Creating an Effective Survey

There are five steps necessary to crafting a survey that is valuable to enhancing organizational improvement.

This article accompanies Beyond Engagement.

Thursday, March 1, 2012
Write To The Editor Reprints

"The devil is in the details." When it comes to employee surveys, this popular saying is indeed correct. Understanding the details behind crafting relevant and actionable survey items is essential when compiling the most informative data to foster improvement at your organization.

Successful surveys hinge on the following five facets:

* Item Design;

* Item Scale;

* Item Selection;

* Survey Order; and

* Benchmarking.

Item Design

The wording of items within your survey has a strong impact on the results. Compose survey items in a constructive fashion and make sure the tone of the surveys is always positive. In this way, you will be more likely to generate opportunities for improvement, rather than creating problems or unrealistic expectations.

You should also present items as statements rather than questions. Doing so significantly lowers the reading level of the surveys and enables more respondents to successfully participate, increasing the statistical significance.

A good rule of thumb is to write questions appropriate for a grade 6.5 reading level, since this level is assumed to be applicable to a broad range of multicultural populations.

Avoid double-barreled items which address two subjects within one statement, since responses to these items only allow for a single answer.

For example, when responding to the statement "My manager is friendly and encourages my growth," participants may agree with only part of this statement, but they are forced to either agree/disagree with the entire statement.

Resulting data may not be indicative of participants' true feelings on both subjects, and you will be unable to fully address the challenge without asking another question to separate the two facets.

Item Scale

The response scale also factors into the data you receive. Studies, including an Infosurv white paper, show 71 percent of respondents prefer a five-point -- rather than a six-point -- Likert Scale, including a neutral response, when assessing opinions.

Presenting a scale without a neutral midpoint can introduce respondent bias, since participants are forced to choose a more positive or negative response, even though they may feel truly neutral. In many cases, respondents will accentuate the negative in an experience, which ultimately skews the data.

A five-point scale also provides a nice midpoint and a standard point of comparison when calculating the mean weighted average.

Item Selection

As a rule, you should strategically pick items which reflect the current environment within your organization. In doing so, the survey will reveal how employees feel at the time of survey administration, especially as it relates to events within the organization.

For example, incorporating items addressing any current changes will allow you to understand how employees are adjusting and create effective action plans to address any challenges.

In designing an efficient instrument, it is fundamental that scores on similar items be internally consistent (related), but also contribute some unique information. For example, consider these two statements:

* My organization is a good place to work.

* I would proudly recommend this organization to my friends and family as a good place to work.

The second item expands on the first and provides information about the employee's likelihood to recommend his/her organization. Statements that measure similar information can then be grouped into dimensions, and the construct validity of each item can be calculated to ensure the item measures what it is intended to measure.

Newsletter Sign-Up:

HR Technology
Talent Management
HR Leadership
Inside HR Tech
Special Offers

Email Address

Privacy Policy

Survey Order

Good survey design focuses the employee's attention on the items themselves, rather than a section header. Perhaps counter-intuitively, effective surveys are those that have very little organization when it comes to item placement.

If given categories, survey respondents tend to answer to the headings or quickly answer similar items without much cognitive thought. Randomly placing survey items throughout the survey encourages participants to put more effort into thinking about each survey item, resulting in more useful and accurate information managers can use as part of their action-planning process.


In order to fully understand survey data, you must have a frame of reference. Using items which are comparable to items contained within a normative database is ideal.

Normative databases and norms allow organizations to benchmark against other organizations in the same industry and evaluate and compare survey results within the organization. In using benchmarking data, you will have a more complete picture of your organization internally, as well as how it compares to others.

A well-designed survey can truly have an impact, if implemented effectively. Pay attention to survey details from the start, and data interpretation will be much easier down the road.

One final piece of advice: An employee survey is worth next to nothing if the results are not used for change. The most important part of survey design comes after the survey is administered, when the results are turned into actionable improvement plans.

After all, if you take the time to create and administer the survey, why wouldn't you want to use what you have learned?

Kevin Sheridan is senior vice president of HR optimization at Avatar HR Solutions, Inc., and New York Times best-selling author of Building a Magnetic Culture.

Copyright 2017© LRP Publications