Figures on employee engagement, and who produces them

The most cited engagement percentages come from surveys conducted by consultancies, which also sell services intended to improve the indicator they measure. The figure then circulates in the press, detached from its producer and its method.

That configuration does not invalidate the results, but it calls for the usual caution about conflicts of interest, and it explains part of the remarkable stability of the published figures from year to year. An indicator that suddenly worsened or improved would raise questions about the instrument itself.

Engagement questionnaires combine items on satisfaction, on attachment to the organisation and on willingness to make an extra effort. Some add questions on the concrete conditions of work, such as having the necessary equipment or receiving regular feedback.

Those three dimensions are distinct, and aggregating them into a single score produces a measure whose meaning depends entirely on the weighting used, which is rarely published. Two people can obtain the same score for opposite reasons, one highly satisfied and weakly attached, the other the reverse.

The notion nonetheless has academic origins: William Kahn proposed it in 1990 to describe the degree to which a person invests themselves physically, cognitively and emotionally in their role. Validated scales derive from it, such as the one developed by Wilmar Schaufeli and colleagues in the early 2000s.

Researchers in the field have long noted that the notion covers constructs already measured by validated scales, notably satisfaction and organisational commitment. The debate is about its added value: does it measure something earlier instruments did not, or does it rename an existing set for commercial purposes.

Syntheses find associations between engagement scores and unit performance indicators, of moderate size. Some of those syntheses are in fact produced by the same organisations as the surveys, which does not invalidate their calculations but justifies distinguishing them from independent work.

The direction of those associations is uncertain: a unit doing well produces favourable answers, as much as high engagement might produce results. The few longitudinal designs find effects in both directions, which rules out presenting the score as a lever whose rise would mechanically drive performance.

European working conditions surveys and national surveys on industrial relations measure neighbouring dimensions with documented methods and probability samples. Their questionnaires are published, which makes it possible to know exactly what each figure covers.

They are less cited because they produce less spectacular results and because they are not promoted by actors with an interest in their uptake.