i

Description

Un pneu dont le flanc est usé jusqu'aux fibres après avoir roulé sous-gonflé.

Source

Christopher Ziemnowicz (CZmarlin) · original page

Licence

PD Public domain

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

Self-serving bias

When it works, it is thanks to us; when it breaks, it is the fault of the context, the client or the team next door.

In brief

• What: successes are explained by one's own qualities, failures by external causes.
• Why: to protect self-esteem and one's image in front of others; we also look for fewer explanations when the result pleases us.
• Spotting it: a post-mortem whose list of causes contains only external factors, or “we” for success and “they” for failure.
• Avoiding it: describe the facts before the causes, look for internal and external causes in both cases, have a third party review.

On this page

1. Definition
2. The mechanism
3. Where it shows up
   1. In project management
   2. In management
4. Famous cases
   1. Ford and Firestone: each blames the other 2000-2001
   2. Letters to shareholders 1983
5. Warning signs
6. Avoidance reflexes
7. With AI: detect, remedy, mitigate
   1. Detect
   2. Fix
   3. Mitigate
   4. Prompts to build into your practice
   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#

Self-serving bias is the tendency to take credit for one's successes (talent, effort, good decisions) and to shift responsibility for one's failures onto external factors (bad luck, circumstances, other people). It is an asymmetry: the same person, faced with the opposite outcome, changes explanation.

The reference article is by Dale Miller and Michael Ross (1975), “Self-serving biases in the attribution of causality: Fact or fiction?”. It poses a question that has remained famous: is the asymmetry motivated (we want to feel good) or cognitive (we expect to succeed, so we explain failure as an accident)? The debate has continued, and both mechanisms are now considered to be present.

It differs from the fundamental attribution error, which concerns how we explain the behavior of others: here, it is the explanation of one's own results that is distorted.

The mechanism#

• Protecting self-esteem. Admitting a mistake is uncomfortable; an external cause leaves one's image intact.
• Managing one's image. In front of a committee, a client or a manager, we explain things so as to present ourselves in an acceptable light, sometimes without realizing it.
• Information asymmetry. We know our own efforts, constraints and good intentions from the inside; we only see the results of others. The unfavorable circumstances we endured are vivid in memory, while those that helped are forgotten.
• Expectations. If we expected to succeed, a failure is judged “abnormal” and therefore looks for a cause outside ourselves.

Where it shows up#

In project management#

• Post-mortem and lessons learned: the list of causes of a delay contains the client, the vendors, the “unforeseen complexity”, almost never the initial estimate or the team's own choices.
• Production incident: “the change was correct, it's the environment”; when the incident comes from the neighboring team, it is “symptomatic of a lack of rigor”.
• Relations with a vendor or a client: each attributes the slippage to the other (see the case below). The contract becomes a defensive tool.
• Sprint review: goals that are met are met “thanks to the team”, goals that are missed are missed because of “dependencies” and “interruptions”.

In management#

• Annual self-assessment: results achieved are presented as personal accomplishments; missed objectives, as consequences of a context (reorganization, budget, tools).
• Executive decisions: a good year is an effect of the strategy; a bad one is an effect of the market.
• Feedback to staff: a manager who takes credit for the team's success and blames the team for its failures turns the same bias into a habit, which they then attribute to others.

Famous cases#

Warning signs#

• An incident analysis that cites no internal cause, or no external cause for a success.
• “We” for what worked, “they” or “the context” for what didn't.
• Causes that are unverifiable (“the client didn't know what they wanted”) with no dated fact or document.
• Post-mortems where nobody has anything to change in their own practice.
• The same type of incident coming back, each time explained by a new external circumstance.
• Two teams or two organizations whose reports on the same event contradict each other, each exonerating itself.

Avoidance reflexes#

1. Facts first, causes second. Start the lessons-learned session with a dated, neutral timeline, with no judgmental verbs. Causes come afterward.
2. Two columns, in both cases. After a success as after a failure, list what was down to us (decisions, choices, avoidable mistakes) and what was down to the context (luck, help, circumstances).
3. Look for what we could have done differently. “What, within our own scope, would we change next time?” is the question to ask before discussing external factors.
4. Go beyond the first cause. The five whys method forces you past “the client changed their mind”: why hadn't we anticipated this change? Why didn't the process absorb it?
5. A third party to facilitate. The post-mortem is led by someone with no stake in the result, and it adopts a blameless framework so that acknowledging a mistake costs nothing.
6. Lead by example. The manager or sponsor names their own share first (late decision, questionable trade-off): this makes the gesture normal.
7. Verify each external cause. For each one: what fact proves it, and what action on our part could have limited its effect?

With AI: detect, remedy, mitigate#

Prompts to build into your practice#

1. Tally of internal and external causes. On a post-mortem report, anonymized.

Here is the report of a project or an incident (names removed). Make two lists:
1. the causes the text attributes to the team or its work (internal);
2. the causes it attributes to the context, the client, other teams or bad luck (external).
Then indicate, for each external cause, the cited fact that supports it, or the absence of a fact.
Conclude with a sentence on the balance between the two lists, without blaming anyone.

Report: [text]

2. Factual timeline, then questions. To prepare a blameless lessons-learned session.

Here are the raw notes of an incident or a delay. Rewrite them as a dated, neutral timeline: observable facts only, with no adjectives or judgments, and flagging what is an interpretation.
Then propose ten open questions for the lessons-learned session, of the form “how”, “what made it possible”, “what would we change within our scope”, without the question “who”.
Do not conclude on the cause: just list the leads to verify.

Notes: [text]

How to build them in#

• Before the meeting: each participant sends their notes to prompt 2, rereads the timeline and corrects the facts.
• After drafting: prompt 1 serves as a balance check, before the report is circulated.
• Keep the decision in human hands: the tool proposes a classification, the group decides what is true.

Precautions#

• AI does not establish facts: it reorganizes what it is given. Do not use it to point to a culprit.
• An already slanted account will slant the answer: give the raw notes, not your interpretation.
• Do not paste personal data, employee names or confidential information into an unauthorized tool.

Limits#

• The bias is not universal: it varies with culture, mood, personality and social context (in front of an audience, people appear more modest or more defensive).
• Attributing a failure to external causes is not always wrong: the context exists, and sometimes the fault really does lie elsewhere. The bias lies in the systematic asymmetry, not in an isolated explanation.
• Part of the asymmetry is cognitive (expectations, available information) and not defensive; it is therefore not bad faith.
• Taking all the responsibility for every failure is also a bias, and does not help learning either.

Sources and verification#

• Miller & Ross (1975), “Self-serving biases in the attribution of causality: Fact or fiction?”, Psychological Bulletin, 82(2)
• Zuckerman (1979), “Attribution of success and failure revisited, or: The motivational bias is alive and well in attribution theory”, Journal of Personality, 47(2)
• Mezulis, Abramson, Hyde & Hankin (2004), “Is there a universal positivity bias in attributions? A meta-analytic review…”, Psychological Bulletin, 130(5)
• Staw, McKechnie & Puffer (1983), “The justification of organizational performance”, Administrative Science Quarterly, 28(4)
• Bettman & Weitz (1983), “Attributions in the board room: Causal reasoning in corporate annual reports”, Administrative Science Quarterly, 28(2).
• Modern Tire Dealer, “Firestone/Ford partnership ends acrimoniously” (May 21, 2001) — letter from J. Lampe, reciprocal accusations.
• Modern Tire Dealer, “Ford continues to blame Firestone tires as it prepares for its own recall” (May 21, 2001) — tire replacement, 174 deaths and over 6,000 reports.
• EBSCO Research Starters, “Bridgestone/Firestone Announces Massive Recall” — recall of August 9, 2000.

See also#