Effective Use of Employee Survey Research

Most organizations regularly survey employees to better assess and understand the state of the employee/employer relationship. Usually this is done with the specific aim of identifying strengths and weaknesses in the relationship in an effort to improve levels of employee engagement and loyalty.

Conducting employee survey research isn’t inexpensive. Given the costs involved from consulting fees to employee time and effort, organizations need to get value from the information generated. In short, the information needs to get used. Too often, it isn’t, at least not effectively.

Three Reasons Why Survey Research Isn’t Used Effectively

Three barriers stand in the way of effective use of employee survey research. More than any other, it is these three factors that rob organizations of the value of their employee survey research program.

1. Using statistical significance to identify areas of practical importance.

The  use of statistical significance as a measure of practical importance has had a profoundly negative impact on organizations as we have documented previously (see The Cult of Statistical Significance). Organizations misinterpreting statistical significance simply end up addressing the wrong things — fixing items that don’t need fixing and ignoring areas that do.

Beyond the cost of wasteful effort, however, the misinterpretation of statistical significance also prevents employee survey research information from being used. Given that the reliance on statistical significance has likely led to faulty conclusions anyway, this is usually a good thing. Nevertheless, it is instructive to see precisely how a reliance on statistical significance works to inhibit the use of information and change.

People don’t use information they don’t trust. Employee survey research that uses statistical significance testing in one form or another (see ), inevitably produces incorrect conclusions. When these conclusions are then feedback to the organization, they are seen as being inconsistent with the personal experience of employees and managers. That breeds a lack of confidence or trust in the research (rightfully so).

Of course, people don’t know why the results seem wrong, especially when they are assured by Human Resources that the survey was scientific because the results are statistically significant (just the opposite is true!). This in turn can breed cynicism concerning the capability or motives of the Human Resources function and the people within it.

Nevertheless, employees and managers won’t use the information. They may create action plans, make commitments or sign accountability agreements to initiate some changes highlighted in the research as being necessary because that is what the organization either expects or requires them to do. They won’t implement, however, because they don’t really believe it.

2. Failing to provide actionable information

Actionable information is the right information, provided at the right place, at the right time. Providing the information at the right time is not usually the biggest issue in employee research (although we recommend no more than 2 weeks between wind up the survey and the feedback of at least some preliminary results). Failing to get the right information to the right place, however, is a major factor contributing to a failure to use survey data effectively.

An example is providing global corporate level data to managers of small departments and expecting them to do something with it. What can a manager with a small group of ten people be expected to do with a finding that states that corporately, employee morale is low? Such a finding doesn’t mean that employee morale is low in that managers specific department. In fact, addressing low morale in a team where the morale is high, will likely destroy morale.

High level, global results can be useful to the senior management team and the Human Resources function in the identification of global, corporate wide issues. Such analysis may drive change to corporate policies for example. But they should not be deployed down to lower levels of the organization on the assumption that if there is a problem corporately, that all departments, functions and teams share equally in the issue or can do anything about it.

If you want managers to take action, you must provide managers with the information that is relevant to them. This requires breaking down the data to a department or team level (so long as the number of people in the team isn’t so small as to violate confidentiality). The manager in Field Services should get Field Services data, Accounts Payable should get Accounts Payable data as well as some meaningful comparative data.

3. Failing to deploy responsibility.

Whose job is it to interpret the research, make conclusions and decide what should be done? Most organizations assume it is the responsibility of the survey research consultant and whoever contracted for the research to be done, usually human resources.

It typically unfolds something like this. The consultant takes responsibility for technical aspects of the research, implementing the survey, analyzing the data, (usually employing calculations of statistical significance) and drawing conclusions. A draft report is prepared which is then reviewed by Human Resources who in turn, develop some corporate action priorities (with or without the involvement of others in the organization). These priorities then form part of the feedback report and presentation to the organization at large. Managers will be expected to take the initiatives as presented and figure out how to support them at a departmental or team level. Wrong, wrong, wrong. Wrong, because this approach fails to effectively deploy responsibility for the research.

To effectively deploy responsibility, managers must be given the information relevant to them (point 2 above) and then taught how to analyze the data for themselves. Analysis should never be left to the sole discretion of the consultant or those conducting the research. The consultant should show managers how to analyze their own data (it isn’t that hard) and then give them the opportunity to do so. Managers can then consider their data, bring their own experience, expertise and knowledge of their operation to bear on the survey results and develop improvement initiatives that are relevant to them.

In this way, responsibility for analyzing the results is deployed to those who will be making change happen. That builds confidence, capability and credibility to the entire employee survey research process.

Summary: Some Lessons for Organizations

So what should organizations do? What steps should you take to see your employee research information actually get used and provide value to the organization? Here are three lessons corresponding to each of the three barriers discussed above:

1. Never confuse statistical significance with practical importance. If you do, you will reach the wrong conclusions and make bad recommendations. Worse, you will build cynicism and distrust in the organization and certainly with HR.

2. Breakdown analysis to a level where the information is relevant to managers. Different people with different roles in the organization all have different information needs. The information delivered must be right for them.

3. Give managers the responsibility for their data. Breakdown analysis (described in point 2) should never be presented as a fait d’accompli with conclusions already developed for the manager. Managers must be given their data, taught how to interpret that data, and then given the responsibility to draw their own conclusions.

If you enjoyed this post, please consider leaving a comment or subscribing to the RSS feed to have future articles delivered to your feed reader.

Filed Under: Survey ResearchVoice of the Employee

About the Author:

RSSComments (0)

Trackback URL

Leave a Reply

    Why products get better and service gets worse

    Posted on: 2012-07-25

    Service standards are really a means of avoiding accountability CALGARY, AB, Jul 17, 2012/Troy Media/ – Remember when getting a 100,000 miles of your car was considered a miracle of sorts? Those invited to bear witness performed the ritual of peering into the driver side window to confirm all those zeros on the odometer. The

    Read More…

- More From Our Blog-