Robert Elliott University of Strathclyde

Person-Centered Science: What We Know and How

We Can Learn More about Humanistic/Person-

Centered/Experiential Psychotherapies

Robert ElliottUniversity of Strathclyde

OutlineHistorical IntroductionQuestion 1: What have we learned from

existing quantitative research on Humanistic/Person-Centred/Experiential therapies?

Question 2: What have we learned from existing qualitative research on Humanistic/Person-Centred/Experiential therapies?

Question 3: How can we learn more?

Context: Carl Rogers as Psychotherapy Research

PioneerInnovations:

Use of voice recording technology Psychotherapy process research Controlled outcome research Modern process-outcome research

Humanistic Therapy in Eclipse

Rogers gave up scientific research when he moved to La Jolla

Lack of research 1965 - 1990 hurt scientific & academic standing of humanistic therapy Led to humanistic therapies being

marginalized

Humanistic Therapy Revival

Since 1990: Rise of qualitative research Re-engagement in quantitative research Newer therapies (e.g., Focusing-oriented,

Process-Experiential/Emotion-Focused Therapy, Pre-therapy)

Available outcome research has tripled

Current situationDanger of split between:

Practitioners and training schools: reject quantitative research in favor of qualitative research

Small cadre of academic researchers: doing quantitative outcome research in order to gain official recognition

Question 1a: What Does Positivist Outcome Research

Tell Us?Humanistic/Person-Centred/Experiential

(HPCE) meta-analysis projectMeta-analysis: analysis of results

Effect size = standardized difference statistic

Creates a common for comparing results

Change E.S. =

m

pre

− m

post

sd

( pooled )

The HPCE Meta-Analysis Project

1st Generation: Greenberg, Elliott & Lietaer, 1994 (n= 36 studies) ….

5th Generation: Elliott & Freire (2008): Supported by a grant from the British Association

for the Person-Centred Approach 180+ studies 200+ samples of clients >13,000 clients 60 controlled studies (vs. no therapy or waitlist) 110 comparative studies (vs. HPCE therapies)

Elliott & Freire (2008) Meta-analysis Preliminary Results

1. HPCE therapies associated with large pre-post client change Effect size: 1.03 sd [standard deviation units] = a very large effect

2. Clients’ large posttherapy gains are maintained over early & late follow-ups Post: .95sd => early follow-up: 1.08sd => late

follow-up (12+ months): 1.14

Elliott & Freire (2008) Meta-analysis Preliminary Results

3. Clients in HPCE therapies show large gains relative to untreated clients Effect size: .81 sd = a large effect size Proves therapy causes client change.

Elliott & Freire (2008) Meta-analysis Preliminary Results

4. HPCE therapies in general are clinically and statistically equivalent when compared to other treatments (combining CBT and other therapies) Effect size: .01 sd = no difference in amount of change Held true even when we only considered

randomized (“gold standard”) studies

Elliott & Freire (2008) Meta-analysis Results

5. Comparison to Cognitive-Behavior Therapy (CBT): HPCE therapies as a group slightly

but trivially less effective than CBT: Effect size: -.18 sd =trivially worse (a small effect)

But…

Elliott & Freire (2008) Meta-analysis Results

6. Researcher theoretical allegiance effects strongly predict comparative ES: Correlation between comparative ES and

theoretical allegiance of researcher: -.52 CBT-oriented researchers => worse effects for

HPCE Small negative effect for HPCE therapies

vs. CBT disappears after statistically controlling for researcher allegiance

Where does researcher allegiance effect come

from?Big differences in how different HPCE therapies

do in comparison to CBT

Type HPCE Therapy N Comparative ES

Nondirective/ supportive

37 -.36 (=worse)

Person-centred 22 -.09 (=equivalent)

Emotion-Focused 6 +.60 (=better)

Other experiential 10 -.14 (=equivalent)

What is “Nondirective/ Supportive” Therapy?

Nondirective/supportive: 87% studies carried out by CBT Researchers

(40/46 in total sample) 65% explicitly labelled as “controls” (30/46) 52% involve non bona fide therapies (24/46) 76% of researchers are North American (35/46) 61% involve depressed or anxious clients (28/46)

The Moral of this Story:We don’t have to be afraid of

quantitative research or RCTsBut if we let others define our reality, we

are going to be in trouble.

Therefore, we need to do our own outcome research… including RCTs

Question 1b: What does Quantitative Process-Outcome

Research Tell Us?Process-outcome research predicts outcome

from in-therapy process measures, e.g., therapist empathy

Best-known process variable is Therapeutic Alliance Most common measure: Working Alliance

InventoryMeta-analyses show that alliance predicts

outcome: e.g., Horvath & Bedi, 2002; n = 90 studies: mean r = .21

Process-Outcome Research on Therapist

Empathy Therapist empathy is one of the

strongest predictors of outcome Bohart et al. (2002) meta-analysis 47 studies: mean r = .32

Accounts for about 10% of the variance in outcome

Interpretation of r = .321. Optimist’s view: 10% is a lot!

One of the best predictors of outcome Maybe even better that therapeutic alliance

Interpretation of r = .322. Pessimist’s view: The glass is 90%

empty! Rogers’ “necessary & sufficient” predicts

perfect correlation (r = 1.0) r = .32 decisively refutes Rogers’

hypothesis

Interpretation of r = .323. Optimist’s rebuttal: 10% is almost 100% of

what we can reasonably expect from the real world Client individual differences in problem severity

and resources predict most of outcome Measurement error Restriction of range (not enough unempathic

therapists!) Other stuff

Interpretation of r = .324. Pessimist’s plea: I still want the other

90%…

Question 2: What does Qualitative Research

Tell Us?Rogers’ Process Equation was based on proto-qualitative research: Years of careful observation of productive

and unproductive therapy sessionsSystematic qualitative research is a

relatively recent developmentBut mature enough now to allow a few

small qualitative meta-analyses

1. Helpful and Hindering Factors

Greenberg et al. (1994)Reviewed 14 studies of HPCE therapiesSelected 5 most frequent helpful and 3

most frequent hindering aspects14 categories of Helpful aspects,

grouped into 4 larger domains

Most Common Helpful Aspects in HPCE therapies1. Positive Relational Environment (7 out of

14 data sets; e.g., empathy) =>2. Client's Therapeutic Work (13 sets)

Most common : Self-Disclosure, Involvement =>3. Therapist Facilitation of Client's Work (6

sets; e.g., fostering exploration) =>4. Client Changes or Impacts (12 sets)

Most common: Understanding/ Insight, Awareness/Experiencing

Most Common Hindering Aspects

Much less common; difficult to studyMost common: Intrusiveness/

Pressure Even in person-centered therapy

Also present: Confusion/Distraction (derailing the client's

process) Insufficient Therapist Direction

2. Client Post-therapy Changes

Qualitative outcomeJersak, Magana and Elliott (2000; in

Elliott, 2002)5 studies, mostly Process-Experiential

for depression or trauma

Jersak et al. (2000)

Vitalizing the Self: Internal change4 subprocesses:

Leaving Distress Behind => Increased Contact with Emotional Self => Improved Self-esteem => Increased Sense of Personal

Power/Coping/Self-control Describe the first phase of a metaphorical

journey

Jersak et al. (2000)Changes in the Self’s Relationships to

Others/World:3 subprocesses:

Defining Self with Others/Asserting Independence

Engaging with Others, Experiencing the World More/Mobilizing Self to

Act in the World Describe the outward phase of the client’s

journey

3. Effects of significant therapy events

Timulak (2007)7 studies, most HPCE9 common categoriesAll 7 studies:

Awareness/Insight/Self-Awareness Reassurance/Support/Safety

More than half the studies: Behavior Change/Problem Solution Exploring Feelings/Emotional Experiencing Feeling Understood.

Implication: Qualitative Studies of HPCE

May be possible to integrate these 3 types of research into a model of HPCE change process

Framework: Helpful (hindering) aspects => Immediate effects (significant events) => Qualitative outcome

Question 3: How Can We Learn More?

1. Be Methodologically Pluralist

Most sensible course of action:To encourage both kinds of research

Render politically expedient quantitative data to the government and professional bodies (“Caesar”)

Simulaneously carry out qualitative research that completely honors person-centered principles

Even in the same study

2. Follow Person-Centred Research

PrinciplesE.g., Mearns & McLeod (1984) (1) Empathy. Understand, from the inside, the

research participant’s (client or therapist) lived experiencing

(2) Unconditional Positive Regard. Accept/prize the research participant’s experiencing,

(3) Genuineness. Be an authentic/equal partner with the research participant: participant = co-researcher; researcher = a fellow human being.

(4) Flexibility. Creatively and flexibly adapt research methods to the research topic and questions at hand

Applying Person-centred principles to different

types of researchFairly easy to see application to qualitative

research, e.g., Clarifying expectations and other researcher pre-

understandings; Negotiating nature of participation with informant

in a transparent, collaborative manner; Carrying out data collection in a careful, intentional

manner, including helping informant stay focused and clarifying their meanings; etc.

Person-Centred Principles Apply Equally to

Quantitative ResearchAlways put the participant’s needs

ahead of yoursTreating participants disrespectfully and

inconsistently leads to resentment and sloppy, invalid data

A questionnaire is a form of relationship

Person-Centred Principles Apply Equally to Quantitative

ResearchA research participant will feel misunderstood

and uncared for by a confusing questionnaire layout or an overly hot or noisy research room

An ill-prepared research packet or an anxious interviewer can betray a lack of genuine commitment by the researcher

All of our criticisms of quantitative research are really criticisms of bad research, of any kind

3. Focus on Change Process Research

Much current research on HPCE therapies does not focus on how change occurs

Needed as complement to outcome research & improve therapy

Select from different genres of change process research

a. Important preliminary: Basic outcome research

What are the effects of HPCE therapies with specific client populations?

Can be quantitative or qualitativeSingle client or group of clientsStandard questions or individualizedSee Elliott & Zucconi (2006) for suggestions to

implement in practice and training settings

Necessary starting point for Change Process research

b. Process-Outcome Research

Quantitative genre: Measure process (e.g., empathy) => predict outcome

HPCE’s not studied enough with this approach: Only 6 out of 47 studies in Bohart et al.

(2002) empathy-outcome meta-analysis were HPCE therapies

Highly appropriate to naturalistic samples

c. Helpful Factors Research

Qualitative genre: Interview (e.g., Change Interview) Helpful Aspects of Therapy (HAT) Form Analyze with variety of methods, e.g.,

Grounded Theory, discourse analysis

d. Micro-analytic Sequential Process

ResearchExamine turn-by-turn interaction

between client and therapistQuantitative: client and therapist

process measures (e.g., client experiencing and therapist empathy)

Qualitative: Task analysis or Conversation analysis

e. Complex Change Process Research MethodsCombine genres to develop richer pictureBalance strengths, limitationsExamples:

Assimilation Model (Stiles et al., 1990) Task Analysis (Rice & Greenberg, 1984) Comprehensive Process Analysis (Elliott, 1989) Hermeneutic Single Case Efficacy Design (Elliott,

2002)

4. Get Involved!Elliott & Zucconi (2006): International

Project on Psychotherapy and Psychotherapy Training (IPEPPT)

The project is to stimulate practice-based research, especial in training centres

Have developed a set of sample research protocols to choose form

Further Suggestions (Elliott & Zucconi, 2006)

(1) Contribute to dialogues on how to measure therapy and training outcomes within HPCE therapies

(2) Set an example for students and colleagues by carrying out simple research procedures with your own clients and in your own training setting

(3) Help to develop specialized research protocols for particular client populations (e.g., people living with schizophrenia)

Further Suggestions (Elliott & Zucconi, 2006)

(4) Contribute to method research aimed at improving existing quantitative and qualitative instruments

(5) Take part in more formal collaborations with similarly-inclined training centers to generate data for shared research

Robert Elliott: [email protected]: pe-eft.blogspot.com