Role Clarity and Goal-Setting — Research and Evidence

Clarity of expectations is the first of Frank Newman’s “5 Cs” of people management, and this page is the science vein — the mechanisms and evidence behind it. It reviews what research says about goals, progress, role clarity, and performance measurement, in a neutral, sourced register. For how to put this into practice see Setting Expectations and Goals — Practices, and for the human and dollar cost of getting it wrong see Role Ambiguity and Poor Communication — Costs and Consequences.

One theme runs through the evidence below: clarity is necessary but rarely sufficient. Clarifying roles removes a drag on performance; specific goals and feedback supply the thrust; and aggressive, badly designed goals can do real harm. The confidence behind each claim varies — from the verified, hundreds-of-studies core of goal-setting theory to single-source vendor figures — so each section carries its own confidence label.

Why do specific, challenging goals beat vague "do your best" ones?

Decades of replicated experiments show that specific, difficult (but attainable) goals produce significantly higher task performance than vague “do your best” goals or no goals, because “do your best” has no external referent — a specific high goal defines what counts as success. Locke and Latham report meta-analytic effect sizes for specific-difficult versus “do your best” goals ranging from about d = .42 to .82.

Goal-setting theory was built by Edwin Locke and Gary Latham over roughly 35 years. It holds that conscious goals regulate action and that two attributes drive performance: specificity and difficulty. The core finding — replicated across hundreds of lab and field studies — is that specific, difficult (but attainable) goals beat vague “do your best” goals or no goals. As Locke and Latham (2002) put it, when people are asked to do their best they do not do so, because “do your best” has no external referent; a specific high goal defines what counts as success.

Goals work through four mechanisms: they direct attention to relevant activity, energize effort, increase persistence, and prompt the discovery of task-relevant strategies. This is one of the most replicated paradigms in organizational psychology — a verified construct.

The effect is moderated: it depends on ability (you cannot exceed your skill ceiling), commitment, feedback, and task complexity. Wood, Mento and Locke’s (1987) meta-analysis found goal effects strongest for simple tasks and weakest for complex ones (down to roughly d = .48–.67). So the existence of the effect is verified, but its size shrinks substantially as work gets more cognitively complex.

For an Ontario small or mid-sized business, “do your best” instructions waste the cheapest motivational lever available. The caveat: much foundational evidence is US-based and lab-heavy, and for genuinely novel, complex knowledge work (software, advanced manufacturing, professional services), a rigid difficult outcome goal can underperform a learning goal (“find three ways to cut defect rates”) — a refinement Locke and Latham themselves endorse.

Source: Locke & Latham, Building a Practically Useful Theory of Goal Setting and Task Motivation: A 35-Year Odyssey, American Psychologist (2002); Locke & Latham, A Theory of Goal Setting & Task Performance (book, 1990); Locke & Latham, New Directions in Goal-Setting Theory, Current Directions in Psychological Science (2006); Wood, Mento & Locke, Task Complexity as a Moderator of Goal Effects: A Meta-Analysis, Journal of Applied Psychology (1987).

Confidence: verified.

Does goal commitment and alignment matter — and do cascading goals or OKRs deliver?

The underlying principles are evidence-based, but the branded methods are not equally well supported. Goal commitment is a verified moderator of the goal–performance link, and Management by Objectives shows large meta-analytic productivity gains — but those gains are contingent on top-management commitment, and OKRs specifically remain a practitioner framework with little independent peer-reviewed evidence.

A goal only works if the person accepts and commits to it. Klein, Wesson, Hollenbeck and Alge’s (1999) meta-analysis of 83 samples established goal commitment as a key moderator of the goal–performance link — and it matters more for difficult goals, which only pay off if people do not abandon them at the first setback. The two levers for raising commitment are the goal’s importance (public commitment, leader buy-in) and self-efficacy (the belief it is achievable). This is solid, replicated science.

“Alignment” or “line of sight” is the management application. The strongest quantitative evidence is Rodgers and Hunter’s (1991) meta-analysis of Management by Objectives (which combines goal-setting, participation and feedback): across 70 studies, 68 showed productivity gains. The striking contingency: when top-management commitment was high, the average gain was 56%; when low, just 6% — a roughly nine-fold difference. Cascading goals succeed or fail on visible leadership commitment, not paperwork.

OKRs deserve honesty: they are a practitioner framework, not an independently validated intervention. They descend from MBO and goal-setting theory, so their principles inherit that evidence — but OKRs as a branded system have little rigorous, independent peer-reviewed support; systematic reviews find mostly case studies and consultancy white papers. Claims like “a McKinsey study found OKRs raise performance” trace to marketing, not peer review. Treat OKR adoption as well-supported principles in unproven packaging — directional at best for the branded method.

For an Ontario SMB, this is reassuring and cautionary: the principles (clear aligned goals, leader commitment, feedback) are evidence-based and scale down well to a 20–200-person firm where line-of-sight is naturally short. But buying an OKR platform is no substitute for the leadership commitment the MBO evidence shows is the actual active ingredient. Cited magnitudes are US/large-org; SME-specific effect sizes are not established.

Source: Klein, Wesson, Hollenbeck & Alge, Goal Commitment and the Goal-Setting Process: Conceptual Clarification and Empirical Synthesis, Journal of Applied Psychology (1999); Rodgers & Hunter, Impact of Management by Objectives on Organizational Productivity, Journal of Applied Psychology (1991); Locke & Latham, American Psychologist (2002); Surveying the Academic Literature on the Use of OKR, XIX Brazilian Symposium on Information Systems (2023) (representative of the thin OKR evidence base).

Confidence: industry-consensus for the goal-commitment and MBO evidence; directional for OKRs as a branded method.

Can goals backfire? The dark side of goal-setting

Yes. Well-evidenced research shows aggressive, specific goals can narrow focus to neglect non-goaled areas, distort risk preferences, erode cooperation, reduce intrinsic motivation, and increase unethical behaviour — especially when people fall just short and the goal carries a reward. The remedy is goal design, not abandoning goals, and the question remains a live scholarly debate.

The counterweight to goal-setting enthusiasm is itself well-evidenced. Ordóñez, Schweitzer, Galinsky and Bazerman (2009), in “Goals Gone Wild,” argue the benefits have been overstated and the harms under-examined. They identify side effects: a narrow focus that neglects non-goaled areas (you get what you measure, and only that); distorted risk preferences; eroded cooperation and culture; reduced intrinsic motivation when over-specified extrinsic goals crowd out interest; and more unethical behaviour.

The unethical-behaviour claim rests on experiment, not anecdote. Schweitzer, Ordóñez and Douma (2004) found people with specific unmet goals were significantly more likely to misrepresent their performance than people told to “do their best” — and the effect was strongest when people fell just short, and stronger when goals carried rewards. That is the mechanism behind real sales-quota and banking scandals: a sharp target, plus a near-miss, plus a reward, manufactures temptation.

Honesty requires noting this is a live debate, not a verdict against goals. Latham and Locke (2009) replied forcefully that critics leaned on anecdote and that most pitfalls were already known and manageable; Ordóñez and colleagues countered that goal-setting is “like a potent medication” needing careful prescription. The reasonable synthesis: the side effects are real and replicated under specific conditions (aggressive, narrow, single-metric, rewarded, near-miss), and the cure is goal design — balanced metrics, ethical guardrails, learning goals for complex work, and avoiding stretch targets tied to punitive stakes.

For an Ontario SMB, the risk is asymmetric: a small firm chasing one aggressive revenue or production number can do reputational or compliance damage it cannot absorb. The evidence is largely US lab and case work, so exact incidence in Canadian SMBs is unknown — but the design lesson transfers directly. Clarity must include clarity about how to win, not just what number to hit.

Source: Ordóñez, Schweitzer, Galinsky & Bazerman, Goals Gone Wild: The Systematic Side Effects of Overprescribing Goal Setting, Academy of Management Perspectives (2009); Schweitzer, Ordóñez & Douma, Goal Setting as a Motivator of Unethical Behavior, Academy of Management Journal (2004); Latham & Locke, Has Goal Setting Gone Wild, or Have Its Attackers Abandoned Good Scholarship?, Academy of Management Perspectives (2009).

Confidence: industry-consensus, with the side effects established under specific conditions and the overall question still a live scholarly debate.

Why does knowing where you stand (progress and expectation feedback) matter for clarity?

Goals deliver their full effect only when paired with feedback, and a large diary study shows that making progress in meaningful work is the single biggest day-to-day booster of motivation and performance. This section is about clarity-as-progress — knowing the target and how far along you are — not the interpersonal craft of delivering corrective feedback, which is covered in the feedback-and-performance research.

Within goal-setting theory, feedback is a necessary condition, not optional garnish. Locke and Latham (2002) are explicit that goals and feedback work interactively: a goal says what to aim at; feedback (knowledge of results) says whether to adjust effort or strategy. Neither works as well alone — a goal without progress information leaves people unable to calibrate, and feedback without a goal lacks a referent.

The complementary evidence is Teresa Amabile and Steven Kramer’s The Progress Principle (2011), built on a large diary study: nearly 12,000 daily diary entries from 238 employees across 26 project teams in 7 companies. Their central finding is that of all the events that brighten “inner work life,” the most powerful is making progress in meaningful work — even small wins. A positive loop runs progress → better inner work life → higher performance → more progress; setbacks run it in reverse. Tellingly, the managers studied ranked “progress” last among the motivators they thought mattered — a documented blind spot. The book is practitioner-facing but rests on a serious longitudinal dataset, so confidence is industry-consensus, with the caveat that the diary method is correlational and the sample is white-collar.

For an Ontario SMB, the implication is cheap and high-leverage: pair every clear goal with a visible, regular progress signal — a weekly number, a check-in, a simple dashboard — and frame the work as meaningful. Small firms have an advantage: leaders can see and acknowledge progress directly. Clarity is not just stating the target; it is letting people see their progress toward it.

Source: Locke & Latham, Building a Practically Useful Theory of Goal Setting and Task Motivation, American Psychologist (2002); Amabile & Kramer, The Progress Principle, Harvard Business Review Press (2011).

Confidence: industry-consensus; the diary evidence is correlational and the sample is white-collar.

Does role clarity reduce stress and improve performance — and what does role ambiguity cost?

Role ambiguity is reliably associated with lower job performance and higher strain, but the performance correlation is modest (about r = −.21) and moderated by job type and who rates performance; role conflict, by contrast, mainly harms wellbeing rather than output.

“Role ambiguity” — lacking clear information about what is expected, how to achieve it, and what success looks like — was operationalized by Rizzo, House and Lirtzman (1970), whose scale still dominates 55 years later. This section covers the mechanism linking ambiguity to outcomes; the human and dollar cost of confusion is handled in the costs-and-consequences material.

The construct is well established, and the performance effect is real but modest. Tubré and Collins’s (2000) meta-analysis — the methodologically strongest synthesis — found a role-ambiguity-to-performance correlation of r = −.21, moderated by job type and rating source (self-ratings show stronger relationships than supervisor or objective ratings, hinting at common-method inflation). In the same study, role conflict was essentially unrelated to performance (r ≈ −.07). The clean reading: ambiguity about expectations modestly drags performance; conflicting demands mostly harm wellbeing rather than output.

On strain, Örtqvist and Wincent’s (2006) meta-analysis (~300 articles) tied role ambiguity to higher tension and burnout (emotional exhaustion, depersonalization, reduced personal accomplishment) and to lower satisfaction and commitment — it is the strongest predictor of the reduced personal accomplishment component of burnout.

Two caveats matter. Magnitudes are modest, so clarifying roles is necessary but not sufficient — it removes a drag rather than supplying thrust (goals and feedback supply the thrust). And much evidence is cross-sectional and self-report, so causal claims should be tempered.

For an Ontario SMB, a relevant Canadian signal: Statistics Canada’s 2018 job-quality study found small-firm workers (under 20 employees) markedly over-represented in lower-quality jobs, and only 58.5% of workers overall had a formal performance assessment in the past year — role-clarity infrastructure is often weakest in the smallest firms. The fix should target clarity of tasks and standards, not just titles.

Source: Tubré & Collins, Jackson and Schuler (1985) Revisited: A Meta-Analysis of Role Ambiguity, Role Conflict, and Job Performance, Journal of Management (2000); Rizzo, House & Lirtzman, Role Conflict and Ambiguity in Complex Organizations, Administrative Science Quarterly (1970); Örtqvist & Wincent, Prominent Consequences of Role Stress: A Meta-Analytic Review, International Journal of Stress Management (2006).

Confidence: industry-consensus; magnitudes are modest and much evidence is cross-sectional and self-report.

Do performance reviews and ratings actually improve performance?

The evidence that any particular formal performance-management apparatus — the annual review, forced ranking, numeric rating scales — improves how well employees do their jobs is surprisingly weak after a century of research. The “abolish ratings” movement is a genuinely two-sided debate, not a settled answer; what is actually evidenced is narrower than “the review” — clear expectations, specific goals, and useful feedback.

The honest state of the evidence is uncomfortable for anyone selling a performance-management (PM) “system”: after a century of research, there is surprisingly little evidence that any particular formal PM apparatus — the annual review, forced ranking, numeric rating scales — actually improves how well employees do their jobs. In their 100-year review, DeNisi and Murphy (2017) concluded that we have learned something about improving individual performance but still know very little about whether formal appraisal systems drive it, or whether individual improvements translate into firm-level results. Pulakos and O’Leary (2011) argued the formula for effective PM “remains elusive” because PM has been reduced to burdensome administrative steps disconnected from the day-to-day behaviours — clear expectations, short-term objectives, continual guidance — that actually move performance.

The ratings debate is genuinely two-sided. Adler et al. (2016) is the canonical artifact: a structured, peer-reviewed debate from a standing-room-only 2015 SIOP session, with one side (Colquitt, Murphy, Ollander-Krane) arguing to scrap ratings and the other (Adler, Campion, Grubb) arguing to keep them. The case against ratings: disappointing intervention effects, rater disagreement, weak criteria, the weak relationship between actual performance and the ratings people receive, and conflicting organizational purposes. The case for keeping them: some discomfort and a measured gap between “where you are” and “where you should be” is what stimulates behaviour change; without a defensible metric you lose the ability to make fair, comparable decisions about pay, promotion, and development. Both sides explicitly agreed that goals, regular feedback, and a development focus have value — the dispute is narrowly about whether to collect a quantitative performance index. The debate should not be collapsed into either “ratings are debunked” or “ratings work.”

The corporate “kill ratings” stories are practitioner accounts, not evidence. The widely cited redesigns — Deloitte (Buckingham & Goodall, HBR 2015), Adobe’s “Check-in” (2012), GE, Microsoft’s 2013 abandonment of “stack ranking,” Accenture — are company self-reports and press, not controlled studies. They are genuinely informative about cost and morale. Deloitte, tallying the hours spent “completing the forms, holding the meetings, and creating the ratings,” found it “consumed close to 2 million hours a year.” Adobe estimated that its stack-ranked annual reviews required roughly 80,000 hours of its ~2,000 managers’ time each year (about 40 full-time-equivalent staff) and “saw a spike in voluntary attrition every year in the months following the review … disappointed employees leaving after receiving ratings below their expectations”; after “Check-in” launched in fall 2012, Adobe reported voluntary attrition fell about 30% (while involuntary departures rose ~50%). These stories also illustrate the specific harms of forced distribution ranking. But “we changed our system and we feel better about it” is not evidence that the new system improves performance.

The counter-evidence is real, too — removing ratings can backfire. A CEB (now Gartner) study, reported in its November 2016 release, found that “employee performance drops 10 percent when ratings are absent from the review process,” alongside a 14% fall in the quality of manager performance conversations, a 6% drop in engagement, and an 8% decline in top performers’ perception of pay-for-performance; CEB’s HR practice leader Brian Kropp attributed it to employees, lacking “the tangible symbol of a rating,” putting forth less effort and disengaging. A 2018 Gartner re-study found the performance decline had shrunk to about 4%. These are consultancy survey findings, not peer-reviewed — useful to show the debate is genuinely two-sided, not to prove either system “works.”

What is actually evidenced is narrower than “the review.” Two things have real empirical support and are distinct from the annual-review ritual: (1) specific, difficult, accepted goals beat “do your best” (goal-setting theory), and (2) useful, task-focused feedback can help — though feedback is not reliably positive and sometimes backfires (feedback science). The practical implication for the ratings debate: the payoff lives in clear expectations and good feedback conversations, which can be delivered with or without a numeric scale.

For an Ontario SMB, the implication is not to spend scarce money or political capital chasing a “best-practice” PM system, and to be sceptical of vendors who promise that buying their ratings-or-no-ratings platform will lift performance — the evidence does not support that promise. The better investment is in the cheap, evidenced fundamentals: clear role expectations, specific goals, and frequent honest feedback. Separately, in Ontario a documented, fair appraisal record has a legal value independent of whether it improves performance — it is often the central evidence in a wrongful-dismissal or human-rights dispute (see Difficult Conversations, Underperformance, and Termination).

Source: Adler, Campion, Colquitt, Grubb, Murphy, Ollander-Krane & Pulakos, Getting Rid of Performance Ratings: Genius or Folly? A Debate, Industrial and Organizational Psychology (2016); Pulakos & O’Leary, Why Is Performance Management Broken?, Industrial and Organizational Psychology (2011); DeNisi & Murphy, Performance Appraisal and Performance Management: 100 Years of Progress?, Journal of Applied Psychology (2017); Buckingham & Goodall, Reinventing Performance Management, Harvard Business Review (April 2015) (practitioner account).

Confidence: industry-consensus for the academic evidence; the corporate redesigns and CEB/Gartner figures are practitioner and consultancy accounts, not peer-reviewed.

How common is clarity of expectations in Canadian workplaces and SMBs?

There is no direct Statistics Canada measure of role clarity, but adjacent job-quality data show small-firm workers are over-represented in lower-quality jobs and less likely to receive a formal performance assessment, while Gallup reports only about half of workers strongly agree they know what is expected of them and just 21% of Canadian employees are engaged.

Statistics Canada does not publish a standalone “role clarity” or “I know what is expected of me” indicator. Its job-quality framework measures autonomy, managerial support and performance assessment, but not clarity of expectations as a discrete variable. That gap is itself worth recording.

The closest StatCan proxies come from Chen and Mehdi’s 2018 job-quality study (2016 General Social Survey, n ≈ 10,680). Roughly two-thirds to three-quarters of workers reported high autonomy, and 58.5% had a formal performance assessment in the past year. The SME-relevant finding is a firm-size gradient: among large-firm workers (100+ employees), about 70% were in the best/good job-quality classes and 23% in the worst; for small firms (under 20 employees) it was 40% best/good and 33% worst. Small-firm workers were “overrepresented in jobs that offer fewer employment benefits and a less desirable social environment.” Ontario workers specifically showed 28.4% high-quality and 31.4% poor-quality jobs. The clear-expectations infrastructure is weakest in exactly the smallest firms.

The direct “know what is expected” metric comes from Gallup, not government: its Q12 item Q01 is “I know what is expected of me at work,” and Gallup reports only about half of workers strongly agree — and that clarity of expectations has been among the largest-declining items. That figure is US/global and proprietary. The cleanest Canada-specific datapoint is engagement: Gallup’s 2025 report found 21% of Canadian employees engaged, versus 32% in the US. These are vendor figures (single-source); the StatCan data are government-grade but measure proxies rather than role clarity directly.

For an Ontario SMB, the opportunity is the point: clarity of expectations is neither measured by government nor guaranteed in Canadian small firms, making it a low-cost, under-exploited lever.

Source: Chen & Mehdi, Assessing Job Quality in Canada: A Multidimensional Approach, Statistics Canada Analytical Studies Branch (2018; 2016 General Social Survey); Gallup, State of the Global Workplace 2025 Report (Canada figures), Gallup (2025).

Confidence: single-source; the StatCan data measure proxies rather than role clarity directly, and the Gallup figures are proprietary vendor data.

---

This page is general information drawn from published research, not professional or legal advice. Effect sizes and study findings come with the caveats noted in each section — much of the foundational evidence is US-based, lab-heavy, or cross-sectional — so treat the figures as directional and confirm any specific claim against the cited source before relying on it.