Russia Finally Has a Validated Alliance Measure – and Bordin's Three Factors Collapsed Into One

The therapeutic alliance – the collaborative bond between client and therapist – is among the most reliable predictors of psychotherapy outcome anywhere it has been studied. Yet Russian-speaking clinicians have had no psychometrically validated instrument to measure it in the psychologist-client dyad. A seven-author team spanning RANEPA (Moscow), Kuban State University (Krasnodar) and the Institute of Psychology of the Russian Academy of Sciences has now published the first such adaptation – and the data hold a structural surprise.

A single factor, not three

The team translated the Working Alliance Inventory-Short (WAI-S) and administered both client- and therapist-report versions after the fourth session, alongside the Session Rating Scale (SRS) for convergent validity. The sample was 226 clients (mean age 24) and 24 therapists, who returned 228 ratings; the alliance correlations were computed on 139 complete client-therapist dyads.

Bordin's canonical three-factor model – task, bond and goal – did not fit. In the client version the confirmatory model was frankly poor (RMSEA = 0.237, CFI = 0.658); the therapist version fared better but still fell short of conventional thresholds. Exploratory analysis pulled a single general-alliance factor out of both versions, describing 61.7% of the variance in the 11-item client scale and 68.1% in the 10-item therapist scale. Internal consistency was very high (Cronbach's alpha = 0.945 and 0.954 respectively) – so high that it points to a largely unidimensional construct rather than three separable subscales.

Two convergent signals were more sober. Client and therapist ratings of the same alliance agreed only moderately (Spearman rho = 0.413), and correlations with the SRS were modest (rho 0.305 to 0.346). Both are consistent with the wider literature, in which client- and therapist-rated alliance routinely diverge and the client's view predicts outcome better.

What it means for the field

The headline is not the factor structure but the fact that the measure exists at all. A validated Russian WAI-S is infrastructure: it gives services, trainers and researchers a shared yardstick, which is the precondition for any serious move toward routine outcome monitoring or measurement-based care in Russian-language settings. The collapse to one factor also echoes another short-form adaptation (the French), where the WAI's subscales are so intercorrelated that they behave as one.

The honest limits are large. This is a cross-sectional, single-time-point validation – it establishes that the instrument measures something coherent, not that the alliance predicts outcome in Russian practice; that study still has to be done. The sample is a young convenience sample drawn from mixed modalities with only 24 therapists, and a unidimensional scale cannot separate a task rupture from a bond rupture – exactly the distinction clinicians use when they repair an alliance. For now, Russian practitioners gain a number they can trust; the mechanism and the outcome link remain open.