Why does the same knowledge land differently?
Early-career work often depends on someone else understanding, trusting, questioning or acting on your contribution. This essay follows those interactions instead of treating “soft skills” as a single trait.
Making room for an answer
Jenn’s October 13 notes identify a wish to communicate with warmth and structure, and an experiment in asking one question without planning the next while the other person answers.
Evidence of an intention and chosen practice. The resulting behavior change remains to be demonstrated.
Three outcomes. Three different explanations.
Choose a view. Each sequence is a proposed mechanism, with evidence and a limit.
Make expertise usable
A technically correct answer can still fail to help if a colleague cannot understand the decision or knows something the analyst never asked about. This is the proposed mechanism to test.
Observe: Can the colleague accurately restate the recommendation, its limit and the next action?
Related evidence: Adhvaryu, Kala & Nyshadham · JPE, 2023 · A bundled program in one labor-market setting. Neither an isolated listening effect nor a forecast for MBA wages. Figures reported in the published abstract.
What could change in one conversation?
A junior analyst can have a correct answer to the wrong version of a colleague's problem. Listening matters when it reveals the deadline, constraint or disagreement that the analysis needs to address.
The example below is hypothetical. It illustrates the proposed mechanism; it is not an anecdote about Jenn or an observed treatment effect.
A careful analysis
An analyst has compared two options and arrives ready to explain the recommendation.
What is still unknown: which decision the colleague needs to make today.
A constraint surfaces
The analyst asks what would make the recommendation usable. The colleague explains that approval must happen before Friday.
What changed: the available decision window.
A different next step
The analyst checks whether the preferred option can meet that deadline and revises the recommendation if it cannot.
Evidence to seek: a documented change in the work, not just a pleasant exchange.
Would a smoother conversation prove the skill worked?
It would be a useful observation, but the colleague might simply be polite or already familiar with the topic. Stronger evidence would name the information that surfaced and show how it changed a decision, reduced a misunderstanding or improved the handoff. A single interaction still cannot establish a career effect.
Experience needs a feedback check
Soyer and Hogarth distinguish environments where feedback is timely and accurate from those where important outcomes are hidden or delayed. Repetition can strengthen the wrong lesson when the evidence is incomplete.
Misra's account of reworking a flawed paper adds a concrete example: criticism became useful after he investigated the error and changed the work. The story illustrates a process; it does not estimate the average return to humility.
Course readings: Soyer & Hogarth, The Myth of Experience, Introduction (2020); Sanjog Misra, The Incredible Power of Being Wrong (2025).
A small experiment to carry forward
- Choose one upcoming interaction and one question that could reveal a missing constraint.
- Record what the answer added and whether the work changed.
- Name another explanation for the response.
- Adjust the next attempt using the observation.
Adapted from the course's Observe → Experiment → Reflect → Apply loop. This practice prompt is analyst-proposed.
Behavior → information → changed work → career outcome
A useful question may reveal information. A changed decision needs its own evidence. Promotion and pay also depend on access, evaluation and opportunity.Evidence that changes the story
These quantities have different units and designs. They belong beside one another as evidence, never added into one “leadership impact” score.
Training can change output
+13.5% productivityThe published study also reports a 256% net return to the firm eight months after completion. Wages and retention did not change; promotion probability increased.
What this does and does not establish
A bundled program in one labor-market setting. Neither an isolated listening effect nor a forecast for MBA wages. Figures are those reported in the published abstract.
Listening is associated with performance
r = .21, 95% CI [.13, .29]The full review covers 664 effect sizes and 400,020 observations. Removing self-report-only data reduces the job-performance association from the overall performance estimate of r = .36.
What this does and does not establish
Primarily observational; the large total is observations across the review, not necessarily unique workers in this subset. Correlation is not a percentage gain or training effect.
Analytical and social skills work together
Nearly +12 percentage pointsHigh-social-interaction occupations grew as a share of employment. Growth in employment and wages was especially strong in jobs demanding both analytical and social skills.
What this does and does not establish
Historical labor-market pattern, not a present-day forecast or causal estimate of taking a leadership course.
Teams need usable candor
51 work teamsPsychological safety was associated with learning behavior; the analysis connected learning behavior with team performance.
What this does and does not establish
Team-level observational evidence. A warm exchange does not establish psychological safety or prove a causal performance gain.
Feedback can make performance worse
More than one-third negativeAverage performance effect was positive (d = .41), but more than one-third of interventions reduced performance.
What this does and does not establish
Do not equate observations with unique people or studies. Content and attention matter; more feedback is not automatically better. Original 1996 abstract and authors’ 1998 review checked.
Skills need to be visible to employers
A distinct matching channelGiving both sides credible information about non-cognitive skills changed matching and increased earnings conditional on employment.
What this does and does not establish
Certification, not skill training. No universal wage premium inferred; no numerical wage effect is claimed here.