Employee satisfaction and engagement surveys used to be about receiving employee feedback on what's working and what isn't in the employee/employer relationship from the employee’s perspective. Information that could be used to build better, more productive workplaces.

However, the employee survey has evolved into a process of misusing statistical technique in making exaggerated claims – literally creating statistical fairy tales about the level of employee engagement in the business and the impact this engagement has on performance. These surveys then, have moved from getting information useful for improvement to calculating meaningless statistical engagement indexes.

Garbage in, garbage out. Employee engagement surveys typically use statistical significance tests to identify important findings – such as whether engagement is improving or getting worse, or whether departments have particularly high, or low levels of engagement.

But statistical significance isn't practical importance. When used in this way, statistical significance gets over 90 per cent of the answers wrong.

The result, management takes action on things that don't matter to employees while ignoring the things that do. It doesn't take long before employees conclude that management, or the HR department, just doesn't get it. This starts the decline in engagement.

Even worse, however, is when effective, efficient and highly engaged departments get labeled as having 'low engagement' determined by a statistical significance test. People don't know what's going on or what they should do. The numbers seem wrong, but people believe they are right, because statistical significance testing makes it look so scientific.

You are right. 'Stats' was the class most people simply muddled through in college. This is what has made it so easy for large HR consultancies to sell this statistical junk-science.

For example, engagement surveys today claim to be able to predict productivity, profitability, sales, absenteeism and even accident rates. Being able to relate employee engagement with business performance in this way may be described as the “Holy Grail” for HR. But few managers or HR professionals seem to take the time to think about how this could even be remotely possible. It isn't.

The best defense for managers and business owners is a little skepticism combined with that old bit of common folk wisdom: 'If it seems too good to be true, it is probably is.'

Absolutely. Edward Tufte, author of Beautiful Evidence, said that "a presentation is a moral act as well as an intellectual activity." When statistical technique is misused in the way it is, organizational performance suffers, real people are hurt. But the ethics of this go both ways. It is easy to point at those selling this junk science and say it shouldn't be done. But too many organizations are buying it and casting a blind eye to the ethical downside.

Perhaps an example will illustrate just how bad things have become. The statistical methods used to analyze data in engagement surveys is precisely the same as the analytical methods used in the 1994 best seller 'The Bell Curve' to 'prove' that blacks were intellectually inferior to whites. I am not saying sort of the same, or kind of like, but exactly the same.

I'd like to think that most organizations would be offended in knowing that this is the kind of thing they are inflicting on their people. They owe it to their people to do a better job at soliciting and analyzing employee feedback.

Get back to the basics. Stop trying to calculate statistically significant engagement indexes highly correlated to critical business performance metrics. Start asking your people what matters to them, what would help make this a better, more productive, and enjoyable workplace.

That means custom designing an employee survey. Steal questions you think are good from other surveys and develop a few of your own. Base these on informal conversations you have had with your people about what is working and what isn't.

Employee engagement and feedback surveys are great tools for getting feedback. Every business should use them. But what comes back must be given the respect it deserves. That means business leaders must listen to what is being said.

I know that not every business can afford a high priced consultant. To help, Converge has published a do it yourself guide on our Voice of the EmployeeTM website. This is a good guide even for those that can afford a consultant as it outlines what to do to conduct an employee survey properly and what to watch out for.

As you can probably tell, I'm rather numbers oriented. My focus has been on improving operational performance. But my business partner in Converge, Shelley McLean, is a HR expert, specializing in change leadership.

She introduced me to employee engagement surveys and asked me what I thought from the scientific/statistical perspective. I was floored. I couldn't believe people were doing this. She challenged me to do it better.

It wasn't hard. Just adopt well known and valid statistical analysis techniques to the problem of gathering and gauging employee feedback. Real science doesn't use proprietary methods, magical black boxes or secret decoder rings. The techniques of sound data analysis are published and available.