked to the success of your business
This principle may seem evident, but we have reviewed countless surveys and found that most surveys are designed without focusing on business outcomes or job performance. This flaw explains many of the surprising survey failures. The dilemma is most often triggered by the inability to pick questions systematically. We use a two-stage system to prevent this. Firstly, we consult with informed contributors and ask them to identify the significant issues and the reasons for them. Then we examine the published research in order to identify known problems and causes. Only then we selectively pick questions that reflect all of the identified issues.
Guidelines for Format
- Maintain unlabeled and uninterrupted portions of the survey by page breaks
Boxes, category labels, and other harmless-looking survey information can subtly and significantly skew responses. The cause is relatively straightforward: as extensive research reveals, respondents prefer to answer questions related to each other similarly.By simply removing the boxes, tags, and page breaks that disrupted some pages, we solved the problem. The formatting changes allowed respondents to interpret the question on their own merits; although the changes were minimal, they had a significant impact on the results of the survey. - Format parts containing a similar number of items and asking a similar number of words
Studies and our own experience indicate that the more questions you ask, the higher the resulting scores for the entire section tend to be. Similarly, questions that contain more words and require more time for reflection are often given higher ratings by respondents. Maintaining fairly equal lengths of questions and segments provides the highest probability of receiving compatible survey responses across all sections. - Let respondent demographics be the last question in your employee surveys and the first on in performance surveys
An optional demographics section is a regular part of customer surveys and its value is unquestionable. Demographic questions also frequently appear in employee surveys as managers believe that the information generated can yield useful detailed information about trends in the workforce. Of course, minimizing demographic questions that may seem intrusive or insignificant is crucial. The rates of response of employees can dramatically reduce if they feel that anonymity is being compromised by including demographic questions. In employee surveys, putting demographic questions at the end is generally best, making them optional, and minimizing their number.
Guidelines for Language
- Ignore words with strong associations
This is one of the most ignored principles of language. Metaphor plays a prominent role in management descriptions, but it can also trigger bias-responding associations. While these sentences are commonly used to describe leadership qualities, they are counterproductive in surveys because they can trigger gender favoring associations.Here, simple revisions in wordings can solve the problem: “Has a strong grasp of complex problems” seems more inclined towards males, while “Discusses complex issues with effectiveness and clarity.” is more gender-neutral. Similar results were observed when words were changed that triggered ethnic and religious associations. - Change the wording so that the desired answer is negative in about one-third of the questions
The respondents’ tendency to agree on questions is one of the best-documented response inconsistencies, which increases with the length of the survey. It is best to introduce questions that are articulated negatively. One good way to prepare readers for this possibility is to add a simple reversed question early on. It reflects the presence of such questions throughout the survey. In our experience, we have found that changing the wording into about one-third of the questions is a good rule. - Avoid combining two separate topics into one subject
Most questions in a survey combine two or more components. If items are related, there is a sense in merging them to reduce the duration of the survey, but it may other times be troublesome to combine two or more components. When deciding if two similar elements should be included in the same question, it should be determined if they require to be treated similarly when an issue is identified.
Guidelines for Measurement
- Create an answer scale with regularly spaced numbers and words only at each end of the range
Yet, because they were affected by a variety of external factors, the outcomes of these tests were notoriously unreliable. The major problem is that each response option on the scale contains different terms, and it is, therefore, difficult to place the answers in a spaced mathematical process so that statistical tests can be carried out. Although the labels can be plausible, the distance between each pair of continuum classifications remains unknown.These and other distortions created with word labels can be avoided by using a scale with only two-word markings, with one on each end. Questions answered with numerical scales may not appear to be very different from those answered by the word, but their answers are much more reliable and can be analyzed in more detail. - Only use a single answer scale with an unusual number of options
One response scale, like “never” to “always” with numerical ratings in between, makes a quick response comparison and is better for respondents. Single-scale surveys take less time than multiple-scale surveys, are more reliable, and provide much simpler statistical comparisons of different items. We think it is advisable to provide an unusual range of responses so that people can register a neutral opinion. We are also in favor of including a “don’t know” or “not applicable” reply. Without this option, respondents may feel obliged to provide unworthy answers. Including this choice increases response rates and reduces the likelihood of respondents leaving blanks or ending the survey in the middle - Avoid questions involving ranking
Numerous surveys require respondents to rank in order of preference a number of items. A host of factors skew the answers to such questions— most notably, the number, order, and object range. Respondents will best remember the first and last items of a list and tend to assign the top and bottom ranks to them. Furthermore, other research shows that a ranking question may influence ratings on subsequent questions, possibly as respondents are sensitized to the ranking question subject.
Guidelines for Administration
- Make sure that employee surveys are anonymous and ensure that they remain so
As we have already pointed out, if respondents are confident that personal anonymity is guaranteed, they are much more likely to participate in surveys. Many employees are smart enough to know that the fingerprint in every device is unique and that passwords can be decrypted or overridden easily. Employees are reluctant to provide honest reviews and ratings for their leaders and peers as they know that their answers are stored somewhere on the mainframe of the business. The respondents’ desire for anonymity explains why many companies prefer to use paper-based surveys, even though all employees have access to a computer network. - Make the department the primary unit of analysis for corporate surveys in large organizations
Although the need to maintain anonymity is paramount, large corporations still need to coordinate and analyze department-or operating unit-level results of internal surveys as they measure success at these levels. Clearly, the utility of surveys that are undifferentiated by departments is poor. This helps you compile customized feedback reports that divide the cluster departments into the specific groups you need. - Ensure that the survey is completed in about 20 minutes by employees
Employees are busy, and no one really likes surveys and evaluations. If a survey appears to be unreasonably time-consuming, only people with a lot of free time will be involved, and the response quality will drop significantly. We’ve already seen that respondents’ responses become automatic and overly positive when surveys are long. Generally speaking, we found that surveys that can be completed in 20 minutes can provide a company with substantial results.Employee focused strategy has become the norm, and consumers look favorably on people-centric organizations rather than profit-focused businesses. You will recognize what goes well and what needs improvement only by communicating closely with your employees.
Now you know the secret recipe to the best employee surveys.
