Watched factory workers hid their most productive improvisations from observers, and reducing observability by enclosing production lines in privacy curtains raised defect-free output by as much as 10–15 percent, sustained over five months.

• supportsprimary-checkedResults, p. 22 of the HBS preprint (ASQ 57(2): 181–216)

  “Performance on each of the four lines surrounded by curtains, measured in defect-free units per hour (UPH), increased by as much as 10–15 percent after the first week and maintained a lead over the 28 control lines for the remaining five months of the experiment.”

  defect-free units per hour vs control lines: up to +10–15% (the paper's own upper-bound phrasing), lead held 5 months (n = 4 curtained lines vs 28 control line-shifts, one Chinese factory)

  Bernstein, E. (2012). The Transparency Paradox: A Role for Privacy in Organizational Learning and Operational Control. Administrative Science Quarterly, 57(2), 181-216. doi:10.1177/0001839212453028

• supportsprimary-checkedAbstract

  “maintaining observability of workers may counterintuitively reduce their performance by inducing those being observed to conceal their activities through codes and other costly means; conversely, creating zones of privacy may, under certain conditions, increase performance”

  Bernstein, E. (2012). The Transparency Paradox: A Role for Privacy in Organizational Learning and Operational Control. Administrative Science Quarterly, 57(2), 181-216. doi:10.1177/0001839212453028

• supportsprimary-checkedAbstract (via PMC6030579)

  “the volume of face-to-face interaction decreased significantly (approx. 70%)”

  face-to-face interaction after open-plan conversion: approximately −70%; email sent +56% in Study 1 (n = Study 1 N = 52, Study 2 N = 100 (two Fortune 500 HQs))

  Bernstein, E., Turban, S. (2018). The impact of the 'open' workspace on human collaboration. Philosophical Transactions of the Royal Society B, 373(1753), 20170239. doi:10.1098/rstb.2017.0239

• contextualizesprimary-checkedAbstract, opening sentence (publisher-deposited abstract via the OpenAlex record for doi:10.1177/0001839212453028, verified verbatim 2026-07-09) — grounds the essay's description of the factory's scale, replacing the seed report's unverifiable ~14,000-employee figure

  “Using data from embedded participant-observers and a field experiment at the second largest mobile phone factory in the world, located in China”

  Bernstein, E. (2012). The Transparency Paradox: A Role for Privacy in Organizational Learning and Operational Control. Administrative Science Quarterly, 57(2), 181-216. doi:10.1177/0001839212453028

Counter-evidence searched: The folk counter — the Hawthorne effect, i.e. observation raises output — is itself contested (Levitt & List 2011, encoded as clm.surveillance-gain-error.hawthorne-contested), and social-facilitation evidence shows observation can help simple well-learned tasks (aiello-kolb-1995, bond-titus-1983, encoded in the observation-gain claim). Bernstein told the night shift the curtains belonged to a day-shift project, ruling out a being-studied effect. Held at moderate per the seed report: one rigorous single-site experiment with only two treatment line-pairs, an upper-bound effect phrasing, and no at-scale replication.

Two meta-analyses by independent teams (94 and 70 samples, aggregating overlapping primary literatures) agree that electronic performance monitoring shows no average worker-performance benefit while reliably elevating stress.

• supportsprimary-checkedAbstract of the Microsoft NFW working-paper version (published as Personnel Psychology 76(1))

  “We also found that monitoring without an explicit purpose was not associated with any increases in performance. Finally, we found that regardless of monitoring characteristics, individuals tend to find EPM as stressful.”

  EPM presence → stress & strain (working-paper table, labeled as such): rc = .19 (n = k = 94 samples, N = 23,461 (published version, per seed report))

  Ravid, D., White, J., Tomczak, D., Miles, A., Behrend, T. (2023). A meta-analysis of the effects of electronic performance monitoring on work outcomes. Personnel Psychology, 76(1), 5-40. doi:10.1111/peps.12514

• supportsreport-derivedSeed report, Evidence map

  “k = 94 samples, N = 23,461: "no evidence that EPM improves worker performance," stress up regardless of monitoring characteristics”

  Ravid, D., White, J., Tomczak, D., Miles, A., Behrend, T. (2023). A meta-analysis of the effects of electronic performance monitoring on work outcomes. Personnel Psychology, 76(1), 5-40. doi:10.1111/peps.12514

• supportsprimary-checkedAbstract (publisher-deposited abstract retrieved via the OpenAlex record for doi:10.1016/j.chbr.2022.100227, 2026-07-07)

  “Results indicate that electronic monitoring slightly decreases job satisfaction, r = −0.10, and slightly increases stress, r = .11 … the current meta-analysis found no relationship between electronic monitoring and performance, r = −0.01, but a small positive relationship with CWB, r = 0.09.”

  meta-analytic correlations with electronic monitoring: performance r = −.01; stress r = .11; satisfaction r = −.10; CWB r = +.09 (n = 70 independent samples, 233 effect sizes)

  Siegel, R., König, C., Lazar, V. (2022). The impact of electronic monitoring on employees' job satisfaction, stress, performance, and counterproductive work behavior: A meta-analysis. Computers in Human Behavior Reports, 8, 100227. doi:10.1016/j.chbr.2022.100227

• contradictsprimary-checkedAbstract (publisher-deposited abstract retrieved via the OpenAlex record for doi:10.1111/peps.12046, 2026-07-07)

  “Results indicate that more frequent supervisory use of EPM is associated with better task performance and OCB.”

  Bhave, D. (2014). The Invisible Eye? Electronic Performance Monitoring and Employee Job Performance. Personnel Psychology, 67(3), 605-635. doi:10.1111/peps.12046

• contextualizesprimary-checkedAbstract, Study 1 (publisher-deposited abstract via the OpenAlex record for doi:10.1111/peps.12046, verified verbatim 2026-07-09) — the abstract names call quality metrics as the task-performance measure; it does not name calls handled

  “results reveal that shorter time lags between 2 consecutive employee performance assessments are related to better task performance as indicated by call quality metrics”

  Bhave, D. (2014). The Invisible Eye? Electronic Performance Monitoring and Employee Job Performance. Personnel Psychology, 67(3), 605-635. doi:10.1111/peps.12046

Counter-evidence searched: Searched 2026-07-07 for positive EPM-performance evidence. Found Bhave (2014): two call-center field studies where more frequent supervisory use of EPM correlated with better task performance and citizenship behavior — a within-firm frequency-of-use effect, non-experimental, and inside the primary literature the later meta-analyses aggregate. Aiello & Kolb's skilled-worker facilitation is a second positive pocket. Kept at strong because the claim is about the population-level average, on which both meta-analyses (k = 94 and k = 70, aggregating overlapping primary literatures — not additive samples) agree; the pockets are encoded in the boundary-condition claims.

Monitoring signals distrust and strips agency: most experimentally controlled agents cut discretionary effort in response, and monitored employees were more likely to break rules and cheat, with perceived fair treatment the buffer.

• supportsprimary-checkedAbstract (AEA page; the page strips the apostrophe in "principal's")

  “Our results show that control entails hidden costs since most agents reduce their performance as a response to the principals controlling decision.”

  agents reacting negatively to imposed control: 64%/57% in C5/C10 conditions; effort under trust up to 2x controlled level (per seed report / IZA dp1203) (n = N = 804, one-shot principal–agent game)

  Falk, A., Kosfeld, M. (2006). The Hidden Costs of Control. American Economic Review, 96(5), 1611-1630. doi:10.1257/aer.96.5.1611

• supportsreport-derivedSeed report, §Evidence landscape

  “Thiel and colleagues found monitored employees more likely to take unapproved breaks, work slowly, and cheat on tasks, mediated by lost sense of agency and moral disengagement — with perceived fair treatment as the buffer”

  Thiel, C., Bonner, J., Bush, J., Welsh, D., Garud, N. (2023). Stripped of Agency: The Paradoxical Effect of Employee Monitoring on Deviance. Journal of Management, 49(2), 709-740. doi:10.1177/01492063211053224

• contextualizesprimary-checkedAbstract (publisher-deposited abstract retrieved via the OpenAlex record for doi:10.1257/aer.96.5.1611, 2026-07-07); per the seed report / IZA dp1203, the control-aversion reaction also fades in the strong-control C20 condition

  “Overall, the effect of control on the principal's payoff is nonmonotonic.”

  Falk, A., Kosfeld, M. (2006). The Hidden Costs of Control. American Economic Review, 96(5), 1611-1630. doi:10.1257/aer.96.5.1611

• contextualizesprimary-checkedAbstract (publisher-deposited abstract retrieved via the OpenAlex record for doi:10.1016/j.chbr.2022.100227, 2026-07-07)

  “the current meta-analysis found no relationship between electronic monitoring and performance, r = −0.01, but a small positive relationship with CWB, r = 0.09.”

  monitoring → counterproductive work behavior: r = +.09 (n = 70 samples, 233 effect sizes)

  Siegel, R., König, C., Lazar, V. (2022). The impact of electronic monitoring on employees' job satisfaction, stress, performance, and counterproductive work behavior: A meta-analysis. Computers in Human Behavior Reports, 8, 100227. doi:10.1016/j.chbr.2022.100227

Counter-evidence not yet searched.

When stakes attach to a watched measure, the measured game it: forensic methods find cheating in at least 4–5% of Chicago classrooms annually, English NHS ambulance response-time distributions spiked at exactly the eight-minute target (with A&E waits peaking at the four-hour target, less dramatically), and hospital hand hygiene ran roughly 2.5x higher only where auditors could see.

• supportsprimary-checkedAbstract (NBER w9413 version)

  “serious cases of teacher or administrator cheating on standardized tests occur in a minimum of 4-5 percent of elementary school classrooms annually”

  classrooms with detected test-answer manipulation: ≥4–5% annually (explicit lower bound; only blatant manipulation detected) (n = Chicago Public Schools, 1993–2000)

  Jacob, B., Levitt, S. (2003). Rotten Apples: An Investigation of the Prevalence and Predictors of Teacher Cheating. Quarterly Journal of Economics, 118(3), 843-877. doi:10.1162/00335530360698441

• supportsprimary-checkedAbstract (NBER w9413 version)

  “the observed frequency of cheating appears to respond strongly to relatively minor changes in incentives”

  Jacob, B., Levitt, S. (2003). Rotten Apples: An Investigation of the Prevalence and Predictors of Teacher Cheating. Quarterly Journal of Economics, 118(3), 843-877. doi:10.1162/00335530360698441

• supportsprimary-checkedAbstract, Conclusions (via PubMed 25002555)

  “This is consistent with the existence of a Hawthorne effect localised to areas where the auditor is visible and calls into question the accuracy of publicly reported hospital hand hygiene compliance rates.”

  hand-hygiene events per dispenser-hour, auditor visible vs not: 3.75 vs 1.48 (p = 0.001); no change inside rooms auditors could not see (n = 780,777 dispensing events, single hospital)

  Srigley, J., Furness, C., Baker, G., Gardam, M. (2014). Quantification of the Hawthorne effect in hand hygiene compliance monitoring using an electronic monitoring system: a retrospective cohort study. BMJ Quality & Safety, 23(12), 974-980. doi:10.1136/bmjqs-2014-003080

• supportsprimary-checkedpp. 530–531 and Figure 7 (full text, published-article PDF fetched 2026-07-07)

  “found evidence that in a third of ambulance trusts, response times had been 'corrected' to be reported to be less than eight minutes … a 'corrected' pattern with a curious 'spike' at 8 minutes – with the strong implication that times between 8 and 9 minutes have been reclassified to be less than 8 minutes”

  ambulance trusts with 'corrected' Category A response times: one-third of trusts; frequency distribution shows a sharp spike at exactly 8 minutes (n = Commission for Health Improvement audit evidence, England)

  Bevan, G., Hood, C. (2006). What's measured is what matters: Targets and gaming in the English public health care system. Public Administration, 84(3), 517-538. doi:10.1111/j.1467-9299.2006.00600.x

• supportsprimary-checkedp. 531 (full text, published-article PDF fetched 2026-07-07)

  “a study of the distribution of waiting times in A&E found frequency peaked at the four-hour target (Locker and Mason 2005) – although this pattern was much less dramatic than that for ambulance response times”

  Bevan, G., Hood, C. (2006). What's measured is what matters: Targets and gaming in the English public health care system. Public Administration, 84(3), 517-538. doi:10.1111/j.1467-9299.2006.00600.x

• supportsprimary-checkedpp. 528–529 (full text, published-article PDF fetched 2026-07-07)

  “in 2002/03, officially, in 139 out of 158 acute trusts 90 per cent of patients were seen in less than four hours, but only 69 per cent of patients reported that experience in the survey”

  Bevan, G., Hood, C. (2006). What's measured is what matters: Targets and gaming in the English public health care system. Public Administration, 84(3), 517-538. doi:10.1111/j.1467-9299.2006.00600.x

• contextualizesprimary-checkedAbstract (NBER w15531)

  “NCLB generated statistically significant increases in the average math performance of 4th graders (effect size = 0.22 by 2007)”

  4th-grade NAEP math under high-stakes accountability: +0.22 SD by 2007; no reading gains in either grade (n = state-level NAEP panel, comparative interrupted time series)

  Dee, T., Jacob, B. (2009). The Impact of No Child Left Behind on Student Achievement. NBER Working Paper 15531. doi:10.3386/w15531not peer-reviewed

Counter-evidence searched: Searched 2026-07-07 for evidence that high-stakes measurement produces genuine rather than gamed improvement. Dee & Jacob's NCLB analysis finds real 4th-grade math gains (+0.22 SD) alongside gaming documented in the same style of high-stakes regime — note the scope: Jacob & Levitt's cheating window is Chicago's own 1993–2000 accountability regime, while Dee & Jacob measure the later national NCLB program (2002–2007), so the two results share a regime style, not a system. Measurement-with-stakes corrupts the signal at the margin without necessarily faking the whole trend; the claim text is scoped to gaming of the measure, not to a claim that all measured improvement is fake. Jacob & Levitt's 4–5% is a lower bound of blatant manipulation only.

In large vendor surveys of hybrid work, 87% of employees report being productive while only 12% of leaders express full confidence in it, and employees under surveillance tools admit to two to three times more performative 'productivity theater' than unmonitored peers.

• supportsprimary-checkedWorkLab special report (2022-09-22), definition callout

  “Productivity paranoia: where leaders fear that lost productivity is due to employees not working, even though hours worked, number of meetings, and other activity metrics have increased.”

  employee-reported productivity vs leader full confidence: 87% vs 12% (85% of leaders say hybrid makes confidence in productivity challenging) (n = 20,006 knowledge workers, 11 countries (Edelman DxI survey))

  Microsoft (2022). Hybrid Work Is Just Work. Are We Doing It Wrong? (Work Trend Index Special Report). Microsoft Work Trend Index. linknot peer-reviewed

• supportsprimary-checkedBlog write-up of the February 2023 survey

  “Eighty-three percent of respondents admitted to engaging in at least one of the most common performative work behaviors”

  employees admitting performative work: 83% (n = 1,000 US full-time employees)

  Visier (2023). Half of employees say they engage in 'productivity theater,' new survey shows. Visier blog. linknot peer-reviewed

• supportsprimary-checkedBlog write-up of the February 2023 survey

  “Employees with employers who use surveillance tools were also more than twice (and in some cases three times) as likely to commit the most egregious performative behaviors”

  Visier (2023). Half of employees say they engage in 'productivity theater,' new survey shows. Visier blog. linknot peer-reviewed

Counter-evidence not yet searched.

In randomized trials at Ctrip/Trip.com, replacing presence-based oversight with metered outcomes left performance intact or better — home working raised measured performance 13% and roughly halved attrition, and hybrid cut quits by one-third with null performance effects — while managers' presence-based productivity estimates proved miscalibrated.

• supportsprimary-checkedAbstract (QJE 130(1), checked via NBER w18871 page and RePEc/IDEAS mirror)

  “Home working led to a 13% performance increase, of which 9% was from working more minutes per shift (fewer breaks and sick days) and 4% from more calls per minute”

  call-center performance under WFH (phone-system metered): +13% (9% more minutes per shift + 4% more calls per minute); attrition halved and gains 'almost doubling to 22%' after opt-in reselection (abstract); the 35%→17% attrition figures are per seed report (n = 249 randomized by birthdate (131 T / 118 C), from 503 volunteers of 996 asked; 9 months)

  Bloom, N., Liang, J., Roberts, J., Ying, Z. (2015). Does Working from Home Work? Evidence from a Chinese Experiment. Quarterly Journal of Economics, 130(1), 165-218. doi:10.1093/qje/qju032

• contradictsprimary-checkedAbstract (QJE 130(1); confirmed verbatim via the publisher-deposited OpenAlex abstract record for doi:10.1093/qje/qju032, 2026-07-07)

  “Home workers also reported improved work satisfaction, and their attrition rate halved, but their promotion rate conditional on performance fell.”

  promotion rate conditional on performance, home workers vs office: fell (direction only in the abstract; magnitude in the paywalled body) (n = same RCT: 249 randomized (131 T / 118 C), 9 months)

  Bloom, N., Liang, J., Roberts, J., Ying, Z. (2015). Does Working from Home Work? Evidence from a Chinese Experiment. Quarterly Journal of Economics, 130(1), 165-218. doi:10.1093/qje/qju032

• supportsprimary-checkedAbstract (via PubMed 38867040)

  “hybrid working improved job satisfaction and reduced quit rates by one-third”

  quit rate under hybrid WFH: −1/3; manager productivity estimates swung from −2.6% pre-experiment to +1.0% after (abstract-confirmed) (n = N = 1,612, six-month RCT with two-year follow-up)

  Bloom, N., Han, R., Liang, J. (2024). Hybrid working from home improves retention without damaging performance. Nature, 630, 920-925. doi:10.1038/s41586-024-07500-2

• supportsprimary-checkedAbstract (via PubMed 38867040)

  “Null equivalence tests showed that hybrid working did not affect performance grades over the next two years of reviews.”

  Bloom, N., Han, R., Liang, J. (2024). Hybrid working from home improves retention without damaging performance. Nature, 630, 920-925. doi:10.1038/s41586-024-07500-2

• supportsprimary-checkedAbstract (confirmed verbatim via the publisher-deposited OpenAlex abstract record for doi:10.1038/s41586-024-07500-2, 2026-07-07)

  “We also found that the 395 managers in the experiment revised their surveyed views about the effect of hybrid working on productivity, from a perceived negative effect (−2.6% on average) before the experiment to a perceived positive one (+1.0%) after the experiment.”

  manager-estimated productivity effect of hybrid WFH, before vs after outcome data: −2.6% → +1.0% (n = 395 managers in the N = 1,612 RCT)

  Bloom, N., Han, R., Liang, J. (2024). Hybrid working from home improves retention without damaging performance. Nature, 630, 920-925. doi:10.1038/s41586-024-07500-2

• contextualizesprimary-checkedAbstract (AEA page)

  “Before COVID-19, remote workers answered 12 percent fewer calls per hour than on-site workers.”

  remote vs on-site calls per hour: −12% pre-COVID; an 8% gap persisted when everyone went remote, indicating negative selection into remote jobs (n = Fortune 500 call centers)

  Emanuel, N., Harrington, E. (2024). Working Remotely? Selection, Treatment, and the Market for Remote Work. American Economic Journal: Applied Economics, 16(4), 528-559. doi:10.1257/app.20230376

• contradictsprimary-checkedAbstract (IZA DP 14336)

  “Total hours worked increased by roughly 30%, including a rise of 18% in working after normal business hours. Average output did not significantly change. Therefore, productivity fell by about 20%.”

  output per hour under pandemic-forced WFH: −~20% (hours +30%, output flat) (n = 10,000+ IT professionals, one large Asian firm, non-randomized)

  Gibbs, M., Mengel, F., Siemroth, C. (2021). Work from Home & Productivity: Evidence from Personnel & Analytics Data on IT Professionals. IZA Discussion Paper 14336. linknot peer-reviewed

Counter-evidence searched: Searched 2026-07-07 for remote-work productivity declines. Gibbs, Mengel & Siemroth (pandemic-forced WFH at an Asian IT services firm, non-randomized, coordination-heavy): productivity per hour fell ~20%. Emanuel & Harrington: remote call-center jobs attract negatively selected applicants. Both bound generalization rather than refute the RCTs: the claim is scoped to randomized, volunteer, outcome-meterable settings — which is exactly the substitution (outcomes for presence) the claim describes. The 2015 promotion-rate penalty conditional on performance is also real and belongs in any honest telling — it is now encoded as a primary-checked contradicts entry (the abstract carries the direction; the magnitude sits in the paywalled body).

Monitoring can be modeled as a gain applied to the sensor path of the management control loop, with legibility bias and gaming noise growing in the gain so that the signal-to-noise ratio of what managers observe falls past a threshold.

• supportsreport-derivedSeed report, Key sources table

  “Presence of others speeds simple performance, impairs complex accuracy; effects small (0.3–3% of variance)”

  social facilitation by task complexity: positive for simple/well-learned, negative for complex; 0.3–3% of variance (per seed report) (n = 241 studies, ~24,000 participants)

  Bond, C., Titus, L. (1983). Social facilitation: A meta-analysis of 241 studies. Psychological Bulletin, 94(2), 265-292. doi:10.1037/0033-2909.94.2.265

• supportsprimary-checkedAbstract (via PubMed 7797458); the less-skilled pattern reversed, and monitored individuals reported the most stress

  “Highly skilled monitored participants keyed more entries than highly skilled nonmonitored participants”

  Aiello, J., Kolb, K. (1995). Electronic performance monitoring and social context: Impact on productivity and stress. Journal of Applied Psychology, 80(3), 339-353. doi:10.1037/0021-9010.80.3.339

• supportsreport-derivedSeed report, §Quantitative representation

  “The worker side is a multitask allocation in the Holmström–Milgrom tradition. Total effort E = e_v + e_ℓ splits between true value e_v and legible performance e_ℓ; rewards attach to the monitored signal”

  Holmström, B., Milgrom, P. (1991). Multitask Principal-Agent Analyses: Incentive Contracts, Asset Ownership, and Job Design. Journal of Law, Economics, and Organization, 7 (special issue). doi:10.1093/jleo/7.special_issue.24

Counter-evidence not yet searched.

Monitoring improves outcomes when the monitored variable is the actual objective and hard to fake: theft-monitoring IT raised restaurant revenue, camera-verified attendance pay halved rural teacher absence and raised test scores 0.17 SD, and audit floors deter a tempted minority in piece-rate work.

• supportsprimary-checkedAbstract (AEA page)

  “Teachers' attendance was monitored daily using cameras, and their salaries were made a nonlinear function of attendance.”

  Duflo, E., Hanna, R., Ryan, S. (2012). Incentives Work: Getting Teachers to Come to School. American Economic Review, 102(4), 1241-1278. doi:10.1257/aer.102.4.1241

• supportsprimary-checkedAbstract (AEA page)

  “Teacher absenteeism in the treatment group fell by 21 percentage points relative to the control group”

  teacher absence / child test scores: absence 42%→21% (−21 pp); test scores +0.17 SD; ~$2.20 per extra teaching day (per seed report) (n = 113 rural Indian NGO schools (57 treatment))

  Duflo, E., Hanna, R., Ryan, S. (2012). Incentives Work: Getting Teachers to Come to School. American Economic Review, 102(4), 1241-1278. doi:10.1257/aer.102.4.1241

• supportsprimary-checkedAbstract (via RePEc/IDEAS mirror); study uses theft and sales data from 392 restaurant locations at five firms

  “significant treatment effects in reduced theft and improved productivity”

  weekly revenue per location after theft-monitoring adoption: +$2,982 (~7%), driven by redirected effort rather than firings (per seed report; figure sits in the paywalled body, not the abstract) (n = 392 restaurants, 5 firms, diff-in-diff)

  Pierce, L., Snow, D., McAfee, A. (2015). Cleaning House: The Impact of Information Technology Monitoring on Employee Theft and Productivity. Management Science, 61(10), 2299-2319. doi:10.1287/mnsc.2014.1927

• supportsprimary-checkedAbstract (AEA page)

  “A significant fraction of employees behave as the 'rational cheater model' predicts. A substantial proportion of employees, however, do not respond to manipulations in the monitoring rate.”

  fake pledges when perceived audit rate dropped to zero: ~2.5x against a mean rate of 0.02; median employee never cheated (per seed report) (n = 16 call centers, 14-week field experiment)

  Nagin, D., Rebitzer, J., Sanders, S., Taylor, L. (2002). Monitoring, Motivation, and Management: The Determinants of Opportunistic Behavior in a Field Experiment. American Economic Review, 92(4), 850-873. doi:10.1257/00028280260344498

• contextualizesprimary-checkedAbstract (AER 93(4); confirmed verbatim via the publisher-deposited OpenAlex abstract record for doi:10.1257/000282803769206322, 2026-07-07)

  “Estimates for 1997 indicate that advanced on-board computers (OBCs) have increased capacity utilization among adopting trucks by 13 percent.”

  capacity utilization among trucks adopting advanced OBCs: +13% (1997 estimates); industry-wide +3% (n = US trucking industry, 1992/1997 survey data)

  Hubbard, T. (2003). Information, Decisions, and Productivity: On-Board Computers and Capacity Utilization in Trucking. American Economic Review, 93(4), 1328-1353. doi:10.1257/000282803769206322

• contextualizesreport-derivedSeed report, §Evidence landscape (the dispatch-coordination attribution sits in the paper's body, not the abstract)

  “On-board computers raised trucking capacity utilization ~13% among adopters — though Hubbard attributes most of that to better dispatch coordination, not discipline”

  Hubbard, T. (2003). Information, Decisions, and Productivity: On-Board Computers and Capacity Utilization in Trucking. American Economic Review, 93(4), 1328-1353. doi:10.1257/000282803769206322

Counter-evidence searched: This claim is itself the strongest counter-weight to the transparency-paradox thesis, so the search ran the other way: looked for evidence that monitoring backfires even in simple verifiable work. Falk & Kosfeld's hidden-costs-of-control result and Nagin's own finding that the median worker never cheated (cheating concentrated among workers who felt unfairly treated) bound the mechanism to deterring a tempted tail, not motivating the median. The boundary is coherent across all four studies: terminal, ungameable metrics with genuine misconduct risk. The hardest case for the boundary is Bhave (2014), encoded under the population-null claim: identified, evaluative supervisory EPM over semi-gameable call metrics with a positive performance association — non-experimental, but it resists the clean quadrant story and is plotted as such in the essay's matrix figure.

High-resolution observation coexists with — and can improve — safety and performance when it is de-identified, non-punitive, or reciprocal, as in aviation FOQA data legally barred from enforcement use, surgical checklist monitoring, and mutual customer-worker visibility.

• supportsprimary-checked14 CFR §13.401 (via Cornell LII); protection excludes criminal or deliberate acts

  “the Administrator will not use an operator's FOQA data or aggregate FOQA data in an enforcement action against that operator or its employees when such FOQA data or aggregate FOQA data is obtained from a FOQA program that is approved by the Administrator”

  Federal Aviation Administration (2001). 14 CFR § 13.401 — Flight Operational Quality Assurance Program: Prohibition against use of data for enforcement purposes. Code of Federal Regulations, Title 14. linknot peer-reviewed

• supportsprimary-checkedAbstract (via PubMed 19144931)

  “The rate of death was 1.5% before the checklist was introduced and declined to 0.8% afterward (P=0.003).”

  surgical deaths / inpatient complications: deaths 1.5%→0.8%; complications 11.0%→7.0% (P<0.001) (n = 7,688 patients, 8 hospitals, before-after)

  Haynes, A., Weiser, T., Berry, W., Gawande, A. (2009). A Surgical Safety Checklist to Reduce Morbidity and Mortality in a Global Population. New England Journal of Medicine, 360(5), 491-499. doi:10.1056/NEJMsa0810119

• supportsprimary-checkedAbstract (Methods and Results; open NEJM abstract via PubMed 17192537, re-verified verbatim 2026-07-09)

  “A total of 108 ICUs agreed to participate in the study, and 103 reported data. [...] The median rate of catheter-related bloodstream infection per 1000 catheter-days decreased from 2.7 infections at baseline to 0 at 3 months after implementation of the study intervention”

  Pronovost, P., Needham, D., Berenholtz, S. (2006). An Intervention to Decrease Catheter-Related Bloodstream Infections in the ICU. New England Journal of Medicine, 355(26), 2725-2732. doi:10.1056/NEJMoa061115

• supportsprimary-checkedAbstract (publisher-deposited abstract retrieved via the OpenAlex record for doi:10.1287/mnsc.2015.2411, re-verified verbatim 2026-07-09)

  “The introduction of this transparency contributed to a 22.2% increase in customer-reported quality and reduced throughput times by 19.2%.”

  Buell, R., Kim, T., Tsay, C. (2017). Creating Reciprocal Value Through Operational Transparency. Management Science, 63(6), 1673-1695. doi:10.1287/mnsc.2015.2411

• contextualizesreport-derivedSeed report, §Boundary conditions

  “GitHub's activity traces let 24 interviewed developers coordinate large collaborations, judge competence, and learn conventions without any boss watching”

  Dabbish, L., Stuart, C., Tsay, J., Herbsleb, J. (2012). Social coding in GitHub: transparency and collaboration in an open software repository. Proceedings of CSCW '12, 1277-1286. doi:10.1145/2145204.2145396

• contradictsprimary-checkedAbstract, Conclusions (via PubMed 24620866)

  “Implementation of surgical safety checklists in Ontario, Canada, was not associated with significant reductions in operative mortality or complications.”

  operative mortality/complications after mandated checklist adoption: no significant reduction (n = 101 Ontario hospitals, 215,000+ procedures)

  Urbach, D., Govindarajan, A., Saskin, R., Wilton, A., Baxter, N. (2014). Introduction of Surgical Safety Checklists in Ontario, Canada. New England Journal of Medicine, 370(11), 1029-1038. doi:10.1056/NEJMsa1308261

Counter-evidence searched: Searched 2026-07-07 for checklist-null results and found Urbach et al. (2014): population-wide mandated checklist introduction in Ontario produced no significant mortality or complication reduction — suggesting the Haynes/Pronovost effects depend on implementation intensity and voluntary engagement, not the observation artifact alone. Aviation attribution is likewise program-level (the 83% fatality-risk decline the report cites is ecosystem-wide). Held at moderate: the design lesson (decouple observation from individual punishment) is consistent across domains, but causal attribution to the non-punitive design itself is untested.

The canonical Hawthorne-effect story — that being watched raises output — is not supported by the rediscovered original illumination data, in which the celebrated patterns prove fictional and only subtle observation effects remain.

• supportsprimary-checkedAbstract (AEA page)

  “Existing descriptions of supposedly remarkable data patterns prove to be entirely fictional.”

  Levitt, S., List, J. (2011). Was There Really a Hawthorne Effect at the Hawthorne Plant? An Analysis of the Original Illumination Experiments. American Economic Journal: Applied Economics, 3(1), 224-238. doi:10.1257/app.3.1.224

• contextualizesprimary-checkedAbstract — Bernstein frames his qualitative findings as a reverse Hawthorne effect

  “maintaining observability of workers may counterintuitively reduce their performance by inducing those being observed to conceal their activities through codes and other costly means; conversely, creating zones of privacy may, under certain conditions, increase performance”

  Bernstein, E. (2012). The Transparency Paradox: A Role for Privacy in Organizational Learning and Operational Control. Administrative Science Quarterly, 57(2), 181-216. doi:10.1177/0001839212453028

Counter-evidence not yet searched.