Big Assumptions in Immunity to Change: A Coaching Guide

Big assumptions in Immunity to Change are beliefs a person treats as true that make a competing commitment feel necessary. Coaching helps the client name the belief and investigate its conditions. The purpose is better information and a wider choice of action, not persuading someone that every concern is irrational.

Begin with an improvement the client wants, connect it to observable behavior and protection, then investigate a specific prediction. This guide develops all four map areas, a five-step experiment checklist and an original worked case with genuine criticism, useful learning and a later recurrence.

Understand the map before changing behavior

Robert Kegan and Lisa Laskow Lahey developed Immunity to Change. Minds at Work, their developer organization, describes four areas: an improvement goal, actions that obstruct it, competing commitments and the assumptions sustaining them. Its subject-to-object language concerns gaining perspective on something previously taken for granted, rather than assigning a developmental score.

In their Harvard Graduate School of Education interview, the authors distinguish a worry from the active protection that can sustain conflicting commitments. They also explain that their immune-system language does not describe a disease. Here it is a reflective model, not a medical or neurological diagnosis or proof of why every person resists change.

Suppose a manager wants colleagues to prepare updates while repeatedly rewriting them. “Asking for help proves I am incapable” might become a question she can examine. But workload, missing expertise and unclear authority may also explain taking over. The coach checks those possibilities rather than beginning with a hidden-motive conclusion.

A client can decline this approach or leave a map unfinished. A sponsor’s frustration does not establish the client’s improvement goal. Ask what the person wants to do differently and whether exploring this pattern would serve that purpose. Do not require a personally revealing explanation as the price of discussing an ordinary work problem.

Build all four columns with the client

The following people, conversations and observations are fictional. All times are Europe/London. Lena is an operations manager. Her experienced colleague Nell prepares routine project updates; director Gideon retains spending authority, while Lena owns operational decisions and escalation. Mira is Lena’s coach. Lena chooses the goal below because rewriting everyone’s updates displaces her own decision work.

The worry box is a teaching row between the second and third columns, not a fifth column. The worry describes a feared experience; the competing commitment names the protection Lena recognizes herself pursuing. Mira asks whether that distinction fits. Lena changes “avoid criticism” to “prevent anyone questioning my command” because the latter better describes her pattern.

Then check the connection: if Lena were certain that an unanswered question would force preparation back onto her, would taking over make sense? That helps examine the map’s coherence without proving the belief true. The worry-box guide develops this transition; do not rush it merely to finish the table.

The fourth entry contains several links: Nell presents, a question remains unanswered, Gideon judges Lena and preparation ownership changes. Ask which links can actually be observed. A director asking a hard question is not direct evidence of an unspoken judgment. A written change in preparation responsibility would establish something more specific.

“I lack confidence” is too broad for this test. “One unanswered routine question will lead Gideon to take preparation away from Nell” gives Lena a condition and consequence to investigate. It is still her formulation. Mira does not strengthen it into dismissal or permanent reputational damage that Lena never predicted.

Use the authors’ SMART criteria accurately

The authorized HarvardX excerpts of Immunity to Change group S-M as safe and modest, A as actionable soon, and R-T as a research stance while running a test. The authors suggest roughly the next week where feasible. This is their experiment framework, not the familiar performance-goal expansion of SMART.

The same excerpt explicitly allows that an assumption may be accurate. It also separates observable information from inferred reactions and asks whether the actual test, including any departure from the plan, remains valid. Those qualifications matter more than producing a reassuring result.

For Lena, a routine internal update can provide an appropriate opportunity. A safety sign-off, regulatory submission or new financial commitment cannot become an experiment merely because it feels challenging. Keep the actual factual, confidentiality and approval checks. The test changes who writes and presents routine material; it does not suspend Lena’s decision responsibilities.

Before the first test, Lena confirms the normal working arrangement with Nell and Gideon on 2 November 2026. Nell owns preparation and presentation; Lena remains available for decisions and escalation. Both accept that arrangement. Their agreement does not establish how they will respond to every later difficulty, which is part of Lena’s learning question.

Check for a hidden substitution. If Lena rewrites the file overnight but lets Nell read it aloud, that event does not test her prediction about leaving preparation with Nell. Describe the different action accurately and redesign only if Lena chooses.

Follow a complete experiment and debrief

Mira and Lena agree three 50-minute coaching meetings on 2, 12 and 26 November. There is one optional written observation record due by noon on 27 November, with Mira’s response due by 17:00 on 1 December. Closing administration ends on 4 December at 17:00. This is a finite arrangement, not open-ended monitoring of workplace performance.

Lena chooses the ordinary update on 5 November. Nell agrees to present her own wording after the normal factual checks. Lena will not rewrite it or answer a question merely because silence feels uncomfortable. If a question concerns Lena’s authority, confidential information or a material risk, she will act within her role.

Mira: What do you predict if you leave Nell’s wording alone?

Lena: She will have one gap, and Gideon will say I should prepare everything.

Mira: Which part would you be able to check?

Lena: What he actually asks and whether he changes who prepares the next update.

Mira: What would make this unsuitable?

Lena: A new commitment or sensitive information. We would use the ordinary approval process.

On 5 November Nell presents without Lena rewriting. She cannot answer a question about a supplier’s revised date. She says Ivo, the scheduling owner, can verify it. Gideon asks Lena about a separate material delay she learned of on 4 November but had not escalated. He says, “I expect you to flag that delay promptly. Nell can continue preparing the routine updates.”

That criticism is relevant and remains in the record. Lena’s decision role was not delegated to Nell, and the experiment did not excuse missing escalation. Gideon does not require Lena to resume preparation. Nell asks Ivo for confirmation at noon; he accepts at 13:00 that he will answer by noon on 9 November. He replies at 11:00, Nell sends the corrected date at 11:30, and Gideon acknowledges it on 10 November.

Lena records her discomfort and urge to intervene separately from those events. At the 12 November debrief she initially says, “He thought I was incompetent.” Mira asks what supports that interpretation. Lena can identify the criticism, but not an explicit judgment of general incompetence or the predicted change in preparation ownership.

They do not replace her interpretation with “Gideon trusted me.” That would also exceed the observations. Lena concludes that criticism can coexist with Nell retaining preparation and Lena retaining decisions. She still wants to improve escalation. The unanswered question needed an accepted follow-up, not an assumption that every gap had to be filled by Lena in the meeting.

Keep an evidence record that permits uncertainty

This is an original record, not a reproduced publisher worksheet. Keep enough detail to answer the learning question without building a dossier on colleagues. No covert recording is required. If an observer would help, ask what they are willing to notice and share; do not assign them responsibility for proving the client’s belief wrong.

If Gideon had actually withdrawn preparation ownership, that would support the prediction under those conditions. Mira and Lena could investigate the role expectation or stop. A bigger test intended to force a positive result would not be a neutral response to unwelcome information.

Run a second test without declaring the issue solved

Lena chooses the next routine update for 19 November. Nell agrees on 13 November to prepare it and to name unresolved schedule questions rather than hide them. Ivo accepts on 16 November that any question sent after the meeting will be answered by noon on 20 November. Lena keeps normal escalation and factual checks.

At the second update Nell answers one question and refers another to Ivo through the agreed route. Lena handles a decision within her remit without taking over the presentation. Ivo replies at 11:00 on 20 November. Gideon reads the updated schedule on 23 November and leaves the preparation arrangement unchanged. This is a second observation, not a representative sample of all future meetings.

The pattern then recurs. On 24 November Lena rewrites a different routine slide without asking Nell, saying she wanted to prevent another difficult question. Nell objects to the unannounced takeover. Lena acknowledges it on 25 November, returns wording ownership and checks the factual issue with her rather than dictating the whole slide. Nell accepts that limited repair and posts her own revised wording.

At their 26 November coaching meeting, Mira initially calls the two tests “proof that Lena can delegate now.” Lena rejects that conclusion, pointing to the takeover two days earlier and the escalation error. Mira withdraws the claim. They retain the narrower learning that a question or criticism does not necessarily change preparation ownership, while Lena can still over-intervene.

Lena’s chosen next practice is to ask what support Nell wants before offering a rewrite, while flagging material delays promptly. There is no additional trial before the record closes. Whether that practice holds in a higher-stakes situation remains unknown, and no sponsor receives a claim of permanent transformation.

Use a five-step experiment checklist

1. Choose one small action. Specify the situation, the behavior that changes and the authority that stays. Lena leaves routine wording with Nell while retaining her own escalation duty.
2. Write the prediction. Name the expected condition and consequence before the event. Separate a difficult question, criticism and reassigned responsibility rather than treating them as the same outcome.
3. Choose proportionate observations. Record actual words, actions and decisions alongside the client’s experience. Arrange any observer’s participation openly; a facial expression is not proof of a private belief.
4. Compare interpretations. Ask which links occurred, which did not and what remains unknown. Check whether the action actually differed from the usual pattern and whether necessary safeguards stayed intact.
5. Choose the next decision. Continue, narrow the context, clarify a role, redesign or stop. Useful information can support a concern as well as challenge it. Bigger is not automatically better.

Check whether another response is needed

HSE’s Management Standards on Demands connect work demands with agreed hours, skills and the ability to raise concerns. That is a reason to inspect the actual work system. A map cannot supply missing capacity, train a colleague or change a policy that reserves approval to a particular role.

For example, if Lena must authorize an expenditure, refusing to review it is not a sound assumption test. She can ask the authorized policy owner whether a different approval arrangement is possible. Until it is accepted, the existing requirement remains. Personal willingness does not create delegated authority.

A team may also lack the information needed for the task. Establish who can provide it, whether that person accepts the deadline and what happens if they cannot. Calling the gap “resistance” can obscure a real dependency. Lena’s supplier follow-up required Ivo’s actual acceptance, not merely Nell naming him in the meeting.

If the client introduces earlier experiences or relational loyalties, let them decide what is relevant. Do not infer a childhood cause or demand disclosure to complete the map. Nonclinical coaching does not authorize anxiety treatment or trauma processing. Significant distress or a need for clinical treatment calls for an appropriate qualified route and a discussion of coaching scope.

Keep adult-development and team claims in proportion

Kegan’s broader meaning-making framework can inform reflection, but a delegation conversation does not establish a stage or movement to a higher one. The practical evidence here concerns two routine updates, a later takeover and a repaired account. Lena can gain a useful distinction while still valuing approval and still making mistakes.

A collective inquiry should concern shared work people can discuss, such as waiting for every uncertainty to disappear before deciding. Ask each participant what actually constrains the work. Do not combine private individual maps into a team diagnosis or infer that everyone shares a senior leader’s competing commitment.

Suppose a project group wants to decide earlier but finance authorization is missing. The team can examine its decision threshold while the finance owner separately reviews the real authority gap. If one colleague declines to disclose personal fears, that refusal need not prevent them contributing factual information. An invitation from someone who affects careers is not pressure-free by definition.

Agree what leaves the group, who receives it and who carries the work of any experiment. The peer-coaching triad guide provides a different setting for structured reflection. Neither arrangement justifies forwarding a private map to a sponsor without the person’s agreement.

Prepare the map and follow-up in Coachful

We build Coachful, so this reflects our product perspective. Coachful is our #1 recommendation for organizing a client-chosen resource, agreed action and follow-up. This is an editorial recommendation, not an independent market ranking or a claim that software diagnoses an immunity or scores developmental growth.

Create an original worksheet with the map areas and a separate observation record. Agree which responses the client wants the coach to see. Personal reflection can stay in their own notes or an unshared journal page. Where Share with coach is enabled, choosing it intentionally exposes that page; keep sensitive words unshared if the client wants them private. Check the sharing choice for each journal or takeaway page before relying on its privacy.

Deliberately grant the appropriate worksheet and send its private client-access link. A staff preview is not a client response, and a standalone assignment is not guaranteed to appear in My Coaching’s What to do next without the supported session or active-Program context. Check the actual client entry route before giving instructions.

In this case, Lena chooses a short shared action record rather than sending her personal fear notes. She opens the granted link and submits her observations at 11:00 on 27 November. Mira’s internal Coach review record is coach-only; saving it would not deliver the promised response. Mira must send the agreed client-facing explanation separately.

Mira sends that email at 10:00 on 2 December, 17 hours after the agreed deadline. Lena reads at noon and says the delay was frustrating. Mira acknowledges the miss at 15:00. They do not erase it because the three coaching meetings occurred. The 150 minutes counted earlier are live coaching only; preparation, worksheet review, email handling, business meetings and colleagues’ receiving work are not measured by that number.

On 3 December at 09:00 Lena approves a specific limited sponsor summary: the work goal, the two routine tests, the escalation issue, the later takeover and the chosen next practice, with future performance unknown. It excludes the private map and personal fears. Mira sends that approved version at 10:00; Gideon acknowledges it at 15:00. Lena reads the closing confirmation on 4 December at 09:00, before the 17:00 end.

Record the chosen action and review date in the normal client plan; completing a task does not confirm the assumption’s validity. Check session-recap settings before storing sensitive notes because automatic recaps can be enabled by default. A private internal review, client-facing email and agreed sponsor report are different records with different audiences. Explore Coachful’s features for the supported preparation and follow-up tools.

Frequently asked questions

Must the experiment disprove the assumption?

No. It can support a concern, identify conditions or leave uncertainty. Compare the actual prediction with observations rather than defining any discomfort as failure or any polite response as success. The assumption may be partly accurate.

What if the client finds no competing commitment?

Check whether the goal is theirs and whether resources, authority, information or skills explain the behavior. Leave the map unfinished or choose a different approach. Completing the table does not justify inventing a hidden motive.

Can this be used for anxiety treatment?

A coaching adaptation supports a chosen goal within competence and an agreed scope. It does not diagnose or treat an anxiety disorder, authorize trauma work or replace qualified healthcare. Discuss the appropriate support if the client’s needs exceed coaching.

How many tests establish lasting change?

There is no count in this example that establishes permanence. Two routine observations and one later recurrence support a limited account. Review the actual context, safeguards, actions and consequences, then let the client choose whether another test is useful.