Three disengagement figures, pointing the wrong way

Provenance: Attributed, not found. The figure is credited to an organisation whose published work, as far as we can find, does not contain it; the nearest statement there says something different. This describes how the figure is sourced, not whether it is true, and it is not a claim that anyone acted in bad faith.

One sentence carries this trio almost everywhere it appears, and it is worth reading closely before anything else:

In studies by the Queens School of Business and by the Gallup Organization, disengaged workers had 37% higher absenteeism, 49% more accidents, and 60% more errors and defects.

That is from Emma Seppala and Kim Cameron, Proof That Positive Work Cultures Are More Productive, published in Harvard Business Review in December 2015. It is the earliest appearance we could locate.

Two organisations are credited. Neither is given a study title, a year or an author. The five-page HBR reprint we read carries no reference list, no footnotes and no endnotes; we could not check the live web version, which sits behind a paywall and may carry inline links the reprint drops. Either way, the sentence as it circulates travels without a citation, so the only way to check the numbers is to go to each organisation and look.

What Gallup published

Gallup's Q12 meta-analysis is the obvious candidate, and the edition current when the article was written is the 2012 one. It lists nine performance outcomes. Three of them are the relevant ones:

Compared with bottom-quartile units, top-quartile units have: 37% lower absenteeism ... 48% fewer safety incidents ... 41% fewer quality incidents (defects)

Set that beside the claim and three separate things have happened.

One: the direction is reversed

Gallup reports that the most engaged units have 37% lower absenteeism than the least engaged. The claim says disengaged workers have 37% higher absenteeism. Those are not the same statement, and the difference is not cosmetic, because a percentage change does not invert symmetrically.

If the engaged group sits 37% below the disengaged group, then the disengaged group sits 59% above the engaged group, not 37% above. The same arithmetic applies to the other two: 48% fewer safety incidents inverts to roughly 92% more, and 41% fewer defects inverts to roughly 69% more.

Flipping the sign on a percentage and leaving the number alone understates the first figure and has no defensible relationship to the other two.

Two: the unit of analysis changes

Gallup is comparing business or work units in the top quartile of engagement against those in the bottom quartile. The report is explicit about it: "Business or work units that score in the top half of their organization in employee engagement have nearly double the odds of success".

The claim converts that into a property of individual disengaged workers. A quartile comparison between teams does not license a statement about a person, for the same reason our trace of Gallup's 70% manager figure ends up in the same place. Between-unit findings and individual-level findings are different claims, and the word that separates them is the one that goes missing.

Three: two of the three numbers are not Gallup's

49% is close to 48%, and "accidents" is close to "safety incidents" — close enough that this one looks like a slip, and the current edition in fact writes the outcome as "safety incidents (accidents)". 60% is not close to 41%, and we could not find it in either edition we checked.

The current 11th edition, drawn from 183,806 work units and about 3.35 million employees, reports the three outcomes as 78% for absenteeism, 63% for safety incidents and 32% for quality defects. That is worth noticing for its own sake: the percentages move a long way between editions as the sample grows. A figure quoted with no edition attached cannot be checked against anything, because there is no single number to check it against.

What Queen's published

The Queen's attribution points, as far as we can tell, at work by the Queen's Centre for Business Venturing with Aon Hewitt, surveying engagement at Canadian firms of 50 to 399 employees. Queen's own magazine described it in Winter 2014, reporting more than 111,000 completed employee surveys over ten years.

Here are the figures Queen's claims for engaged workforces:

65% greater share-price increase, 26% less employee turnover, 100% more unsolicited employment applications, 20% less absenteeism

Queen's absenteeism figure is 20%, not 37%. Queen's publishes no accidents figure and no errors or defects figure at all. None of the three numbers in the claim appears in Queen's own account of its own research.

The later half of the same HBR paragraph does line up with Queen's, and more faithfully: the 100% more job applications matches Queen's figure with the direction intact. That is consistent with both bodies of work informing the passage, though a matching number is not proof of what the authors had in front of them, and we make no claim about that. What can be said is narrower and enough: the trio does not reconstruct from either source.

How it travels

The sentence propagates intact. EHS Today repeated it in August 2017 as "disengaged workers have 37 percent higher absenteeism, 49 percent more accidents and 60 percent more errors and defects", crediting the same two organisations and naming no study. Each retelling inherits the attribution and none of them inherits a citation, because there was never one to inherit.

This is the pattern our trillion-dollar turnover trace found too. A figure stated once without a reference, in a venue authoritative enough that nobody feels the need to check, becomes permanent.

To be fair to the authors

An HBR column is not a journal article and carries no obligation to footnote. The underlying argument, that engagement tracks with absence, safety and quality, is well supported and Gallup's own data supports it. Nothing here suggests the relationship is imaginary or that anyone acted in bad faith.

What does not survive the check is this particular set of numbers, in this direction, attached to individual workers, credited to these two organisations. If you want a citable version, cite Gallup directly, keep the sign, and keep the work unit: top-quartile units show 37% lower absenteeism, 48% fewer safety incidents and 41% fewer quality defects than bottom-quartile units. That claim is real, checkable, and still makes the point.