Employee Engagement Driver and Priority Assessment for Employees and Human Resources TeamsNot how engaged you are. Which of the things that drive it are short here, and which of those you would actually miss.
Forty short items about one job over the last three months. Six drivers of engagement, each measured two ways: how much of it the job actually supplied, counted in acts and occasions, and how much it matters to you, measured by what you would trade rather than by a rating. No right answers and nothing to know. Reported per driver as a placement and a drawn band, never a number, ranked with ties where the bands overlap, with a refusal of any total in the largest type on the page, who can move each driver, and a ninety-day ladder from the top tie group, ending in a booked re-measurement. One payment covers the sitting and the report together.
An engagement survey that answers the question engagement surveys dodge
The Employee Engagement Driver and Priority Assessment is a fourteen-minute individual survey measuring, for one job over three months, how much of six engagement drivers the job supplied and how much each matters to the person, reported as a placement and a band per driver, a ranking with ties, and a ninety-day plan addressed to whoever can act.
Every engagement survey can tell an organisation that engagement is at some number. Almost none can tell one person which of the things that drive engagement is short in their job, and which of those they would actually miss, and that second question is the whole difference between an action plan that works and one that does not. Acting on a driver that is short but unimportant to the person is how engagement action plans fail, and it is the ordinary way they fail. This instrument exists for that question.
The six drivers are the ones the published work-values and job-resources research keeps finding, anchored on the O*NET Work Values and Work Context taxonomies published by the United States Department of Labor and in the public domain, cited as a public description of what work supplies and with no endorsement claimed: the work itself and what it asks; say over how the work is done; knowing what good looks like and being told; being asked before decisions that affect you; being noticed for work done; and somewhere to go next. Every item is original and no commercial survey's items or scale names appear.
Presence is counted, not felt. Twelve six-point ratings with a named end on each side and no midpoint, eight agreement statements, six plain true-or-false descriptions of the job, and six counts over the last three months on a slider: one-to-ones held, times asked before a decision, times told what good looked like, days of work you chose yourself, times your work was named in front of others, training days you actually took. Each driver is placed against an ordinary respondent computed from a declared answer share on every item, never the scale midpoint, as one of three words, short here, about ordinary or well supplied, with its 68 and 95 per cent band drawn. No driver gets a number: six drivers share thirty-two presence items, five or six each, and a figure on a five-item scale would be a lie of precision, so the refusal is printed in every row where the figure would have gone.
Importance is measured by trade. Eight forced-choice blocks each offer four drivers as four equally good things a job can give you and ask which you would least willingly give up and which you would give up first. Your own average is then removed, so high importance on everything is arithmetically impossible: if one driver rises, another falls. That is the design, because importance on a rating scale is uniformly high and tells nobody anything. The weights run from 0.25 to 1.00, never to zero, because a driver you would trade first still costs you something when the job supplies none of it, and a weight of zero would print no priority for an absent driver, which is arithmetic rather than a finding. The priority on each driver is your weight times its shortfall, read against what the ordinary respondent with your weights would get, and drawn as a 68 and 95 per cent band with no point inside it. Drivers are ranked by that priority, and two drivers whose bands overlap share one rank as a tie group, with the page saying why, so what you read is the separation the measurement supports and never a bare order.
There is no total, and the refusal sits at the top of the page in the largest type on it, beside the two jobs that would have shared one: a job excellent on two drivers and absent on two, and a job moderate on all of them. This is a different object from the Work Environment and Job Demands Assessment in this catalogue, which describes five conditions of a job on bipolar poles and refuses a total because it describes the job and not the jobholder. This one weighs six drivers by what you would trade and refuses a total because drivers do not substitute for one another. Every driver prints who can move it, you, your manager or the organisation, and the top tie group becomes a ninety-day ladder: now, thirty days, ninety days, one if-then action and one observable per stage, each addressed to whoever can take it, ending in a booked re-measurement with the movement that would count as real change printed on it. Where the top group holds three or more drivers the page says the measurement cannot separate them, names the two your own weights put highest, and says what would.
What you walk away with
One row per driver in rank order: a three-way placement against the ordinary respondent, your weight, the priority drawn as a 68 and 95 per cent band with no point inside it and written in plain words, and who can move it. Drivers whose bands overlap share one rank as a tie group, and the page says why. Zero on the scale is nothing short that you would miss.
Eight forced-choice blocks turn into one centred weight per driver, printed with the raw counts behind it: how many blocks it appeared in, how often you kept it, how often you traded it. The page states that high importance on everything is impossible here, because your own average has been removed.
Largest type on the page: there is no total, with a job excellent on two drivers and absent on two drawn beside a job moderate on all of them, sharing one mean. Beneath it, bordered and full size: no driver carries a number either, because five or six items a driver can place it and draw its band but not give it a figure, and every row prints that refusal where the figure would have gone.
Every row names the reader, the manager or the organisation. The top tie group becomes three rungs, now, thirty days and ninety days, each with one if-then action naming a specific situation and a specific behaviour, addressed to whoever can take it, and one thing you would see. A top group of three or more is printed as one the measurement cannot separate, with the two your own weights put highest named.
The ladder ends in a date ninety days out and, per driver, the movement a re-sitting would have to exceed to count as real change rather than measurement error, printed as the width the presence band must move by, computed from a declared reliability and stated as a measurement threshold, distinct from the smaller number whoever spends the budget would care about.
An empty sitting scores exactly zero and describes nothing. One button pressed throughout is refused in words: one press worded two ways is still one press. No driver carries a number on this form and every row says why; a driver too thin even to place is withheld with the reason printed; and a reader for whom everything is present is told that is a good result, not a flat one.
Inside your report
Illustrative sample — your report is generated from your own responses.
The bands support 3 ranks across 5 ranked drivers. Drivers inside a shared rank cannot be ordered by this form; drivers in different ranks can.
Rank 1, shared by 3 drivers: their 95 per cent priority bands all share a common point, so this form cannot say which of them is the higher priority.
5 of 5 presence items answered; a driver needs 8 to carry a number, which none on this form has. Omega 0.729 clears .70 and does not rescue it. A band and a placement, never a figure.
5 of 5 presence items answered; a driver needs 8 to carry a number, which none on this form has. Omega 0.729 clears .70 and does not rescue it. A band and a placement, never a figure.
6 of 6 presence items answered; a driver needs 8 to carry a number, which none on this form has. Omega 0.764 clears .70 and does not rescue it. A band and a placement, never a figure.
Rank 2 on its own: its band shares no point with the band of any driver ranked beside it.
6 of 6 presence items answered; a driver needs 8 to carry a number, which none on this form has. Omega 0.764 clears .70 and does not rescue it. A band and a placement, never a figure.
Rank 3 on its own.
5 of 5 presence items answered; a driver needs 8 to carry a number, which none on this form has. Omega 0.729 clears .70 and does not rescue it. A band and a placement, never a figure.
Only 2 of the 5 presence items in this driver were answered, under the 3 a placement needs. Nothing is printed for this driver.
| Rank | Driver | Placement | Weight | Priority 95% | Mover |
|---|---|---|---|---|---|
| 1 (tie of 3) | The work itself | ▼ short here: less than jobs ordinarily supply | 1.00 | 54 to 90 | the organisation |
| 1 (tie of 3) | Say over how the work is done | ▼ short here: less than jobs ordinarily supply | 0.73 | 25 to 57 | your manager |
| 1 (tie of 3) | Knowing what good looks like, and being told | ▬ about what jobs ordinarily supply | 0.72 | 23 to 51 | your manager |
| 2 | Being asked before decisions that affect you | ▬ about what jobs ordinarily supply | 0.42 | 16 to 34 | your manager |
| 3 | Being noticed for work done | ▲ well supplied: more than jobs ordinarily supply | 0.25 | 6 to 34 | the organisation |
| unranked | Somewhere to go next | withheld | withheld | withheld | your manager |
How to read it: the priority is your weight times the shortfall, drawn as a band because five or six items cannot support a point, and the rank is the order the bands support: two drivers whose 95 per cent bands share a point share one rank. High importance on everything is impossible here, because your own average has been removed, and the weights run from 0.25 to 1.00 rather than to zero, because a driver you would trade first still costs you something when it is absent. Marks: ▼ short, ▬ ordinary, ▲ well supplied, ■ withheld.
The top group has 3 drivers whose 95 per cent priority bands all share a point, and this form cannot separate them at forty items. The plan takes the two your own weights put highest, being asked before decisions and the work itself, and that choice is yours rather than the measurement's. What would separate them: watching both for a fortnight and counting the occasions, or a longer form with eight presence items a driver.
If I first hear of a change to my work after it has been decided, then I ask the person who told me, that day, who made the decision and whether anyone who does the work was asked.
What you would see: One change this week where you know who decided and whether anyone doing the work was asked.
If a piece of work comes up this week that I could take from start to finish, then I ask for the whole of it, by name, before it is split.
What you would see: One piece of work with your name on both its start and its end.
If a decision that changes how a team works is being prepared, then whoever owns the decision asks at least one person who does the work, in writing, before it is final, and records the answer.
What you would see: One decision in the month whose record shows a question to the people who do the work, before it was made.
If the next month's work is being shared out and this person's pieces are all routine or all part-pieces, then the manager assigns one whole piece that needs judgement, and says so when assigning it.
What you would see: One assignment in the month that was given as a whole and needed thought.
If the ask-before step has run for a month, then the organisation writes it into how that kind of decision is made, and names the step in the process document.
What you would see: A named ask-before step in the decision process, and a re-sitting of this form.
If the whole piece was done and it went well, then the manager makes one whole piece a month the pattern, not the exception, and tells the person that is the plan.
What you would see: Three whole pieces in three months, and a re-sitting of this form.
On being asked before decisions, a re-sitting counts as real change only if its presence band has moved by more than 37 points, in either direction. On the work itself, by more than 32 points. No point figure is compared, only the bands.
These are measurement thresholds, not the smaller value-laden difference that would matter to whoever spends the budget.
| Stage | Driver | Addressed to | Observable |
|---|---|---|---|
| Now | Being asked before decisions | you | One change this week where you know who decided and whether anyone doing the work was asked. |
| Now | The work itself | you | One piece of work with your name on both its start and its end. |
| By 30 days | Being asked before decisions | the organisation | One decision in the month whose record shows a question to the people who do the work, before it was made. |
| By 30 days | The work itself | your manager | One assignment in the month that was given as a whole and needed thought. |
| By 90 days | Being asked before decisions | the organisation | A named ask-before step in the decision process, and a re-sitting of this form. |
| By 90 days | The work itself | your manager | Three whole pieces in three months, and a re-sitting of this form. |
| Re-measure | ninety days from the sitting; the presence band must move by more than 37 points on asked before decisions, 32 on the work itself | ||
How to read it: every action is an if-then naming a specific situation and a specific behaviour, addressed to whoever can take it, which is you only where the act is to ask. The ladder closes into a second sitting: the threshold printed on the re-measurement is what a real change would have to exceed, and a movement inside it is drawn flat, never as an arrow.
Built for
- Any employee who wants to know which of the things that drive engagement is short in their job, which of those they would actually miss, and who can change it
- Human resources and people teams who need a per-person priority they can act on, addressed to the right person, rather than an aggregate engagement number for the organisation
- Line managers who want to know, before the next one-to-one, which driver to move first and what one action in thirty days would look like
- Anybody considering whether to stay, who wants a ranked answer to what is missing here rather than a feeling about it
Find out which driver is short here, and which of those you would actually miss
40 items across five formats · about 14 minutes · one payment covers the sitting and the report together: a placement and a band per driver, a ranking with ties, a refused total, who can move each driver, and a ninety-day ladder with a booked re-measurement · ₹499 in India inclusive of GST, or US$4.99 elsewhere, one time. Sit it again at ninety days.
₹499 (incl. GST) · assessment and full report, nothing further to pay
Frequently asked questions
No. There is no engagement score and no total on the page, and the refusal of a total sits at the top in the largest type. The report measures six drivers of engagement two ways: how much of each the job supplied over the last three months, counted in acts and occasions, and how much each matters to you, measured by forced choice. From those it produces, per driver, a placement and a drawn band, never a number, a ranking with ties where the bands overlap, and a ninety-day ladder from the top tie group. It does not measure how engaged you are, your performance, your personality, your health or your wellbeing.
Importance is measured by trade. Eight blocks each offer four drivers as four equally good things a job can give you and ask which you would least willingly give up and which you would give up first. Each driver's weight is how often you kept it minus how often you traded it, over the blocks it appeared in, with your own average removed and the result rescaled onto a band from 0.25 to 1.00: the driver you would least willingly give up sits at 1.00 and the one you would give up first at 0.25, not zero, because it still costs you something when the job supplies none of it. Asked on a rating scale, importance comes back uniformly high, everything matters, and that is noise against a ceiling. With your own average removed, high importance on everything is arithmetically impossible, which is the point.
That instrument measures five psychosocial conditions of a job on bipolar poles and refuses a total because it describes the job, not the jobholder. This instrument is a different object: it pairs six engagement drivers' presence with your own centred importance weights to produce a ranked priority with ties, and it refuses a total for a different reason, that drivers do not substitute for one another, so a job excellent on two drivers and absent on two would share a total with a job moderate on all four. The report prints both refusals in plain words, and the two instruments are built to be sat separately.
The ladder is built from the top tie group only, because a plan with six workstreams is not a plan; where that group holds three or more drivers the page says the measurement cannot separate them and names the two your own weights put highest. It has three rungs, now, thirty days and ninety days, and each rung carries one if-then action naming a specific situation and a specific behaviour, and one thing you would see if it happened. Each action is addressed to whoever can actually take it: you at the first rung where the act is to ask, your manager where the act is to assign, state or name, and the organisation where the act is a decision process or a training day. The ladder ends in a booked re-measurement date with the movement that would count as real change printed on it.
One payment covers the sitting and the report together: ₹499 in India inclusive of GST, or US$4.99 elsewhere, one time. Forty items across five formats take about fourteen minutes. An unanswered item leaves the denominator rather than scoring zero, so a short sitting is short rather than poor. Fewer than twenty-four of forty answered is not reportable. No driver carries a number on any sitting, because six drivers share thirty-two presence items and a figure needs eight; every driver is a placement, a band and a rank, with the refusal printed where the figure would have gone. A driver with fewer than three answered presence items is withheld with the reason printed, and fewer than four trade blocks answered withholds every weight and every rank while the placements still print. You can sit it again at ninety days, and the report names that date and its threshold.
Each one takes a single capability, puts you inside the situations where it is actually tested, and scores your choices against published evidence — with a report designed for that capability alone, not a template. They span hiring, compliance, education, operations and personal skill.
Browse the catalogue →Methodology: Forty original items across five formats: twelve six-point ratings with a named end on each side and no midpoint, eight six-point agreement statements, eight forced-choice most-and-least blocks, six counts on a slider and six true-or-false descriptions of the job. DECLARED RESPONSE INSTRUCTION: one instruction is declared for the whole instrument and it is a SITUATION-AND-PREFERENCE REPORT. Thirty-two presence items ask what has happened in the reader's job over a declared recall window, the last three months, counted in acts and occasions rather than feelings. Eight importance blocks ask what the reader would trade. The two strands are declared separately and never mixed inside one scale; no item asks what should be done, none tests what the reader knows, and none is a hypothetical about somebody else's job. CONSTRUCT STATEMENT: this measures, for one job over one three-month window, how much of each of six engagement drivers the job supplied, and how much the reader weighs each driver against the others; from those two it produces a ranked priority. It does not measure how engaged the reader is, the reader's performance, ability, personality, health, wellbeing, burnout or stress, and it is not a verdict on the manager or the organisation. HOW IT DIFFERS FROM THE WORK ENVIRONMENT AND JOB DEMANDS ASSESSMENT IN THIS CATALOGUE: that instrument measures five psychosocial job conditions on bipolar poles and refuses a total because it describes the job and not the jobholder. This instrument is a different object: it pairs six drivers' presence with the reader's own ipsatised importance weights to produce a RANKED PRIORITY WITH TIE GROUPS, and it refuses a total for a different reason, that drivers do not substitute for one another, so a job excellent on two and absent on two would share a total with a job moderate on all four. Both refusals are printed on their own pages in the reader's words. FRAMEWORK ANCHOR: the six drivers are anchored on the O*NET Work Values and Work Context taxonomies, published by the United States Department of Labor, Employment and Training Administration, and in the public domain: achievement, independence, recognition, support, relationships and working conditions among the Work Values, and freedom to make decisions, structured versus unstructured work and the frequency of contact among the Work Context descriptors. O*NET is the most widely used published description of what work supplies to the people doing it; it is cited as a public taxonomy, no endorsement by its publisher is claimed or implied, and none of its survey items is reproduced. The influence-at-work and role-clarity constructs are additionally anchored on their published definitions in COPSOQ III, the Copenhagen Psychosocial Questionnaire, free by licence for non-commercial research use; only the construct definitions are cited and not one of its items is reproduced or adapted. SCORING DESIGN, C25, importance-weighted shortfall with an ipsatised importance vector and a refused total. Every answered presence item is reduced to a quality q in [0,1] toward the driver-present pole, honouring reverse keying: a six-point rating or agreement is (v - 1) / 5, a true-or-false description is 1 at the present pole, a slider count is its position in its range. present_k is 100 times the mean q over the ANSWERED presence items in driver k; ref_k is 100 times the mean of E[q] under each item's DECLARED marginal, value_base_rate on ratings, statements and sliders and option_base_rate on true-or-false claims, never the scale midpoint and never a uniform draw; departure_k = present_k - ref_k, so zero is the ordinary respondent; shortfall_k = 100 - present_k, continuous, with no clamp and no floor pile-up. IMPORTANCE comes from the forced-choice block only: w_raw_k = (most_k - least_k + n_k) / (2 n_k), where n_k is the number of answered blocks in which driver k appeared, five or six of the eight; w_c_k = w_raw_k - mean(w_raw), IPSATISED so the six centred weights sum to zero by construction; w_k = 0.25 + 0.75 (w_c_k - min(w_c)) / (max(w_c) - min(w_c)) rescales w_c onto a band from 0.25 to 1.00 and is the printed weight, with every w_k set to 0.625, the band's midpoint, when the six are equal. The floor is not zero on purpose: a plain zero-to-one rescale puts the driver the reader would give up first at exactly zero on every sitting, so its priority reads as nothing even when the job supplies none of it, a mechanical artefact of the rescale printed as a finding; a driver you would trade first still costs you something when it is absent. Importance on a rating scale is uniformly high, everything matters, and is noise against a ceiling; forced choice with within-person centring is the published answer to that, and the report prints that high importance on everything is arithmetically impossible here because the reader's own average has been removed. PRIORITY: priority_k = w_k times shortfall_k is the ranked quantity; priorityRef_k = w_k times (100 - ref_k) is the same reader's weights against the ordinary respondent's presence; vsOrdinary_k = priority_k - priorityRef_k is the chance-corrected quantity, which isolates what is short about THIS job rather than what is short about jobs. All of these are computed inside the builder because the ORDERING depends on them, and none is printed as a number. WHAT REACHES THE PAGE PER DRIVER: a three-way placement against the ordinary respondent, short here, about ordinary or well supplied, decided at ten presence points either side of the reference and printed as a decision; a drawn 68 and 95 per cent presence band; a drawn 68 and 95 per cent priority band with no point inside it; the weight, as a number, because it is the reader's own trade over eight blocks and not a measurement of the job; and a rank. ORDERING WITH TIE GROUPS: drivers are ranked by vsOrdinary, largest first, and drivers whose 95 per cent priority bands share a common point are printed as one tie group with one rank and a caption saying why; the separation the bands support is printed, never a bare rank, and a ladder in which everything ties is a legitimate outcome that reads as one. The TOTAL is REFUSED, and the refusal occupies the top of the page in the largest type on it, beside the two opposite profiles that would have shared the number. Every driver prints who can move it, the reader, the manager or the organisation, and its action is addressed to that person. RELIABILITY, AND WHY NO DRIVER CARRIES A NUMBER: McDonald's omega in its Spearman-Brown form from the answered presence-item count and an assumed average inter-item correlation of .35, printed as an assumption on the unrounded value. The subscore rule is conjunctive: a driver with fewer than eight answered presence items carries a three-way placement and no figure, AND a driver with omega under .70 carries no figure. Six drivers share thirty-two presence items, five or six each, so no driver reaches the eight-item floor on this form, by design and not as a defect: every driver is a placement and a band, and the refusal of a figure is printed in every row where the figure would have gone, with what would change it, eight presence items a driver on a longer form. The priority band's half-width is the weight times the presence standard error, sd_k times the square root of one minus omega_k, with sd_k a declared modelled standard deviation printed as an assumption and the weight treated as known; every priority band is drawn and written at 68 and 95 per cent. RE-MEASUREMENT: the ladder is built from the TOP TIE GROUP, one or two members directly; three or more means the page says the measurement cannot separate them at this length, names the two the reader's own weights put highest, and says what would separate them. The booked re-sitting carries, per plan driver, the movement threshold 1.96 times sd_k times the square root of 2(1 - omega_k) from the Reliable Change Index, printed as a width the presence band must move by rather than as a target figure, and distinguished in words from the smaller, value-laden difference that would matter to whoever spends the budget. REFUSAL RULES: an unanswered item leaves the denominator and the reference together; an empty sitting scores exactly zero on every presence, departure, shortfall, weight and priority and is not reportable; fewer than twenty-four of forty answered is not reportable; the flatness test runs on the RAW pressed values of the twenty six-point items, never on the keyed or centred ones, because eight of those items are worded at the absent pole and one button pressed throughout would produce a shape entirely of our authoring, so such a sitting is refused outright rather than scored; the flat-baseline gate runs on vsOrdinary and on the priority spread, never on priority alone, because a reader for whom everything is present scores near zero priority everywhere and that is a valid good result; fewer than four answered importance blocks, or any driver appearing in fewer than two answered blocks, refuses the weight vector and with it every priority band and every rank, while the placements and presence bands still print. No percentile appears anywhere, no driver-level number appears anywhere, and no reader is ranked against anybody. SOURCES: O*NET Work Values and Work Context taxonomies, United States Department of Labor, Employment and Training Administration, O*NET Resource Center, public domain, and Dawis and Lofquist, A Psychological Theory of Work Adjustment (University of Minnesota Press, 1984), on which the Work Values are built; Burr, Berthelsen, Moncada et al., The Third Version of the Copenhagen Psychosocial Questionnaire (Safety and Health at Work, 2019), for the influence-at-work and role-clarity construct definitions; Bakker and Demerouti, The Job Demands-Resources model: state of the art (Journal of Managerial Psychology, 2007), and Demerouti, Bakker, Nachreiner and Schaufeli, The job demands-resources model of burnout (Journal of Applied Psychology, 2001), for the treatment of autonomy, feedback, support and growth as resources; Hackman and Oldham, Motivation through the design of work (Organizational Behavior and Human Performance, 1976), for skill variety, task identity, autonomy and feedback from the job; Deci and Ryan, The what and why of goal pursuits (Psychological Inquiry, 2000), and Gagne and Deci, Self-determination theory and work motivation (Journal of Organizational Behavior, 2005), for autonomy, competence and relatedness; Steelman, Levy and Snell, The Feedback Environment Scale (Educational and Psychological Measurement, 2004), for the feedback-environment construct, cited as a construct only; Morrison, Employee voice behavior: integration and directions for future research (Academy of Management Annals, 2011), and Kahn, Psychological conditions of personal engagement and disengagement at work (Academy of Management Journal, 1990), for voice and for psychological availability; Christian, Garza and Slaughter, Work engagement: a quantitative review and test of its relations with task and contextual performance (Personnel Psychology, 2011), and Harter, Schmidt and Hayes, Business-unit-level relationship between employee satisfaction, employee engagement, and business outcomes: a meta-analysis (Journal of Applied Psychology, 2002), for engagement and its antecedents; Gollwitzer and Sheeran, Implementation intentions and goal achievement: a meta-analysis of effects and processes (Advances in Experimental Social Psychology, 2006), for the if-then action design; Jacobson and Truax, Clinical significance: a statistical approach to defining meaningful change (Journal of Consulting and Clinical Psychology, 1991), for the Reliable Change Index behind the re-measurement threshold; Brown and Maydeu-Olivares, How IRT can solve problems of ipsative data in forced-choice questionnaires (Psychological Methods, 2013), Hicks, Some properties of ipsative, normative, and forced-choice normative measures (Psychological Bulletin, 1970), and Salgado and Tauriz, The five-factor model, forced-choice personality inventories and performance: a comprehensive meta-analysis (European Journal of Work and Organizational Psychology, 2014), for the ipsative weight vector and its limits; Kraut (editor), Getting Action from Organizational Surveys (Jossey-Bass, 2006), Church and Waclawski, Designing and Using Organizational Surveys (Jossey-Bass, 2001), and Macey and Schneider, The meaning of employee engagement (Industrial and Organizational Psychology, 2008), for why engagement action plans fail, chiefly by acting on what is short rather than on what is short and matters, and by addressing the action to the wrong person; Schwarz, Self-reports: how the questions shape the answers (American Psychologist, 1999), for the fixed recall window and counting behaviour rather than rating feeling; Baumgartner and Steenkamp, Response styles in marketing research (Journal of Marketing Research, 2001), and Weijters, Baumgartner and Schillewaert, Reversed item bias (Psychological Methods, 2013), for balanced keying and the raw-press flatness rule. Every item is an original work written for this instrument. No commercial engagement survey's items, scale names, driver names or report sections are reproduced, and no vendor, instrument or certification is named on any candidate-facing or store surface.