According to Bordin's 1979 formulation, what are the three components of the working alliance?
Choose an answer, then check it.
Social Work & Human Services · Communication
Helping Relationships
On this page 9 sections
In 30 seconds
A Helping relationship A bounded professional partnership formed for the benefit of the person seeking help, defined by an agreed purpose, an explicit structure, external accountability, and disclosure that runs mostly one way. Full entry → is a working partnership formed for someone else's benefit, carrying a purpose, a structure, and an accountability that friendship does not. Research across helping settings finds that the quality of that partnership - agreement on goals, agreement on tasks, and the bond between two people - predicts how the work goes, reliably but modestly. The relationship is not the point. It is what makes the work possible, and it has to be built, repaired, and ended on purpose.
Why this matters
Every method you learn later - assessment, case planning, crisis response, referral - is delivered through a relationship, and it fails when the relationship does not hold. Courses treat this as an early chapter, but field placements, licensing exams, and supervision return to it constantly, usually by asking what you did when someone disengaged or when your own authority was part of the problem. Knowing the actual size of the alliance evidence, and its limits, also protects you from the two standard errors: dismissing the relationship as soft, or treating warmth as though it were an intervention. One of those errors gets people ignored. The other gets them befriended instead of helped.
The college version
What makes it a professional relationship
A helping relationship is organized around one person's stated purpose and another person's professional role. Four features separate it from friendship, mentorship, or family support. First, purpose: it exists to accomplish something identifiable, and it ends when that work ends or stops being useful. Friendship has no completion condition. Second, structure: time, place, frequency, documentation, and often payment or a statutory mandate are set in advance rather than negotiated by mood. Third, accountability: the practitioner answers to an employing agency, a funder, a licensing board, and a professional code. The NASW Code, for instance, makes clients' interests primary but says plainly that legal obligations and responsibility to the wider society can on limited occasions supersede loyalty to a client, and that clients should be told this. A friend answers to no one. Fourth, asymmetry of disclosure: the client is expected to say difficult things about their own life, and the practitioner is not. Information runs mostly one direction, and so does the exposure that goes with it. None of this makes the relationship cold; it makes it legible. Someone who knows what the relationship is for, how long it lasts, who reads the record, and what the practitioner is required to report can make real decisions inside it. Warmth without that structure is not help. It is a pleasant conversation with unclear stakes, and the person leaves without knowing what they agreed to.
Rogers' core conditions and what the evidence actually shows
Carl Rogers argued in 1957 that a set of relationship conditions were necessary and sufficient for constructive personality change. Three are the ones you will hear named: empathic understanding of the client's internal frame of reference, communicated back to them; Unconditional positive regard Nonjudgmental valuing of a person that does not rise or fall with approval of their choices or their progress. Full entry →, meaning a nonjudgmental prizing of the person that does not fluctuate with approval of their choices; and Congruence Genuineness on the practitioner's part - being an aware, integrated person in the relationship rather than operating a professional persona. Full entry →, meaning the practitioner is genuine and integrated in the relationship rather than operating a professional persona. The evidence since supports the contribution and not the sufficiency. Three updated meta-analyses published in 2018 put each condition in the same modest range. Empathy, across 82 independent samples and 6,138 clients, correlated with outcome at r = .28, which the reviewers state accounts for roughly nine percent of outcome variance. Positive regard across 64 studies gave g = .28, rising to .36 under a multilevel model. Congruence across 21 studies and 1,192 patients gave r = .23. These are replicated associations, and they are not sufficiency. Something explaining under a tenth of the variation in outcomes cannot be the whole causal account. The empathy reviewers add a detail worth keeping: at least one study found clients who preferred a business-like practitioner to a warm one, and a technically empathic response can be intrusive when it misjudges how much closeness a person wants. The skill is attunement, not maximum warmth applied uniformly. Treat the Core conditions Rogers' three practitioner-provided relationship qualities: empathic understanding, unconditional positive regard, and congruence. Full entry → as well-supported contributors most helping work depends on, and treat Rogers' sufficiency claim as a hypothesis the evidence has not established.
The working alliance: goals, tasks, and bond
Edward Bordin proposed in 1979 that the collaborative core of a helping relationship, in any theoretical tradition, has three components: agreement on the goals of the work, consensus on the tasks that make up the work, and the bond between the two people. Two of the three are agreements rather than feelings, and that is the part students skip. A client can like you enormously and still have no idea what the two of you are working toward or why the forms matter. In Bordin's terms that relationship is weak, not strong. Bordin also expected the profile to differ by approach: different kinds of work place different demands on the relationship. The alliance is among the most heavily studied variables in the helping literature. Flueckiger and colleagues synthesized 295 independent studies covering more than 30,000 patients published between 1978 and 2017 and found an overall alliance-outcome correlation of r = .278 (95% confidence interval .256 to .299), equivalent to d = .579. The association held across who rated the alliance, which instrument was used, treatment approach, patient characteristics, and country, and was essentially identical for internet-delivered therapy at r = .275. Report that number at its real size. An r of .278 means alliance and outcome share about eight percent of their variance: reliable, not decisive. The causal direction is genuinely debated, because nobody can randomly assign people to a strong or weak alliance, so the causal question is settled only indirectly. The authors tested the leading rival account - that a good alliance is a by-product of improvement that had already started - by comparing zero-order with partial correlations in the 66 studies reporting both, and found no significant difference (about .25 against .22). That weakens the by-product explanation without proving causation. Note the scope, too: this literature is overwhelmingly psychotherapy, mostly North American and European, with ethnic minority clients underrepresented outside substance-use samples. Carrying it over to case management, child welfare, or housing work is a reasonable extrapolation, not a demonstrated result.
Engagement, the first meeting, rupture and repair
Engagement The process of turning a referral or a mandate into a working partnership, concentrated in but not limited to the earliest contacts. Full entry → is the work of turning a referral into a relationship, and the first meeting carries a disproportionate share of it. First contacts arrive in different forms - a crisis, a voluntary request, an involuntary mandate - and knowing which one you are walking into is a practice skill, not a formality. What that meeting has to establish is unglamorous: what the service can and cannot do, how long it lasts, who reads the record, what you are obliged to report, and what the person actually wants out of it. Relationships in this work strain, and the strain has a name. A Rupture A deterioration in the collaborative relationship, appearing as disagreement about goals, breakdown of joint work on tasks, or strain in the bond. Full entry → is a deterioration in the alliance, showing as disagreement about goals, a breakdown of collaboration on tasks, or a strain in the bond. Ruptures are ordinary and frequently invisible unless someone names them: the client who agrees with everything and commits to nothing, the missed appointments after a difficult session. Eubanks, Muran and Safran's 2018 meta-analysis of 11 studies with 1,314 patients found rupture-repair episodes related to better outcome at r = .29, d = .62. A companion meta-analysis of 6 studies asked a different question - whether training practitioners in rupture resolution improved their clients' outcomes - and found no significant effect, r = .11. Both belong in the same sentence. Repair is associated with better outcomes; teaching people to repair has not yet been shown to reliably produce them. That gap is a live research problem, not a reason to stop naming ruptures.
Involuntary and mandated clients
A large share of helping work is not voluntary. People arrive through court orders, child protection cases, probation conditions, school discipline, and conditions attached to benefits or housing. Calling their reluctance resistance misdescribes it. Someone ordered to meet a stranger who can write a report affecting whether they keep their children is behaving rationally when they disclose little and watch what you do with what they say. The honest response is to make the authority explicit rather than soften it with warmth. Ethical practice requires telling clients who receive services involuntarily what the service involves, how far it extends, and how far their right to refuse actually goes - which means naming the consequences of refusing as well. That conversation is often the first genuine thing that happens in the relationship, because it shows the person you will not pretend the power difference away. Bordin's framework still applies, but goal agreement has to be worked for rather than assumed. There is frequently a mandated goal the person did not choose and, separately, a goal they would choose; naming both and writing both down is more workable than pretending they are the same thing. Rapport built by minimizing your authority collapses the first time you exercise it.
Endings are part of the relationship
Termination The ending of services, treated as a practice decision about whether the work is still needed and still useful, with notice and transfer where required. Full entry → is a practice decision, not the paperwork after one. Ethical standards frame it that way: services should end when they are no longer required or no longer serve the person's needs and interests; practitioners should avoid abandoning clients who still need services; and where services are ending, clients should be notified promptly and helped to transfer to another provider when continued service is needed. A planned ending does real work. It reviews what changed and what did not, names what the person did rather than what the service did, identifies what would signal a need to come back, and states clearly whether the door is open. Unplanned endings are constant - funding lapses, caseloads transfer, people move or simply stop coming - and the practitioner's job is to make the ending as non-arbitrary as circumstances allow, including a letter or call that says what happened. For someone whose previous relationships with institutions ended without explanation, a clean ending is itself part of the intervention rather than an administrative courtesy.
Use of self, and the maintenance the work requires
Use of self The practitioner's own attention, manner, reactions, and judgment considered as the working instrument of the intervention. Full entry → names the fact that the practitioner is the instrument. Attention, tone, pacing, tolerance for distress, and willingness to be affected are the working parts, and instruments drift. The occupational risks have distinct names. Burnout is exhaustion and disengagement produced by chronic workload and organizational conditions. Secondary traumatic stress A pattern of trauma symptoms arising from repeated indirect exposure to the traumatic experiences of the people one serves. Full entry → is a trauma symptom pattern arising from repeated indirect exposure to other people's traumatic experiences. Bride's survey of social workers in direct practice reported that secondary exposure through the work is highly likely, that many respondents experience at least some secondary traumatic stress symptoms, and that a significant minority may meet diagnostic criteria for PTSD. Supervision is the maintenance this requires, and it is a structural provision rather than a personal virtue. Federal child welfare guidance describes effective supervision as essential to service quality, to reducing employee stress, and to worker retention, spanning administrative, educational, and supportive functions, and specifically as creating a space in which workers can manage the secondary trauma they encounter. The same guidance notes that a positive supervisor-worker relationship can model the collaborative partnership workers are asked to build with families - the alliance concept applied one level up. Framing self-care as purely an individual duty misplaces the load: the profession's own ethical guidance calls professional self-care central to competent practice while also directing organizations to adopt policies that support practitioner wellbeing.
Eli explains
The same idea, in plain words
Explain it like I’m 10
A helping relationship is a partnership with a job to do, and two things have to be true at once. The person has to feel that you actually understand them and are not judging them - that is the warm half. And the two of you have to agree on what you are trying to change and on what each of you will do about it - that is the working half. Most people assume the warm half is the whole thing. It is not. Being liked is not the same as being useful, and the research says the whole partnership only predicts how things turn out modestly, not magically. This relationship also has rules a friendship does not have. It has an ending. Someone else checks your work. You do not get to bring your own problems into it. And when the partnership goes wrong - which happens routinely - the fix is to say so out loud rather than hope it quietly settles.
Picture it like this
Think of a climbing rope team. The rope is the bond: it is what holds you together when someone slips. But a rope alone gets nobody up the mountain. The team also has to agree on which summit they are heading for and who carries what, or the rope just ties two people together while they walk in different directions.
Where the picture stops working
The analogy breaks in three places. On a rope team both climbers face the same risk, while in a helping relationship the client carries nearly all the exposure and does nearly all the disclosing. Second, climbers choose their mountain, but mandated clients often do not, and the practitioner may be the one holding the court's authority. Third, a rope is either attached or it is not, whereas an alliance is a matter of degree that shifts week to week - and it can be repaired mid-climb, which a snapped rope cannot.
Worked example
A parent is referred to a family support worker by a child protection agency after a substantiated neglect finding. The parent arrives twenty minutes late and answers in single words. A worker reading this as resistance pushes harder for engagement, with more warmth. A worker reading it as rational caution does something different: she names the report she is required to write, states what she must document and what she will not, and separates the agency's mandated goal (the home is safe) from a goal the parent names herself (regaining work hours so she can afford childcare). Both go on the plan, in the parent's own words where possible. Three weeks later the parent cancels twice after a case conference - a lapse in collaboration on the agreed tasks, not a scheduling problem. The worker names it directly at the next visit and asks what the conference was like from the other side of the table. The repair is not a technique. It is the parent discovering that disagreement can be raised without the relationship collapsing, which is also what she needs from every other institution in the case.
Key takeaway
The alliance is not rapport and it is not the goal: it is agreement on goals and tasks plus the bond, and it exists to make the work possible. The evidence for it is reliable but modest, so build it deliberately and never let warmth stand in for structure, honesty about your authority, or a planned ending.
Quick check
3 questions here, of 5 in this lesson’s practice set. Answers stay hidden until you check.
Flueckiger and colleagues' 2018 meta-analysis of 295 studies reported an overall alliance-outcome correlation of r = .278. Which statement reports that finding accurately?
Choose an answer, then check it.
A practitioner and client get along easily and never disagree, but the client cannot say what they are working on together and does not do the tasks in the plan. How should this relationship be judged in alliance terms?
Choose an answer, then check it.
Study tools & related lessonsYou’ll learn to · Common mistakes · Easily confused · Key vocabulary · Related
You’ll learn to
- Distinguish a professional helping relationship from friendship, mentorship, and family support by purpose, structure, accountability, and asymmetry of disclosure.
- Define Bordin's three components of the working alliance and explain why rapport alone does not satisfy them.
- Evaluate the meta-analytic evidence on the alliance and on Rogers' core conditions, including effect size and the limits of causal inference.
- Explain what an alliance rupture is and analyze what the repair evidence does and does not establish.
- Apply informed-consent obligations to work with involuntary and mandated clients, and reframe 'resistance' accordingly.
- Explain why planned endings, supervision, and attention to secondary traumatic stress belong to the relationship rather than to administration.
Common mistakes
Treating rapport as the goal, so that a comfortable, friendly relationship with no agreed goals or tasks is counted as a strong one.
Bordin's alliance has three components and two of them are agreements, not feelings. A warm relationship with no shared goal and no task consensus is a weak alliance and should be assessed as one. The alliance serves the work; it is not the work.
Presenting Rogers' core conditions as sufficient, as though empathy, unconditional positive regard, and congruence produce change on their own.
Rogers claimed sufficiency in 1957. The 2018 meta-analyses support a modest contribution - empathy r = .28, positive regard g = .28 to .36, congruence r = .23 - which is not sufficiency. Say 'well-supported contributors', and note that some clients prefer a business-like practitioner to a warm one.
Reporting the alliance-outcome link as though it were large, or as though it were established as causal.
Flueckiger and colleagues report r = .278 across 295 studies and more than 30,000 patients - roughly eight percent shared variance. Randomly assigning people to alliance levels is not possible, so causal claims rest on indirect evidence, which the authors themselves describe as controversial.
Calling an involuntary client's guardedness 'resistance' and trying to dissolve it with warmth or by downplaying the mandate.
Caution toward someone with authority over your family or your liberty is a rational response, not a personality defect. State the authority, the reporting obligations, and the real limits of the right to refuse, then work toward goals the person can genuinely endorse alongside the mandated one.
Treating termination as an administrative step that happens once the case closes.
Ethical standards make ending a practice decision: end when services are no longer needed or no longer serve the person's interests, avoid abandoning clients who still need help, give prompt notice, and assist transfer where continued service is required.
Easily confused
Friendship vs. Professional helping relationship
Friendship has no completion condition, no external accountability, and reciprocal disclosure. A helping relationship has a purpose that can be finished, a supervisor and a code above it, and disclosure that runs mostly one direction.
Rapport vs. Working alliance
Rapport is the felt ease between two people. The alliance additionally requires agreement on what the work is for and on what each party will do, so a relationship can have abundant rapport and almost no alliance.
Burnout vs. Secondary traumatic stress
Burnout is exhaustion and disengagement produced by chronic workload and organizational conditions. Secondary traumatic stress is a trauma symptom pattern produced by repeated indirect exposure to clients' traumatic material. They often co-occur, but they have different causes and different remedies.
Observed rupture repair vs. Rupture-resolution training
Repair episodes related to outcome at r = .29 across 11 studies, while training practitioners in rupture resolution showed no significant effect on client outcome (r = .11) across 6 studies. An association with a process is not evidence that teaching the process works.
Key vocabulary
- Helping relationship
- A bounded professional partnership formed for the benefit of the person seeking help, defined by an agreed purpose, an explicit structure, external accountability, and disclosure that runs mostly one way.
- Working alliance
- The collaborative element of a professional relationship, made up of agreement on goals, agreement on tasks, and the emotional bond between the two people (Bordin, 1979).
- Core conditions
- Rogers' three practitioner-provided relationship qualities: empathic understanding, unconditional positive regard, and congruence.
- Congruence
- Genuineness on the practitioner's part - being an aware, integrated person in the relationship rather than operating a professional persona.
- Unconditional positive regard
- Nonjudgmental valuing of a person that does not rise or fall with approval of their choices or their progress.
- Rupture
- A deterioration in the collaborative relationship, appearing as disagreement about goals, breakdown of joint work on tasks, or strain in the bond.
- Engagement
- The process of turning a referral or a mandate into a working partnership, concentrated in but not limited to the earliest contacts.
- Termination
- The ending of services, treated as a practice decision about whether the work is still needed and still useful, with notice and transfer where required.
- Use of self
- The practitioner's own attention, manner, reactions, and judgment considered as the working instrument of the intervention.
- Secondary traumatic stress
- A pattern of trauma symptoms arising from repeated indirect exposure to the traumatic experiences of the people one serves.
Sources & references
- The necessary and sufficient conditions of therapeutic personality change (Journal of Consulting Psychology, 21(2), 95-103) — Carl R. Rogers / American Psychological Association
- The generalizability of the psychoanalytic concept of the working alliance (Psychotherapy: Theory, Research & Practice, 16(3), 252-260) — Edward S. Bordin / American Psychological Association
- The Alliance in Adult Psychotherapy: A Meta-Analytic Synthesis (Psychotherapy, advance online publication, 24 May 2018) — Christoph Flueckiger, A. C. Del Re, Bruce E. Wampold and Adam O. Horvath / American Psychological Association, copy posted by the Society for the Advancement of Psychotherapy
- Therapist empathy and client outcome: An updated meta-analysis (Psychotherapy, 55(4), 399-410) — Robert Elliott, Arthur C. Bohart, Jeanne C. Watson and David Murphy / American Psychological Association; accepted manuscript in the University of Strathclyde repository
- Positive regard and psychotherapy outcome: A meta-analytic review (Psychotherapy, 55(4), 411-423) — Barry A. Farber, Jessica Y. Suzuki and Daisy E. Lynch / American Psychological Association
- Congruence/genuineness: A meta-analysis (Psychotherapy, 55(4), 424-433) — Gregory G. Kolden, Chia-Chiang Wang, Sara B. Austin, Yunling Chang and Marjorie H. Klein / American Psychological Association
- Alliance rupture repair: A meta-analysis (Psychotherapy, 55(4), 508-519) — Catherine F. Eubanks, J. Christopher Muran and Jeremy D. Safran / American Psychological Association
- NASW Code of Ethics (2021 revision), including Ethical Responsibilities to Clients — National Association of Social Workers
- Supervision — Child Welfare Information Gateway, Children's Bureau, Administration for Children and Families, U.S. Department of Health and Human Services
- Prevalence of secondary traumatic stress among social workers (Social Work, 52(1), 63-70) — Brian E. Bride / National Association of Social Workers, Oxford University Press
- Social Work Practice, in Introduction to Social Work: A Look Across the Profession — James Langford and Craig Keaton, Mavs Open Press, University of Texas at Arlington
EliExplains lessons are original prose written from the open, credible references above. See Copyright & Licensing.
Researched 2026-08-18
Educational content only. It is not medical, legal or professional advice. Found an error? Tell us.