i

Description

A school building

Source

WiinterU · original page

Licence

CC0 Public domain

Original photo, unretouched. The background is a version of it stylised by the site.

The Pygmalion effect (self-fulfilling prophecy)

The expectations we have of a person, spoken or not, sometimes end up producing what we expected.

In brief

• What: a manager's expectation of a person influences, through the manager's behavior, that person's performance (Pygmalion effect); a negative expectation produces the opposite effect (Golem effect).
• Why: we do not treat those from whom we expect a lot the same way as the others (climate, feedback, difficulty of tasks, opportunities given).
• Detect: the same people always get the rewarding assignments; “he'll never improve” proves true without ever having been tested.
• Avoid: check your expectations against facts, offer the same opportunities, set demanding but realistic goals, measure the effect of your own behavior.

On this page

1. Definition
2. The mechanism
   1. The 1968 experiment
   2. In organizations
3. Where it shows up
   1. In project management
   2. In management
4. Famous cases
   1. “Pygmalion in the Classroom”: the San Francisco school 1964-1968
   2. “Pygmalion goes to boot camp”: the Israeli army 1982
   3. A kindergarten sorted after eight days 1970
5. Warning signs
6. Avoidance reflexes
7. With AI: detect, remedy, mitigate
   1. Detect
   2. Fix
   3. Mitigate
   4. Prompts to build into your routines
   5. How to build them in
   6. Precautions
8. Limits
9. Sources and verification
10. See also
    1. Combines with, or reinforces
    2. Related biases
    3. Counterweights and related reflexes

Definition#

Fictitious illustration

Fictitious illustration

The Pygmalion effect, or Rosenthal effect, is the phenomenon by which one person's expectations of another (pupil, team member, team) change the behaviors they adopt toward that person, and thereby the other person's actual performance. The name comes from the sculptor of mythology, who fell in love with his statue, which came to life. The mechanism is that of the self-fulfilling prophecy, a notion the sociologist Robert K. Merton formalized in 1948: a false definition of the situation provokes a behavior that makes it come true.

The opposite case, the Golem effect, refers to the drop in performance caused by low expectations. Another related, more discreet case is that of trajectory maintenance: the expectation is accurate, but it prevents the person from exceeding what is expected of them.

A useful clarification: this is not a bias of judgment like anchoring, but a bias of conduct. The initial error is not only in the manager's head, it changes the world that the manager then observes, which makes it very hard to correct.

The mechanism#

Rosenthal proposed a four-factor model, based on syntheses of research on teacher expectations:

• Climate: we smile more, listen more attentively, and are warmer with those from whom we expect a lot.
• Feedback: feedback is more frequent, more precise and better substantiated; with the others, it is rarer, vaguer and harsher.
• Input: we give more material, more difficult tasks and more stimulating content to those from whom we expect a lot.
• Output opportunity: we leave them more room, ask them more questions and give them more time to answer.

On the person's side, these behaviors change their confidence, their effort and their opportunities to learn; performance follows. The manager then observes a result consistent with their expectation and feels confirmed (see confirmation bias), without seeing that it was their own behavior that produced it.

The 1968 experiment#

In 1964-1965, Robert Rosenthal (Harvard) and Lenore Jacobson, principal of an elementary school in San Francisco (referred to in the book as “Oak School”), told teachers that a new test, presented as a “test for intellectual blooming,” could identify the pupils who would experience an “intellectual spurt” during the year. About one pupil in five was designated in this way. In reality, the names were drawn at random; the test used was an ordinary intelligence test (Flanagan's Test of General Ability). The pupils were retested after about eight months.

The results reported in Pygmalion in the Classroom (1968): the IQ gains of the “designated” pupils were greater than those of the others, significantly so mainly in the first two grades of elementary school (about 27 points versus 12 in first grade, about 16.5 versus 7 in second grade). For the higher grades, the effect does not appear.

In organizations#

The world of work has produced its own studies. Livingston (1969, Harvard Business Review) drew from them a formula: “what a manager expects, and the way he treats his subordinates, largely determines their performance.” Dov Eden and Avraham Shani (1982) tested it in the Israeli army: course instructors were told, wrongly, that some trainees, drawn at random, had high command potential; these trainees obtained better objective results than the others. McNatt's meta-analysis (2000) pools workplace studies and concludes there is a clear average effect, more marked than in school studies, based on small samples to be interpreted with caution.

Where it shows up#

In project management#

• Task allocation: the “solid” developer gets the complex topics, the other the minor fix tickets. After a year, one has progressed and the other has not, which confirms the expectation.
• Code review: a reviewer who is more attentive and kinder toward profiles they respect; more nitpicking, and less pedagogical, toward the others.
• Integrating a person from a vendor or another team: the label (“they're not as good,” “they cost less”) determines which meetings they are invited to, and therefore what they can contribute.
• Estimation and commitment: if the team is judged “slow,” its estimates are trimmed without discussion, which creates overruns that “prove” its slowness.
• Acceptance testing and quality: “this module is always buggy”: it is tested more, more defects are found in it, which confirms its bad reputation.

In management#

• Onboarding: a newcomer's first impression determines the quality of support they will receive afterwards (see the primacy effect).
• Delegation: those believed capable receive trust and autonomy; the others, control, which erodes their initiative.
• Mentoring and promotions: we invest in “high potentials” identified early, sometimes on the basis of similarity (see similarity bias), and we offer them the opportunities that will make them high potentials.
• Annual review: a performance ranking set in a cold state conditions the goals, opportunities and raises of the following year.
• The whole team: a manager who describes their team in public as “mediocre” or “excellent” influences everyone's expectations, including those of the members toward themselves.

Famous cases#

Warning signs#

• “I knew it, he didn't make it”: a judgment made before the test that was never contradicted by a test.
• The same people always get the rewarding assignments, the same others the routine tasks.
• Feedback to some is precise and encouraging, to others vague or only corrective.
• The gap between “good” and “average” widens over time without any new fact explaining it.
• Team members are introduced by their label (“our star,” “the dead weight”) rather than by their achievements and areas for improvement.
• A person changes team or manager, and their performance changes with it: a sign that context mattered.
• Reviews end with expectations identical to last year's.

Avoidance reflexes#

1. Check your expectations against facts. For each person, write down what you expect and on which observable facts it rests. An expectation without facts is a label.
2. Offer the same opportunities. Make a table of how rewarding assignments, presentations, reviews and training are distributed: the inequality becomes visible and can be discussed.
3. Set demanding but realistic goals. A clear goal, slightly above the current level, accompanied by the means to reach it, is what makes the expectation useful. Overestimating without support does not help.
4. Give feedback of the same quality to everyone. Precise, focused on the facts and the actions that improve things, with the same attention, whether you are addressing a “strong” or a “weak” performer.
5. Allow time to respond. In meetings, wait for the answer from those you know little; call on those who speak less.
6. Change your view of a trajectory. Before concluding that someone lacks ability, ask yourself: “What opportunities did I give? What help did I offer?”
7. Have your own behavior observed. A peer or mentor attends meetings, reviews and interviews, and notes who speaks, who receives what, and in what tone.
8. Be careful with passed-on labels. Introduce a newcomer by their achievements rather than a ranking; do not hand a new manager someone's “profile” as if it were the truth.

With AI: detect, remedy, mitigate#

Prompts to build into your routines#

1. Review a piece of feedback or an evaluation. Use on written feedback, after removing names.

Here is feedback I am about to give a team member (names removed): [text].
1. Identify the statements that rest on no observable fact (labels, generalizations, “always,” “never”).
2. Rewrite each as a dated fact, a consequence and a clear expectation.
3. Say whether the level of demand seems consistent with what I would expect of another team member in the same situation.
Do not judge the person: work only on the text.

2. Test an expectation before concluding. When you think “he won't make it” or “she is excellent.”

I think [person, anonymized role] is [judgment]. Here are the facts I rely on: [facts].
1. Distinguish what is an observed fact from what is an interpretation.
2. Give three plausible alternative explanations of these facts that do not depend on the person's abilities.
3. Suggest a simple way to test my judgment within two weeks (a task, support, an opportunity).
4. Say how my usual behavior toward this person might contribute to the observed result.

How to build them in#

• Before a review: prompt 1 on your notes, then human review.
• When a label takes hold: prompt 2, then a concrete experiment with the person.
• In a quarterly review: an anonymized table of task distribution, reviewed with a peer.

Precautions#

• Evaluations of people are personal and sensitive data: entrust them only to a tool authorized by your organization, anonymized, and follow your company's rules and the applicable law.
• Do not let a tool rank or “score” team members: it would reproduce your labels, with an appearance of objectivity.
• The alternative explanations offered are hypotheses: test them with the person.
• Avoid turning the message into pressure (“believe in yourself!”): a useful expectation comes with means.

Limits#

• The founding study is fragile methodologically; the effect exists, but it is on average modest and varies by context (Jussim and Harber, 2005).
• Expectations are often accurate: a manager who judges a team member on past results does not necessarily create their failure. The bias begins when the expectation rests on irrelevant cues (appearance, origin, similarity, rumor).
• “Believing in someone” does not replace training or resources: high expectations without support can even demoralize.
• The name “Pygmalion effect” is sometimes used vaguely (motivation, confidence). Keep the precise mechanism in mind: the expectation changes the manager's behavior, which changes the person's.
• Workplace studies involve small samples, often in particular settings (army, training): their conclusions should be generalized with caution.

Sources and verification#

• Rosenthal & Jacobson (1968), Pygmalion in the Classroom: Teacher Expectation and Pupils' Intellectual Development, Holt, Rinehart & Winston.
• Rosenthal & Fode (1963), “The effect of experimenter bias on the performance of the albino rat,” Behavioral Science — the preliminary experiment with rats.
• Thorndike, R. L. (1968), “Review of Pygmalion in the Classroom,” American Educational Research Journal, 5(4) — the methodological critique.
• Raudenbush (1984), “Magnitude of teacher expectancy effects on pupil IQ as a function of the credibility of expectancy induction: A synthesis of findings from 18 experiments,” Journal of Educational Psychology
• Jussim & Harber (2005), “Teacher expectations and self-fulfilling prophecies: Knowns and unknowns, resolved and unresolved controversies,” Personality and Social Psychology Review
• Eden & Shani (1982), “Pygmalion goes to boot camp: Expectancy, leadership, and trainee performance,” Journal of Applied Psychology
• McNatt (2000), “Ancient Pygmalion joins contemporary management: A meta-analysis of the result,” Journal of Applied Psychology
• Livingston (1969), “Pygmalion in management,” Harvard Business Review, 47(4).
• Babad, Inbar & Rosenthal (1982), “Pygmalion, Galatea, and the Golem: Investigations of biased and unbiased teachers,” Journal of Educational Psychology
• Rist (1970), “Student social class and teacher expectations: The self-fulfilling prophecy in ghetto education,” Harvard Educational Review, 40(3).
• Merton (1948), “The self-fulfilling prophecy,” The Antioch Review, 8(2).

See also#