.png)
Feedback for a Difficult Boss: How to Say It So It Lands.
Every executive assistant has a sentence they have been carrying for a while. It is the one about the Friday-night emails, or the third "quick favour" that turned into a fortnight, or the meeting that gets moved four times and lands on your lunch. You have rehearsed it in the car. You have said it to a friend, and the friend said "you have to tell him", and you agreed, and then Monday came and the sentence went back in the drawer. I wrote about giving feedback to a difficult boss in 2023 and the advice held: be clear and direct, set expectations early, and look for the compromise when the first answer is no. What I did not have then was the research on why the sentence stays in the drawer, what the executive on the other side of it is actually experiencing, and which versions of it land. It turns out to be a well-studied problem, and the findings are kinder to both of you than the word "difficult" suggests.
What you will get out of this
Why the executive is usually working from a picture of themselves that nobody has corrected, and what happens to the ones who get the correction, from five years of upward feedback data on 252 managers
The landing test: four questions that decide whether feedback is heard, from experiments that varied the message, the source, the setting and the timing
The three-line feedback, a way of saying the sentence that the research on how managers rate what they hear keeps rewarding, with a builder that assembles it from what you type
The expectations ledger, a page for the tasks that fall outside the job, and the finding that you and the executive will disagree about which ones those are
The line, for the rare case where the problem is not the channel, and the two people to speak to before you decide anything
Three printable pages: the landing test, the three-line script and the ledger
The picture nobody has corrected
Start with what the executive can see, because it is less than you think. In a health care organisation where 146 staff were surveyed, fifteen interviewed and six focus groups convened, informal feedback travelling upwards was mostly absent. Where it did travel, it was inaccurately positive. Senior managers were unaware of the distortion, unwilling to consider that it existed, had an exaggerated impression of how much upward feedback they were getting, and discouraged the critical kind. Nobody in that study was a villain. The people at the top believed the flattering version because the flattering version was the only one delivered, and the people below delivered it because the honest one seemed to cost too much.
The cost is the interesting part. When 40 employees were interviewed about issues they had chosen not to raise with a supervisor, most had a story, several could not think of any issue they would feel comfortable raising, and the most common reason was fear of being viewed negatively and, through that, damaging a relationship they valued. A later programme of four studies gave the fear a shape. People carry taken-for-granted rules about when speaking up is dangerous, absorbed long before the current job, and two of the rules are worth naming because this post breaks both on purpose: do not embarrass the boss in public, and do not speak up without solid data or a solution. The rules are, as you will see, close to what the evidence on landing says. The trouble is that they are usually applied as a reason to say nothing at all.
Most executives are working from a flattering picture of themselves that nobody has corrected. Correcting it, kindly and in private, is a service.
Now the other half, which is what happens to the executives who do get the correction. A five-year study followed 252 managers through five annual rounds of upward feedback from their teams. The managers who started poorly or moderately rated improved, beyond what regression to the mean would predict. The managers who met with their direct reports to discuss the feedback improved more than those who did not, and any given manager improved more in the years they discussed the previous round than in the years they let it pass. In a quasi-experiment in an international professional services firm, managers who received upward feedback were rated by their own subordinates as performing significantly better six months later, against both their own baseline and a comparison group, with the effect strongest in the managers who believed they could change. A meta-analysis of 24 longitudinal studies is the sober footnote: improvement after multisource feedback is on average small, and it is largest when the recipient sees that change is needed, believes it is feasible and sets a goal. Which is to say the feedback works when it arrives in a form that can be acted on, in a conversation, from someone the manager wants to keep.
One more, because it is the cheerful surprise in this literature. In a field experiment, leaders randomly assigned to share with their teams the critical feedback they had already received about themselves raised their team's psychological safety a year later. Leaders assigned merely to ask for feedback did not. When the researchers went back two years on, the askers had let the vulnerability dissolve in defensiveness and inaction, while the sharers had made a public commitment that their teams then reciprocated. The executive who repeats your feedback to the wider team is not weakened by it. On this evidence they become the leader people tell things to.
The research behind this section
Tourish and Robson (2004) audited a health care organisation with a survey of 146 staff, fifteen interviews and six focus groups. Informal upward feedback was mostly absent; where present it was inaccurately positive; senior managers were unaware of the distortion, overestimated how much feedback they received and discouraged the critical kind.
Milliken, Morrison and Hewlin (2003) interviewed 40 employees. Most had withheld a concern from a supervisor; the most frequently given reason was fear of being seen negatively and damaging valued relationships. Detert and Edmondson (2011), across four studies, described implicit voice theories, the taken-for-granted rules about when speaking up is risky, including not embarrassing the boss in public and not speaking up without solid data or solutions.
Walker and Smither (1999) followed 252 managers through five annual rounds of upward feedback. Managers rated poor or moderate at the start improved beyond regression to the mean; those who met their direct reports to discuss the results improved more, and more in the years they discussed than the years they did not. Heslin and Latham (2004) found managers given upward feedback were rated higher by subordinates six months later than at baseline and than a comparison group, moderated by self-efficacy. Smither, London and Reilly (2005) meta-analysed 24 longitudinal studies: average improvement after multisource feedback is small and is most likely when the recipient perceives a need to change, believes it feasible and sets goals.
Coutifaris and Grant (2022): CEOs' feedback-seeking and feedback-sharing each predicted board ratings of top-team psychological safety; in a longitudinal field experiment, leaders assigned to share feedback raised team psychological safety one year later, while those assigned to seek it did not; interviews two years on explained why.
The landing test
So the executive benefits, the relationship survives, and the rules in your head are half right. The question becomes which sentence, said where, said when. The research on how managers receive what they are told has been unusually practical about this, because it has run the experiments. In three laboratory studies, researchers varied six things about an employee who spoke up and measured how observers rated that employee's performance afterwards: the speaker's expertise and trustworthiness, whether the message recommended a solution and whether it was framed positively or negatively, and whether it came early enough to act on and into a culture that expected people to speak. At least one factor from each of the three groups (source, message, context) moved the ratings, and most of the movement ran through three judgements the observer was making without noticing: whether they liked the speaker, whether the speaker seemed to be acting for the organisation rather than for themselves, and whether the contribution was constructive.
Setting matters more than most people expect, and in the direction you would hope. Across five studies, correlational and experimental, managers endorsed suggestions raised privately, one to one, more than the same suggestions raised in front of others, and the gap was widest when the relationship was weak. The explanation the studies supported was face: a point made in public reads as a challenge to the manager's image as competent, and the manager defends the image rather than weighing the point. The employees in a separate pair of studies had worked this out for themselves. They took issues to public channels or formal procedures only when three conditions held at once: the issue was important, the manager was known to be open, and the manager was in a good mood at the time. The rule about not embarrassing the boss in public is, on this evidence, an accurate reading of how bosses work.
Timing has two findings. The first is about the manager's tank. In an experience-sampling study of 62 managers over ten working days, and an experiment with 198 more, managers who were depleted, running on empty after a day of self-control, were less likely to weigh a suggestion carefully and more likely to reject it, defaulting to the status quo, and that effect was stronger when they saw the person speaking as less expert. The finding has a cruel symmetry: the moment the executive most needs the feedback is the moment they are least able to take it. Late on a bad day is the worst slot in the week. The second finding is about how often. Across three organisations, managers gave the best evaluations to people who raised concerns at a moderate rate, and marked down both those who rarely spoke and those who spoke constantly, with the shape depending on how good the relationship was. A related programme, a qualitative study, two multi-source field studies and a preregistered experiment, found that people who deliberately held back untimely points, which the researchers called strategic silence, had their eventual contributions rated as higher quality and were evaluated and rewarded better for them. Silence, chosen and temporary, is part of how the sentence lands.
Try this. Before the sentence leaves the drawer, run it through the four questions in the graphic. Is it private? Is there a proposal in it, or only a problem? Is it early enough to change anything, and is the executive's tank full enough to hear it? Have you said something like it recently, or is it the first in a while? Two or more wrong answers means the sentence waits, and waiting is itself a tactic.
The research behind this section
Whiting, Maynes, Podsakoff and Podsakoff (2012) ran three laboratory studies manipulating two source factors (expertise, trustworthiness), two message factors (recommending a solution, positive versus negative framing) and two context factors (timing, organisational norms for speaking up). At least one variable from each category affected performance evaluations of the speaker, mostly indirectly through liking, perceived prosocial motives and perceived constructiveness.
Isaakyan, Sherf, Tangirala and Guenter (2021), across five correlational and experimental studies, found managers endorse voice raised privately more than the same voice raised publicly, especially when the relationship is weak; the mechanism supported was face management. Xu, Huang, Ouyang, Liu and Hu (2020), in two studies, found employees use public and formal channels only when the issue is important, the manager is perceived as open and the manager is in a positive mood, and that public and formal voice was related to success.
Li, Barnes, Yam, Guarana and Wang (2019): 62 managers surveyed about voice events over ten days and an experiment with 198 managers. Depleted managers processed voice less carefully and endorsed it less, more so when the voicer was seen as low in expertise. Huang, Xu, Huang and Liu (2018), across a Chinese bank and two US firms, found an inverted-U between the frequency of prohibitive voice and managers' evaluations, moderated by leader-member exchange, while the frequency of promotive voice was unrelated to promotability. Parke, Tangirala, Sanaria and Ekkirala (2022): a qualitative study, two multi-source field studies and a preregistered experiment showed that strategic silence, the intentional withholding of untimely ideas or concerns, led managers to judge the employee's voice as higher quality and reward it more.
The three-line feedback
The 2023 post said be clear and direct, and present facts rather than opinions. The research agrees and adds the part that decides whether "direct" reads as helpful or as hostile, which is the shape of the sentence. The finding that organises everything else came from a field study and a set of experiments on how managers respond to two kinds of speaking up. When employees challenged how things were done, managers read them as less loyal and more threatening, rated their performance lower and endorsed their ideas less. When employees spoke in ways that supported the existing direction while improving it, the reverse held. A meta-analysis of the voice literature found the same split at scale: suggestions that promote something new are associated with better performance ratings, while voice that only prohibits or objects is associated with worse. And an experiment with 384 German managers found they rated high-quality proposals higher, rated promotive voice above prohibitive, and were not moved by their own mood, which is a useful reminder that the content carries more than the weather.
None of this says soften the message. Quality, in the study that measured it, had four parts: a rationale, a feasible proposal, a focus on the organisation rather than the speaker, and something new. Across four scale-development studies and time-lagged surveys of peers and managers, quality predicted performance and promotion ratings above and beyond how often the person spoke up. The people who did best were the ones whose contributions carried reasons and a workable next step, however rarely they made them. Add the status finding, from a survey and two experiments: people who spoke up were ascribed higher standing than those who stayed quiet, because speaking read as both competent and helpful. The image risk the 40 interviewees feared is real for a complaint and largely reversed for a proposal.
So the three-line feedback. Line one is the observation, specific and recent, with no adjective attached to the executive. Line two is the cost, stated in terms of something the executive cares about: their time, their reputation with the board, the thing that will slip. Line three is the ask, one change, small enough to say yes to in the room. "The Friday-night emails are landing at ten and eleven. I open them and lose the weekend to them, and by Monday I am slower on the things you actually need. Could the ones that are not urgent go in a Monday-morning note instead? I can draft the template." Observation, cost to their goal, proposal. It is direct. It is also, in the vocabulary of the research, promotive, high quality and private, and it lands.
If you would rather build it than write it, the builder below takes the observation, the cost and the ask as you type them, checks them against the landing test, and hands back the three lines in a shape you can read out or paste into a note.
Try this. Take the sentence from the drawer and split it into the three lines. If line one contains a word like "always" or "never", replace it with a date and a number. If line two describes a cost to you and not to the executive, keep it, then add the one that is theirs. If line three asks for more than one thing, keep the smallest.
The research behind this section
Burris (2012), in a field study and experiments, found that managers respond to challenging voice with lower performance ratings and less endorsement, reading the speaker as less loyal and more threatening, and respond to supportive voice in the opposite way. Chamberlin, Newton and LePine (2017), in a meta-analysis of voice and its forms, found promotive voice positively associated with job performance and prohibitive voice negatively. Stumpf (2025) ran a between-subjects experiment with 384 German managers: evaluations were higher for high-quality voice and for promotive over prohibitive content, and managers' mood had no direct or indirect effect.
Brykman and Raver (2021) developed a measure of voice quality (rationale, feasibility, organisational focus, novelty) across four studies and, in time-lagged surveys of peers and managers, found quality predicted performance and promotion ratings above and beyond voice frequency. Weiss and Morrison (2019), in a survey and two experiments, found employees who voiced were ascribed higher status, mediated by judgements of agency in all three studies and communion in two.
The expectations ledger
The 2023 post's second piece of advice was to set expectations early, so that "there are no surprises later when they ask for something more". That advice has since acquired a research programme of its own, about what the extra asks do to the person who absorbs them. The concept is the illegitimate task: work that falls outside what someone in your role can reasonably be expected to do, either because it is unnecessary or because it is unreasonable. In three studies, illegitimate tasks predicted low self-esteem, resentment towards the organisation and burnout, over and above role conflict and unfairness, and in the third, run over two months, they predicted later strain rather than being a symptom of it. A 2026 meta-analysis of 141 studies and 65,124 people confirmed the pattern across commitment, satisfaction, engagement, wellbeing and performance, and compared the two kinds. For this seat, which is defined by breadth, the finding needs handling with care. Most of what you do is legitimate because you chose a job that has no edges. The tasks that erode people are the unreasonable ones, the ones that quietly say "you are not worth someone else's time", and in the 214-employee study below it was those, and not the merely unnecessary, that carried the cost.
The finding that turns this into a tool is about disagreement. In 166 sets of three, an employee, their supervisor and their partner at home, supervisors' assessments of which of the employee's tasks were illegitimate converged only modestly with the employee's own, and converged better when the supervisor had fewer people to manage. The tasks still predicted strain whichever of the two reported them, so the effect is real; the point is that the executive does not see the same list you see. What looks to you like a pattern of impositions looks to them, if it looks like anything, like a handful of one-off asks. Neither of you is lying. You are keeping different books.
The buffer, from 214 early-career employees, is the encouraging bit. Unreasonable tasks explained extra anxiety and low mood beyond ordinary demands and resources, and the relationship was weaker for people who felt supported by their supervisor and had some control over their work. Which is the case for a ledger, kept for a fortnight and then shown: it converts a feeling into a list, and it gives the executive the chance to be the supporting supervisor rather than the source.
Try this. For two weeks, keep the ledger in the graphic: the task, the date, which kind it was (outside the role, or the role but the wrong person or time), what it displaced and what it cost in hours. Then book twenty minutes and take the page in with the three-line feedback already written for the row that matters most. Think of it as reconciling two sets of books rather than presenting a grievance.
The research behind this section
Semmer, Jacobshagen, Meier, Elfering, Beehr, Kälin and Tschan (2015) reported three studies: in two cross-sectional samples, illegitimate tasks predicted low self-esteem, resentment towards the organisation and burnout, controlling for role conflict and for distributive, procedural and interactional justice; in a third, over two months, they predicted later strain controlling for initial strain. Li and Mostafa (2026) meta-analysed 141 studies (N = 65,124) across attitudinal, wellbeing and behavioural outcomes and compared unreasonable with unnecessary tasks.
Meier and Semmer (2018) collected 166 triads of employee, supervisor and partner. Employee and supervisor reports of illegitimate tasks converged only modestly, more when the supervisor's span of control was small; both reports predicted exhaustion, supervisor-rated incivility and partner-rated work-family conflict. Fila and Eatough (2020): in 214 early-career employees, unreasonable tasks explained additional variance in anxiety (9.9%) and depressive symptoms (7.3%) beyond demands and resources, and supervisor support and control buffered the relationship.
The trade, and the line
The third piece of 2023 advice was to look for the compromise when the executive cannot or will not move. The research adds two things. First, on this evidence the compromise is where the three-line feedback should start, because a proposal is what turns a complaint into promotive voice. Second, the compromise you offer should cost them less than the thing you are asking them to give up: a template, a different day, a rule for what counts as urgent. In the depletion study the managers who rejected suggestions were defaulting to the status quo because thinking was expensive; an offer that has done the thinking for them removes the reason to say no.
And then the line. Everything above assumes the executive is demanding, which is a description of the job, and is reachable, which the five-year study says most are. A small number are neither. The research word is abusive supervision, and a meta-analysis of the field found its associations with every outcome measured universally negative. Two things separate it from a hard boss. It is sustained, and it is aimed at you rather than at the work: ridicule, blame for things you did not do, the silent treatment, telling others you are incompetent. If that is the pattern, the landing test does not apply, because the problem was never the channel. Some of the mechanism has been mapped: managers who doubt their own fitness for the role solicit less input, rate the people who speak up more harshly and implement less of what they hear, and in the experiment that isolated it, ego defensiveness was what drove the aversion. Knowing that it is about their picture of themselves and not about you is a useful fact and not a solution.
If you think you may be on the far side of the line, two conversations before any decision. One with HR or the person the executive reports to, taking the ledger, because dates and hours are what those conversations run on. One with someone outside the organisation, a mentor, a friend who has done the job, or a professional if the toll has reached your sleep or your health. This post is a pattern to recognise and not a diagnosis, and nobody should carry the far side of the line alone.
The Feedback Pages below put the three tools on paper: the landing test, the three-line script with the lines blank, and the ledger with a fortnight of rows. Free, no email wall. Print the ledger before Monday and let it fill itself.
Try this. Write the compromise before the conversation, and make it one the executive can accept without thinking. If the pattern you are dealing with is on the far side of the line, take the ledger to HR this week rather than the three-line feedback to the executive.
The research behind this section
Mackey, Frieder, Brees and Martinko (2017) meta-analysed the abusive supervision literature and found its associations with outcome variables universally negative, with magnitudes varying by study design. Fast, Burris and Bartel (2014): in a field study, managers low in managerial self-efficacy solicited less input; in an experiment, inducing low managerial self-efficacy reduced voice solicitation, produced negative evaluations of an employee who spoke up and reduced implementation, driven by ego defensiveness. Li, Barnes, Yam, Guarana and Wang (2019) is the depletion study described in the landing test. Detert and Burris (2007), with 3,149 employees and 223 managers in a restaurant chain, found managerial openness was the leadership behaviour most consistently related to employees speaking up, through psychological safety, and that the effect was strongest for the best performers.
Where to start
Three moves this week. Run the sentence in the drawer through the landing test and decide whether this is the week or a later one. Rewrite it as three lines, observation, cost to their goal, ask, with the compromise built into the ask. Start the ledger, and give it a fortnight before you show anyone.
The rules in your head about speaking up are a description of how to do it, mistaken for a reason not to.
Turning a complaint into a proposal the executive can say yes to, and keeping the record that makes the case, is how influence works in this seat, which is what the Connected Leader Assessment measures under Influence and Stakeholder Trust. It is free, it takes about twelve minutes, and it will show you where you already shape decisions upward and where the sentence is still in the drawer.
And if you want to build the wider skill, saying hard things to senior people in ways that get acted on, that is what our Skill Sprints are for. One topic, built properly, an hour or two, and you come out having made something rather than having been talked at.
Meg ✌️
Free tools
The Connected Leader Assessment, five capability dimensions, twelve minutes, free
The Irreplaceable Audit, forty-eight activities across four kinds of work, free
Our Skill Sprints, one topic built properly, an hour or two, and you come out having made something rather than having been talked at
Frequently asked
What if the executive asks for feedback and then punishes it?
Two findings help. In the field experiment on feedback-sharing, leaders who only asked for feedback tended to let it dissolve in defensiveness, so the invitation on its own is a weak signal. And in 3,149 employees across a restaurant chain, what predicted people speaking up was not the invitation but whether the manager was seen as open in practice, through the safety people felt. Judge the executive by what happened last time, not by the open door. If last time cost you, use the landing test harder: private, early, a proposal, and less often.
Is it ever right to raise something in a meeting?
Rarely, and the employees in the two-study tactics research had it about right: public and formal only when the issue is important, the executive is known to be open and the mood in the room is good. The five-study endorsement research found private voice endorsed more, most of all when the relationship is weak. If the point has to be made in front of others, make it as a question and give the executive the credit for the answer.
What if I have already said it, several times, and nothing changed?
Frequency is part of the landing test for a reason. In the three-organisation study, managers marked down people who raised concerns constantly as well as those who never did. Stop repeating the observation and change the shape: one written three-line version, with the ledger attached and a compromise that costs them nothing. If that does not move it either, you are past the channel problem and into the ledger-to-HR conversation.
Should I put it in writing or say it?
Say it, then write it. The five-year study's improvement came from managers who discussed the feedback, and discussion needs a room. A short written note afterwards, the three lines and the agreed change, gives both of you the record that the ledger research says you will otherwise keep differently.
How do I know whether a task is illegitimate or simply the job?
The research definition is a task that violates what someone in your role can reasonably be expected to do, because it is unnecessary or unreasonable. The seat has no edges, so most tasks pass. The ones to log are the ones you would not ask of anyone else at your level, and the ones that say you are not worth another person's time. And expect the executive to disagree with your list; in the 166-triad study they mostly did, which is why the ledger is dated.
What if the problem is me?
Sometimes the pattern in the ledger is a pattern in the diary, and the fix is a working agreement rather than feedback. That is covered in the onboarding post below, and the two tools sit well together: the agreement for how you will work, the three lines for when it drifts.
P.S. Want to Go Deeper on Speaking Upward?
If this resonated and you want to go further, here is the evidence the piece is built on. These are the papers that shaped my thinking on what executives can see of themselves, what happens when they are told, which sentences land and which do not, what the extra asks do to the person absorbing them, and where the line sits, and they are well worth your time.
Every DOI below has been checked. Where I could not confirm a figure from the paper itself, I have left the figure out rather than guess.
📃 Tourish and Robson (2004) Communication audit of a health care organisation: 146 staff surveyed, fifteen interviewed, six focus groups. Informal upward feedback mostly absent; where present, inaccurately positive; senior managers unaware, overestimating how much they received and discouraging critical feedback. https://doi.org/10.1108/13632540410807628
📃 Milliken, Morrison and Hewlin (2003) Interviews with 40 employees. Most had withheld a concern from a supervisor; the most frequent reason was fear of being labelled negatively and damaging valued relationships. https://doi.org/10.1111/1467-6486.00387
📃 Detert and Edmondson (2011) Four studies on implicit voice theories, the taken-for-granted rules about when and why speaking up at work is risky or inappropriate, and their relationship to self-censorship. https://doi.org/10.5465/amj.2011.61967925
📃 Walker and Smither (1999) 252 managers, five annual rounds of upward feedback. Poorly and moderately rated managers improved beyond regression to the mean; those who discussed the feedback with direct reports improved more, and more in the years they discussed it. https://doi.org/10.1111/j.1744-6570.1999.tb00166.x
📃 Heslin and Latham (2004) Quasi-experiment in an international professional services firm. Managers given upward feedback were rated higher by subordinates six months later than at baseline and than a comparison group; self-efficacy moderated the effect. https://doi.org/10.1111/j.1464-0597.2004.00159.x
📃 Smither, London and Reilly (2005) Meta-analysis of 24 longitudinal studies of multisource feedback. Improvement in ratings over time is generally small, and most likely when recipients perceive a need to change, believe it feasible and set goals. https://doi.org/10.1111/j.1744-6570.2005.514_1.x
📃 Coutifaris and Grant (2022) CEOs' feedback-seeking and feedback-sharing predicted board ratings of team psychological safety; in a longitudinal field experiment, leaders assigned to share feedback raised team psychological safety one year later, while seeking did not. https://doi.org/10.1287/orsc.2021.1498
📃 Whiting, Maynes, Podsakoff and Podsakoff (2012) Three laboratory studies manipulating source (expertise, trustworthiness), message (solution offered, framing) and context (timing, norms). At least one factor per category affected evaluations of the speaker, mostly through liking, perceived motives and perceived constructiveness. https://doi.org/10.1037/a0024871
📃 Isaakyan, Sherf, Tangirala and Guenter (2021) Five studies. Managers endorsed privately raised voice more than publicly raised voice, especially when relationship quality was weak; face management explained the difference. https://doi.org/10.1037/apl0000816
📃 Xu, Huang, Ouyang, Liu and Hu (2020) Two studies of voice tactics. Employees used public channels and formal procedures only when the issue was important, the manager was perceived as open and the manager was in a positive mood; public and formal voice related to success. https://doi.org/10.1002/hrm.21992
📃 Li, Barnes, Yam, Guarana and Wang (2019) 62 managers over ten days and an experiment with 198 managers. Ego-depleted managers processed voice less carefully and endorsed it less, more so when the voicer was seen as low in expertise. https://doi.org/10.1002/job.2370
📃 Huang, Xu, Huang and Liu (2018) Three samples (a Chinese bank, a US technology firm, a US financial services firm). A moderate frequency of prohibitive voice drew the best evaluations, moderated by leader-member exchange; promotive voice frequency was unrelated to promotability. https://doi.org/10.1037/apl0000326
📃 Parke, Tangirala, Sanaria and Ekkirala (2022) A qualitative study, two multi-source field studies and a preregistered experiment. Strategic silence, the intentional withholding of untimely ideas or concerns, led managers to rate the employee's voice as higher quality and to reward it more. https://doi.org/10.1016/j.obhdp.2022.104187
📃 Burris (2012) A field study and experiments on managerial responses to challenging versus supportive voice. Challenging voice drew lower performance ratings and less endorsement, with the speaker seen as less loyal and more threatening. https://doi.org/10.5465/amj.2010.0562
📃 Chamberlin, Newton and LePine (2017) Meta-analysis of voice and its promotive and prohibitive forms. Promotive voice was positively associated with job performance and prohibitive voice negatively; antecedents differed by form. https://doi.org/10.1111/peps.12185
📃 Stumpf (2025) Experiment with 384 German managers. Evaluations were higher for high-quality voice and for promotive over prohibitive content; managers' mood had no direct or indirect effect. https://doi.org/10.1177/23970022241231540
📃 Brykman and Raver (2021) Four studies developing a measure of voice quality (rationale, feasibility, organisational focus, novelty) and time-lagged surveys of peers and managers. Quality predicted performance and promotion ratings above and beyond frequency. https://doi.org/10.1002/job.2509
📃 Weiss and Morrison (2019) A survey and two experiments. Employees who voiced were ascribed higher social status, mediated by judgements of agency in all three studies and of communion in two. https://doi.org/10.1002/job.2262
📃 Semmer, Jacobshagen, Meier, Elfering, Beehr, Kälin and Tschan (2015) Three studies. Illegitimate tasks predicted low self-esteem, resentment and burnout controlling for role conflict and injustice, and over two months predicted strain rather than being predicted by it. https://doi.org/10.1080/02678373.2014.1003996
📃 Li and Mostafa (2026) Meta-analysis of 141 studies, 65,124 people. Illegitimate tasks related to attitudinal, wellbeing and behavioural outcomes; unreasonable and unnecessary tasks compared for predictive validity. https://doi.org/10.1037/ocp0000435
📃 Meier and Semmer (2018) 166 triads of employee, supervisor and partner. Employee and supervisor assessments of illegitimate tasks converged only modestly, more when the supervisor's span of control was small; both predicted exhaustion, incivility and work-family conflict. https://doi.org/10.1080/1359432X.2018.1526785
📃 Fila and Eatough (2020) 214 early-career employees. Unreasonable tasks explained additional variance in anxiety and depressive symptoms beyond job demands and resources; supervisor support and control buffered the relationship. https://doi.org/10.1177/0033294119874292
📃 Mackey, Frieder, Brees and Martinko (2017) Meta-analysis and empirical review of abusive supervision. Associations with outcome variables were universally negative; magnitudes varied with study design. https://doi.org/10.1177/0149206315573997
📃 Fast, Burris and Bartel (2014) A field study and an experiment. Managers low in managerial self-efficacy solicited less input, evaluated a speaker more negatively and implemented less voice; ego defensiveness drove the aversion. https://doi.org/10.5465/amj.2012.0393
📃 Detert and Burris (2007) 3,149 employees and 223 managers in a restaurant chain. Managerial openness was consistently related to improvement-oriented voice, mediated by psychological safety, with the strongest effect on the best performers. https://doi.org/10.5465/amj.2007.26279183
📃 Hao, Zhu, He, Duan, Zhao and Meng (2022) Meta-analysis of employee silence, 168 samples and 63,134 employees. Silence related to leader factors, dispositions and job perceptions; psychological safety accounted for a large share of the explained variance in silence. https://doi.org/10.1007/s10869-021-09788-7
📃 Morrison (2023) Annual Review of the decade's research on employee voice and silence: when and why employees speak up or stay silent, and the consequences for individuals and organisations. https://doi.org/10.1146/annurev-orgpsych-120920-054654
P.P.S. Why Do I Even Have the Nerve to Write This?
Fair question.
I am someone who has spent years working alongside remarkable executives, navigating chaos, translating vision into structure and figuring things out, sometimes beautifully and sometimes the very hard way. I have had the privilege of being the steady right hand to leaders who move fast, take risks and expect a lot. That environment has shaped me more than any classroom ever could.
I also happen to have an MBA, and I am studying psychology because people fascinate me. How we work. Why we disconnect. What actually holds us together when the pressure rises.
But really, none of that is the point.
I am here because I have lived the erosion and rebuilt from it. I have seen the cost of carrying too much and the relief that arrives when you finally stop disappearing inside your own competence.
So I share what I have learned in case it helps someone else. Take what is useful. Leave the rest. And remember this one truth that professionals like you often forget:
You are probably doing far better than you give yourself credit for.
Comments